System

Abstract

To provide a system.SOLUTION: A system includes: means for generating a semiconductor circuit design diagram avoiding a patent using a generative AI model which preliminarily learned patent information; means for confirming whether the generated semiconductor circuit design diagram avoids patent infringement; and means for regenerating the semiconductor circuit design diagram when the patent infringement is confirmed.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] Building and operating high-performance AI requires expensive semiconductors, and there is a need to reduce their costs and avoid dependency risks. In addition, companies want to differentiate themselves from other companies by developing their own semiconductors using semiconductor circuit design drawings that are publicly available in patents. [Means for solving the problem]

[0005] We provide a system that learns patent information in advance and automatically generates semiconductor circuit designs that avoid patents, making it possible to develop and manufacture unique, high-performance, low-cost semiconductors. [Brief explanation of the drawings]

[0006] [Figure 1]1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 2 is a sequence diagram showing a flow of processing in the data processing system according to the first embodiment of the first form example. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1. [Figure 13] FIG. 10 is a sequence diagram showing a processing flow of a data processing system in a second embodiment of the second form example. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Embodiment Example 2. [Figure 15] FIG. 10 is a sequence diagram showing the flow of processing in a data processing system according to a third embodiment of the third embodiment. [Figure 16] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Embodiment 3. [Figure 17] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Form Example 1 when an emotion engine is combined. [Figure 19] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the second embodiment of the second form example when an emotion engine is combined. [Figure 20] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Form Example 2 when an emotion engine is combined. [Figure 21] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the third embodiment of the third form example when an emotion engine is combined. [Figure 22] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Form Example 3 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0007] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0008] First, the terms used in the following description will be explained.

[0009] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, the processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0011] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0012] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0014] [First embodiment]

[0015] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0016] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0017] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0018] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0019] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0021] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0022] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0023] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0024] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0025] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0026] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0027] "Example 1"

[0028] In one embodiment of the present invention, a means for pre-learning patent information is to use an AI that learns from semiconductor circuit design drawings obtained from a patent database. This AI extracts and learns the features of semiconductor circuit design drawings published in patents.

[0029] "Example 2"

[0030] Furthermore, as a means of automatically generating semiconductor circuit designs that avoid patents, the system has the ability to automatically generate new semiconductor circuit designs that do not infringe patents based on the patent information it has learned. Specifically, the AI ​​generates new designs based on patent information and checks whether the designs infringe existing patents.

[0031] "Example 3"

[0032] In addition, as a means of developing and manufacturing unique high-performance, low-cost semiconductors, semiconductors are manufactured based on automatically generated semiconductor circuit design drawings. Specifically, the automatically generated design drawings are input into semiconductor manufacturing equipment, which then manufactures semiconductors based on the design drawings. This makes it possible to build and operate high-performance AI without using expensive semiconductors.

[0033] The processing flow of each embodiment will be described below.

[0034] "Example 1"

[0035] Step 1: Obtain patent information for semiconductor circuit design drawings from a patent database.

[0036] Step 2: The acquired patent information is input into the AI, which then learns from the patent information. Through this learning, the AI ​​extracts the characteristics of the semiconductor circuit design.

[0037] "Example 2"

[0038] Step 1: Based on the learned patent information, the artificial intelligence automatically generates new semiconductor circuit design drawings.

[0039] Step 2: Check whether the automatically generated semiconductor circuit design diagram infringes on any existing patents. If it is confirmed that there is no infringement, the design diagram is adopted.

[0040] "Example 3"

[0041] Step 1: The selected semiconductor circuit design is input into the semiconductor manufacturing equipment.

[0042] Step 2: Semiconductor manufacturing equipment produces semiconductors based on the blueprint. This manufacturing process makes it possible to build and operate high-performance AI without using expensive semiconductors.

[0043] Example 1

[0044] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0045] In conventional semiconductor circuit design, searching and analyzing patent information is often done manually, which is time-consuming and labor-intensive. It is also difficult to automatically generate design drawings that avoid patents, making it difficult to efficiently develop and manufacture high-performance, low-cost semiconductors. To solve these issues, a system that can efficiently learn patent information and automatically generate design drawings that avoid patents is needed.

[0046] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0047] In this invention, the server includes means for pre-learning patent information, means for acquiring semiconductor circuit design drawings from a patent database, means for extracting the acquired design drawings as image data, means for inputting the image data into an artificial intelligence model to extract and learn features of the design drawings, means for receiving prompts from a user and searching for related patent information, and means for providing the search results to the user. This enables efficient learning and analysis of patent information, and realizes the automatic generation of design drawings that avoid patents.

[0048] "Patent information" refers to the technical information and invention content described in patent documents.

[0049] "Pre-training" refers to the process of training a model with relevant data before it performs a specific task.

[0050] "Patent database" refers to an information system that collects, stores, and provides patent documents in a searchable format.

[0051] A "semiconductor circuit design drawing" refers to a drawing that shows the circuit structure and layout of a semiconductor device.

[0052] "Image data" refers to data that represents visual information in digital form.

[0053] An "artificial intelligence model" refers to an algorithm or system built using techniques such as machine learning and deep learning to perform a specific task.

[0054] "Feature extraction" refers to the process of identifying important patterns or attributes from data and converting them into a numerical representation.

[0055] "Prompt sentence" refers to text containing instructions or questions that a user enters into a system.

[0056] "Search Results" refers to the collection of relevant information returned by the system based on a user's query.

[0057] The present invention relates to a system for efficiently learning patent information and automatically generating semiconductor circuit design diagrams that avoid patents. Specific embodiments of this system will be described below.

[0058] The server generates a program for pre-learning patent information. This program retrieves semiconductor circuit design drawings from a patent database and has the AI ​​model learn them. Patent databases used include Google (registered trademark) Patents and Espacenet (registered trademark).

[0059] The server accesses a patent database to retrieve patent information, including semiconductor circuit design drawings. Since the retrieved patent documents are usually provided in PDF format, the server uses a PDF analysis library to extract the design drawings as image data.

[0060] The server then preprocesses the extracted image data and inputs it into an artificial intelligence model. Preprocessing includes image resizing and noise removal. The preprocessed image data is then input into an artificial intelligence model built using a deep learning framework such as TENSORFLOW® or PyTorch.

[0061] The AI ​​model uses a convolutional neural network (CNN) to extract and learn from the features of the blueprint. Specifically, it recognizes the patterns and circuit structures of the blueprint and saves them as feature vectors. The server stores these feature vectors in a database for later use in searches.

[0062] The user inputs a prompt sentence using a terminal, such as "Please tell me patent information on the latest semiconductor circuit designs." The server analyzes this prompt sentence and extracts related keywords and feature vectors.

[0063] The server searches the database for relevant patent information based on the extracted keywords and feature vectors. The search results are provided to the user via their terminal. Specifically, a list of related patent documents and thumbnail images of blueprints are displayed.

[0064] Examples:

[0065] If a user wants to search for patent information on new semiconductor circuit designs, they can input the following prompt into the generative AI model:

[0066] Example prompt sentence:

[0067] "Please tell me the patent information on the latest semiconductor circuit designs."

[0068] When this prompt is input into the generative AI model, the server retrieves the relevant semiconductor circuit design from the patent database, and the AI ​​model analyzes it to extract features, ultimately providing the user with relevant patent information.

[0069] The above is a specific implementation method of the system according to one embodiment of the present invention.

[0070] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0071] Step 1:

[0072] The server accesses a patent database and obtains patent information including semiconductor circuit design drawings.

[0073] Input: API endpoint of patent database, search query (e.g., keywords or classification code)

[0074] Data processing: Send API requests and retrieve patent documents

[0075] Output: Acquired patent document (PDF format, etc.)

[0076] Specific operation: The server uses the patent database API to search for patent documents based on the specified keywords and classification codes, and downloads the relevant patent documents.

[0077] Step 2:

[0078] The server extracts semiconductor circuit design drawings as image data from the acquired patent documents.

[0079] Input: Acquired patent document (PDF format, etc.)

[0080] Data processing: Extracting blueprints as image data using a PDF parsing library

[0081] Output: Extracted image data

[0082] Specific operation: The server uses a PDF parsing library to extract semiconductor circuit design diagrams from patent documents in image format.

[0083] Step 3:

[0084] The server pre-processes the extracted image data and inputs it into an artificial intelligence model.

[0085] Input: Extracted image data

[0086] Data processing: Perform preprocessing such as image resizing and noise removal

[0087] Output: Preprocessed image data

[0088] Specific operation: The server uses an image processing library (e.g., OpenCV) to resize and denoise the image data and convert it into a format suitable for the artificial intelligence model.

[0089] Step 4:

[0090] The server inputs the preprocessed image data into an artificial intelligence model, which extracts and learns the features of the blueprint.

[0091] Input: Preprocessed image data

[0092] Data processing: Feature extraction using artificial intelligence models (e.g., TensorFlow or PyTorch)

[0093] Output: Feature vector

[0094] Specific operation: The server uses a convolutional neural network (CNN) to extract the patterns and structures of the blueprints from the image data as feature vectors and stores them in a database.

[0095] Step 5:

[0096] The user uses the terminal to enter a prompt sentence.

[0097] Input: Prompt statement (e.g., "Please tell me patent information on the latest semiconductor circuit designs.")

[0098] Data processing: Analyze the prompt sentence and extract relevant keywords and feature vectors

[0099] Output: Parsed keywords and feature vectors

[0100] Specific operation: The user inputs a prompt sentence through the terminal interface and sends it to the server.

[0101] Step 6:

[0102] The server searches the database for relevant patent information based on the analyzed keywords and feature vectors.

[0103] Input: Parsed keywords and feature vectors

[0104] Data processing: generating database queries and performing searches

[0105] Output: Search results (list of relevant patent documents and thumbnail images of blueprints)

[0106] Specific operations: The server generates a database query to search for relevant patent information based on the stored feature vectors.

[0107] Step 7:

[0108] The server provides the search results to the user.

[0109] Input: Search results (list of relevant patent documents and thumbnail images of blueprints)

[0110] Data processing: converting search results into a format that can be displayed to the user

[0111] Output: Search results served to the user

[0112] Specific operation: The server sends the search results to the user's terminal and displays them in a format that the user can view.

[0113] (Application example 1)

[0114] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0115] In conventional semiconductor circuit design, it was difficult to utilize patent information to select optimal designs and adjust manufacturing processes in real time. Furthermore, there was a lack of efficient means to develop and manufacture high-performance, low-cost electronic components while avoiding patents. This led to problems such as insufficient optimization of manufacturing processes and quality control, resulting in a decline in product competitiveness.

[0116] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0117] In this invention, the server includes a means for pre-learning patent information, a means for automatically generating patent-avoiding electronic component design drawings, a means for developing and manufacturing unique high-performance, low-cost electronic components, and a means for selecting an optimal design based on patent information and adjusting the manufacturing process in real time. This makes it possible to utilize patent information to select an optimal design and adjust the manufacturing process in real time. Furthermore, it is possible to efficiently develop and manufacture high-performance, low-cost electronic components while avoiding patent infringement, thereby optimizing the manufacturing process and improving quality control, thereby increasing the competitiveness of the product.

[0118] "Patent information" refers to data about patents that is made public by public institutions such as the Patent Office.

[0119] "Pre-training" is the process of training an artificial intelligence model with relevant data before it performs a specific task.

[0120] An "electronic component design drawing" is a drawing or diagram that shows the structure and function of an electronic component.

[0121] "Auto-generation" is the process of using artificial intelligence or algorithms to generate data and blueprints without manual intervention.

[0122] "High performance and low cost" refers to high performance and low manufacturing and operating costs.

[0123] "Development and manufacturing" is the process of designing new products and technologies and actually producing them.

[0124] An "optimal design" is the most efficient and effective design under specific conditions and constraints.

[0125] The "manufacturing process" is the series of steps that go from raw materials to creating a final product.

[0126] "Real-time" refers to immediate processing or response without delay.

[0127] "Adjustment" refers to changing or modifying a system or process to suit specific purposes or conditions.

[0128] The following system configuration and processing procedure will be described as an embodiment of the present invention.

[0129] The server includes a means for learning patent information in advance, a means for automatically generating electronic component design drawings that avoid patents, a means for developing and manufacturing unique high-performance, low-cost electronic components, and a means for selecting the optimal design based on patent information and adjusting the manufacturing process in real time.

[0130] System Configuration

[0131] 1. Hardware:

[0132] Server: A server equipped with a high-performance CPU and GPU.

[0133] Camera: High-resolution camera mounted on factory robot.

[0134] Factory robots: Robots used to carry out manufacturing processes.

[0135] 2. Software:

[0136] TensorFlow: Used for training and inference of AI models.

[0137] OpenCV: Used for image processing.

[0138] Database: A database for storing patent information.

[0139] Processing steps

[0140] 1. Preliminary study of patent information:

[0141] The server uses TensorFlow to pre-train the electronic component design drawings retrieved from the patent database. During this process, the AI ​​model extracts and learns the features of the design drawings.

[0142] 2. Automatic generation of electronic component design drawings:

[0143] The server uses a trained AI model to automatically generate patent-evading electronic component designs, which are optimized based on patent information.

[0144] 3. Selecting the best design and adjusting the manufacturing process:

[0145] Images captured by cameras mounted on factory robots are processed using OpenCV and sent to a server, which uses AI models to select the optimal design and adjust the manufacturing process in real time.

[0146] Specific examples

[0147] For example, an AI model can be created that selects the optimal circuit design based on electronic component blueprints retrieved from patent databases. This model can then be installed on factory robots and used to adjust the manufacturing process in real time.

[0148] Example prompt sentence:

[0149] "Create an AI model that selects the optimal circuit design based on electronic component blueprints retrieved from patent databases. This model can be installed on factory robots and adjust the manufacturing process in real time."

[0150] In this way, patent information can be used to select optimal designs and adjust manufacturing processes in real time, enabling efficient development and manufacturing of high-performance, low-cost electronic components while avoiding patents, improving manufacturing process optimization and quality control, and enhancing product competitiveness.

[0151] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0152] Step 1:

[0153] The server retrieves electronic component blueprints from a patent database. As input, it requires a query of the patent database, and as output, it obtains the retrieved blueprint data, which is used in the subsequent pre-training process.

[0154] Step 2:

[0155] The server pre-trains the acquired electronic component blueprints using TensorFlow. Blueprint data is required as input, and a trained AI model is obtained as output. In this process, the features of the blueprints are extracted and trained into the AI ​​model.

[0156] Step 3:

[0157] The server uses a trained AI model to automatically generate patent-avoidance electronic component design drawings. Patent information and the trained model are required as input, and an optimized design drawing is obtained as output. This design drawing is optimized based on the patent information.

[0158] Step 4:

[0159] The factory robots are equipped with cameras that capture images of the manufacturing process. As input, real-time images of the manufacturing process are required, and as output, captured image data is obtained. This image data is used for subsequent processing.

[0160] Step 5:

[0161] The server processes the captured image data using OpenCV and inputs it into the AI ​​model. Image data is required as input, and image features are extracted as output. This feature data is used to select the optimal design.

[0162] Step 6:

[0163] The server uses an AI model to select the optimal design and adjust the manufacturing process in real time. Image feature data and a trained model are required as input, and the output is adjusted manufacturing instructions, which are then sent to factory robots to optimize the manufacturing process.

[0164] Step 7:

[0165] The factory robot adjusts the manufacturing process based on instructions from the server and produces products according to the optimal design. Instructions from the server are required as input, and an optimized product is obtained as output. This process achieves optimization and quality control of the manufacturing process.

[0166] Example 2

[0167] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0168] In conventional semiconductor circuit design, it has been difficult to generate new design drawings while avoiding patent infringement. There is a need for a method that can properly analyze patent information and automatically generate design drawings that avoid patent infringement.

[0169] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0170] In this invention, the server includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for generating new design drawings using a generative AI model, means for confirming whether the generated design drawings avoid patent infringement, and means for regenerating the design drawings if patent infringement is confirmed. This makes it possible to automatically generate new semiconductor circuit design drawings that avoid patent infringement based on patent information.

[0171] "Patent information" is detailed data about existing patents, including claims, invention details, drawings, etc.

[0172] "Pre-training" is the process of training a model with a specific dataset to improve its predictive accuracy for a specific task.

[0173] A "semiconductor circuit design diagram" is a drawing that shows the structure and operation of a semiconductor device, and describes in detail the layout and connections of the circuit.

[0174] "Auto-generation" is the process of using artificial intelligence and algorithms to generate new designs and data without human intervention.

[0175] A "generative AI model" is a type of artificial intelligence that has the ability to generate new data and designs for specific tasks.

[0176] A "prompt" is text that is input to a generative AI model and contains instructions or conditions for the model to perform a specific task.

[0177] "Patent infringement" is the act of using a product or technology without permission within the scope of an existing patent, which can lead to legal issues.

[0178] "Regeneration" refers to running the generation process again when a design or data that has already been generated does not meet certain conditions.

[0179] This invention relates to a system for automatically generating new semiconductor circuit design drawings that avoid patent infringement based on patent information. Specific embodiments of this system will be described below.

[0180] Hardware and software used

[0181] Hardware:

[0182] High-performance servers (e.g., servers equipped with GPUs)

[0183] software:

[0184] Deep learning frameworks (e.g. TensorFlow, PyTorch)

[0185] Generative AI models (e.g., GPT-4 (registered trademark))

[0186] System Operation

[0187] The server first accesses a patent information database to retrieve existing patent information, which contains detailed data on semiconductor circuit designs. Based on the retrieved patent information, the server then generates a prompt to be input into the generative AI model.

[0188] Example prompt sentence:

[0189] "Please generate a new semiconductor circuit design that does not infringe patent number 123456."

[0190] The server inputs this prompt into a generative AI model to generate a new semiconductor circuit design. The generated design is then compared with patent information to check for patent infringement. This check is performed using a patent text analysis algorithm. Specifically, it checks whether each element of the design is included in the scope of the patent claims.

[0191] If a patent infringement is confirmed, the server inputs the prompt sentence into the generative AI model again to generate a new design. This process is repeated until a patent-free design is generated.

[0192] Specific examples

[0193] For example, to generate a design drawing that does not infringe patent number 123456, the server inputs the following prompt statement into the generative AI model:

[0194] Example prompt sentence:

[0195] "Please generate a new semiconductor circuit design that does not infringe patent number 123456."

[0196] The generative AI model generates a new design based on this prompt. The generated design is compared with patent information to check for patent infringement. If an infringement is confirmed, the server inputs the prompt into the generative AI model again to generate a new design. This process is repeated until a design that does not infringe on a patent is generated.

[0197] The above is a specific embodiment for carrying out the present invention. This system makes it possible to automatically generate new semiconductor circuit design diagrams that avoid patent infringement based on patent information.

[0198] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0199] Step 1:

[0200] Obtaining patent information

[0201] The server accesses a patent information database to retrieve existing patent information. Patent numbers and specific keywords are given as input. The server searches the database based on this and retrieves relevant patent information. The output is patent information including claims, invention details, drawings, etc. Specifically, the server executes SQL queries to extract the required information from the database.

[0202] Step 2:

[0203] Generate prompt statement

[0204] Based on the acquired patent information, the server generates a prompt sentence to be input into the generative AI model. Patent information is given as input. The server analyzes the patent information and generates a prompt sentence that includes conditions for avoiding patent infringement. The output is a prompt sentence to be input into the generative AI model. Specifically, the server extracts the key points of the patent information and constructs a prompt sentence based on them.

[0205] Step 3:

[0206] Generating new blueprints

[0207] The server inputs a prompt statement into the generative AI model to generate a new semiconductor circuit design. The prompt statement is given as input. The generative AI model analyzes the prompt statement and generates a new design. The output is a new semiconductor circuit design. Specifically, the generative AI model generates the design using a deep learning algorithm.

[0208] Step 4:

[0209] Confirmation of patent infringement

[0210] The server compares the generated design with patent information to check for patent infringement. A new design and patent information are given as input. The server uses a patent text analysis algorithm to check whether each element of the design is included in the scope of the patent claims. The output is a result of whether there is patent infringement. Specifically, the server compares the design with patent information and evaluates the risk of patent infringement.

[0211] Step 5:

[0212] Regenerate blueprints (if necessary)

[0213] If patent infringement is confirmed, the server inputs the prompt sentence into the generative AI model again to generate a new design. The patent infringement result and the prompt sentence are given as input. The server adjusts the prompt sentence and inputs it into the generative AI model again. The output is a new semiconductor circuit design. Specifically, the server modifies the prompt sentence and runs the generation process again. This process is repeated until a design without patent infringement is generated.

[0214] (Application example 2)

[0215] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0216] In conventional semiconductor circuit design, it was extremely difficult to design in a way that avoided patent infringement, and the manual task of checking patent information required a great deal of effort. Furthermore, there was a lack of technology to automatically generate design drawings that avoided patent infringement, which placed a heavy burden on designers. Furthermore, there were insufficient means to check whether the generated design drawings infringed on existing patents, which increased the risk of patent infringement.

[0217] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for verifying whether the generated design drawings infringe existing patents, and means for exporting the design drawings. This enables the automatic generation of new semiconductor circuit design drawings based on patent information, enabling efficient design to avoid patent infringement. Furthermore, it is possible to automatically verify whether the generated design drawings infringe existing patents, thereby reducing the risk of patent infringement.

[0218] "Patent information" refers to information about published patent documents and patent applications, which indicates the technical content and scope of rights of the patents.

[0219] "Pre-training" refers to the process by which an artificial intelligence or machine learning model is trained in advance using a specific dataset.

[0220] A "semiconductor circuit blueprint" is a drawing or blueprint that shows the structure and operation of a semiconductor device, and plays an important role in the manufacturing process.

[0221] "Auto-generation" refers to the process of using artificial intelligence and algorithms to automatically generate blueprints and data without human intervention.

[0222] "Verification" refers to the process of verifying whether the generated design drawings infringe on existing patents.

[0223] "Export" refers to the process of outputting the generated design drawings in a format such as PDF or CAD file.

[0224] "Artificial intelligence" refers to systems and programs that mimic human intelligence using technologies such as machine learning and deep learning.

[0225] A system for implementing this invention includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for verifying whether the generated design drawings infringe existing patents, and means for exporting the design drawings.

[0226] System configuration

[0227] The system consists of the following main components:

[0228] 1. Server: Pre-learns patent information and performs processing to generate and verify design drawings.

[0229] 2. Terminal: Provides an interface for users to input design requirements and review and export the generated design drawings.

[0230] 3. Database: Storage for patent information and generated design drawings.

[0231] Hardware and software used

[0232] Hardware: High-performance servers, user devices (smartphones, tablets, PCs, etc.)

[0233] software:

[0234] Generative AI model: Uses OpenAI (registered trademark) API to generate blueprints based on patent information.

[0235] Patent Database API: An API for obtaining patent information.

[0236] Design Tool API: API for exporting generated design drawings in PDF and CAD formats.

[0237] Data processing and calculation

[0238] The server uses a patent database API to obtain the latest patent information and pre-trains the generative AI model. When the user inputs design requirements from their device, the server uses the generative AI model to generate a semiconductor circuit design that avoids patents. The generated design is compared with the patent database to check for patent infringement. Once the check is complete, the design is exported in PDF or CAD format using the design tool API.

[0239] Specific examples

[0240] For example, if a user wants to design a "low-power, high-performance semiconductor circuit," the following prompt sentence is used:

[0241] Prompt Sentence Examples

[0242] Patent information: [Patent information details]

[0243] User requirement: low-power, high-performance semiconductor circuit

[0244] Generate patent-free semiconductor circuit designs.

[0245] Based on this prompt, the server uses a generative AI model to generate a new semiconductor circuit design that bypasses the patent, checks for patent infringement, and then exports the design, allowing the user to efficiently obtain a design that bypasses the patent.

[0246] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0247] Step 1:

[0248] The server uses a patent database API to retrieve the latest patent information. The input is patent information from the patent database API, and the output is a dataset of the retrieved patent information. This dataset contains detailed information about patent documents and patent applications.

[0249] Step 2:

[0250] The server pre-trains the generative AI model using the acquired patent information. The input is a dataset of patent information, and the output is a generative AI model that has learned from the patent information. This process involves data processing based on the patent information and model training.

[0251] Step 3:

[0252] A user inputs design requirements using a terminal. The input is the user's design requirements (e.g., "a low-power, high-performance semiconductor circuit"), and the output is the transmission of the design requirements to the server. The terminal provides an interface that transmits the requirements entered by the user to the server.

[0253] Step 4:

[0254] The server receives the user's design requirements and uses a generative AI model to generate a semiconductor circuit design that avoids patents. The input is the user's design requirements and a pre-trained generative AI model, and the output is the generated semiconductor circuit design. In this process, the generative AI model generates a design based on a prompt.

[0255] Step 5:

[0256] The server checks whether the generated design drawings infringe on existing patents. The input is a dataset of the generated design drawings and patent information, and the output is a result of whether or not there is patent infringement. In this process, the generated design drawings are matched with patent information to assess the risk of patent infringement.

[0257] Step 6:

[0258] The server exports the verified design drawings using the design tool API. The input is the verified design drawings, and the output is a PDF or CAD file. This process outputs the design drawings in a format that can be used by the user.

[0259] Step 7:

[0260] The user downloads and checks the exported blueprints using a terminal. The input is a blueprint file in PDF or CAD format, and the output is the blueprint saved on the user's terminal. The terminal provides an interface for the user to check and use the blueprints.

[0261] Example 3

[0262] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0263] Conventional semiconductor manufacturing processes have the problem of high costs for producing high-performance semiconductors. It is also difficult to design new semiconductor circuits while avoiding patent infringement. Furthermore, the process from creating design drawings to manufacturing is complex, requiring efficient operation.

[0264] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0265] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for transmitting the generated semiconductor circuit design drawings to a manufacturing device, means for the manufacturing device to manufacture semiconductors based on the design drawings, and means for developing and manufacturing high-performance, low-cost semiconductors. This makes it possible to efficiently manufacture high-performance, low-cost semiconductors and design new semiconductor circuits while avoiding patent infringement.

[0266] "Patent information" refers to technical information contained in publicly available patent documents and patent databases.

[0267] "Pre-training" refers to the process of training a model with relevant data before it performs a specific task.

[0268] A "semiconductor circuit design diagram" is a drawing that shows the circuit configuration and layout of a semiconductor device.

[0269] "Automated generation" refers to the process of using artificial intelligence and algorithms to generate blueprints and data without human intervention.

[0270] "Manufacturing equipment" refers to machines and facilities used to manufacture semiconductor devices.

[0271] A "communication protocol" refers to the rules and procedures for data communication.

[0272] "Lithography technology" is a technology that uses light to form fine patterns on silicon wafers.

[0273] "High performance and low cost" means high performance and low manufacturing and operating costs.

[0274] "Artificial intelligence" refers to the technology that enables computer systems to learn and reason by imitating human intelligence.

[0275] MODE FOR CARRYING OUT THE INVENTION

[0276] This invention relates to a system that learns patent information in advance, automatically generates semiconductor circuit design drawings that avoid patents, and manufactures high-performance, low-cost semiconductors based on those design drawings. Specific embodiments of this system are described below.

[0277] 1. System Configuration

[0278] This system consists of three main components: a server, a terminal, and a user. The server includes a means for pre-learning patent information, a means for automatically generating semiconductor circuit design drawings that circumvent patents, and a means for transmitting the generated design drawings to manufacturing equipment. The terminal provides an interface for users to input prompt sentences. The user is responsible for operating the system and inputting the necessary prompt sentences.

[0279] 2. Hardware and Software Used

[0280] The server is a computer system equipped with a high-performance processor and large memory capacity. Specifically, it is equipped with a GPU (graphics processing unit) for running the generative AI model. The generative AI model is implemented using machine learning frameworks such as TensorFlow and PyTorch. Publicly available patent databases are used for pre-training patent information. The generated design drawings are sent to the manufacturing equipment using secure communication protocols (e.g., HTTPS or SFTP).

[0281] Manufacturing equipment is used to manufacture semiconductors using lithography technology. Specifically, EUV (extreme ultraviolet) lithography equipment is used. This equipment is used to form fine circuit patterns on silicon wafers.

[0282] 3. Program Processing

[0283] The server inputs the prompt received from the user into a generative AI model to generate an optimal semiconductor circuit design. The generated design is then sent from the server to the manufacturing equipment, which then manufactures semiconductors based on the design.

[0284] 4. Specific Examples

[0285] If a user wants to build a high-performance AI system, they input a prompt sentence like the following into the generative AI model:

[0286] Example prompt sentence:

[0287] "Design a high-performance, low-cost AI processor."

[0288] When this prompt is entered, the server generates an optimal semiconductor circuit design using a generative AI model. The generated design is then sent to the manufacturing equipment using a secure communication protocol, and the manufacturing equipment manufactures semiconductors based on the design. This enables users to build and operate high-performance AI systems without using expensive semiconductors. The flow of the specific processing in Example 3 will be explained using Figure 15.

[0289] Program processing flow

[0290] Step 1:

[0291] The user enters a prompt statement.

[0292] Input: The user accesses the system's interface and enters a prompt statement into a text box.

[0293] Specific operation: The user enters, for example, "Please design a high-performance, low-cost AI processor" and clicks the submit button.

[0294] Output: The prompt text is sent to the server.

[0295] Step 2:

[0296] The server receives the prompt sentence and inputs it into the generative AI model.

[0297] Input: The prompt the server received from the user.

[0298] Specific operation: The server sends the prompt sentence to the generative AI model as an API request.

[0299] Data processing: Convert the prompt sentence into an appropriate format and input it into the generative AI model.

[0300] Output: The generative AI model generates an optimal semiconductor circuit design and sends it back to the server.

[0301] Step 3:

[0302] The server transmits the generated semiconductor circuit design drawing to the manufacturing device.

[0303] Input: Semiconductor circuit design diagrams returned from a generative AI model.

[0304] Specific operation: The server encrypts the generated design drawing and sends it to the manufacturing device using a secure communication protocol (e.g., HTTPS or SFTP).

[0305] Data processing: Encrypt the blueprint data and convert it into a format suitable for the communication protocol.

[0306] Output: Encrypted blueprint sent to fabricator.

[0307] Step 4:

[0308] Manufacturing equipment produces semiconductors based on blueprints.

[0309] Input: Encrypted semiconductor circuit blueprints sent from the server.

[0310] How it works: The manufacturing equipment receives the design drawings, first decrypts them, then analyzes them, prepares silicon wafers, and forms circuits on them using lithography technology.

[0311] Data processing: Analysis of design data and circuit formation using lithography technology.

[0312] Output: Manufactured semiconductor device.

[0313] Step 5:

[0314] Users receive the manufactured semiconductors and build and operate AI systems.

[0315] Input: Manufactured semiconductor devices.

[0316] What happens next: The user receives the manufactured semiconductors and incorporates them into an AI system, for example by attaching the AI ​​processor to a motherboard and assembling the entire system. Then, the user installs the AI ​​software and begins training and inference of machine learning models.

[0317] Data processing: Embedding semiconductor devices and installing AI software.

[0318] Output: A high-performance AI system is built and operational.

[0319] (Application example 3)

[0320] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0321] Conventional semiconductor manufacturing processes require expensive materials and complex processes to produce high-performance semiconductors, resulting in high costs. Furthermore, there was a lack of means to efficiently manufacture semiconductors while avoiding patent infringement. Furthermore, there was also the issue of insufficient optimization of the manufacturing process, resulting in low manufacturing efficiency.

[0322] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0323] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for optimizing the manufacturing process based on the automatically generated semiconductor circuit design drawings, and means for manufacturing semiconductors according to the optimized manufacturing process, thereby enabling the efficient manufacturing of high-performance, low-cost semiconductors.

[0324] "Means for prior learning of patent information" refers to a means for studying the vast number of publicly available semiconductor circuit design drawings and patent documents, and accumulating knowledge to avoid patent infringement.

[0325] The "means for automatically generating semiconductor circuit design drawings that avoid patents" refers to a means for automatically generating new semiconductor circuit design drawings that avoid patent infringement using artificial intelligence.

[0326] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to the development and manufacturing of high-performance semiconductors at low manufacturing costs based on semiconductor circuit design drawings that avoid patents.

[0327] "Means for optimizing the manufacturing process based on automatically generated semiconductor circuit design drawings" refers to means for optimizing each step of the manufacturing process based on automatically generated semiconductor circuit design drawings, and for efficiently manufacturing semiconductors.

[0328] The "means for manufacturing a semiconductor according to an optimized manufacturing process" refers to a means for actually manufacturing a semiconductor according to an optimized manufacturing process.

[0329] A system for implementing this invention includes a means for prior learning of patent information, a means for automatically generating semiconductor circuit design drawings that circumvent patents, a means for developing and manufacturing unique high-performance, low-cost semiconductors, a means for optimizing the manufacturing process based on the automatically generated semiconductor circuit design drawings, and a means for manufacturing semiconductors in accordance with the optimized manufacturing process.

[0330] 1. Program Generation

[0331] The server collects and studies a vast number of publicly available semiconductor circuit designs and patent documents to learn about patent information. It then uses artificial intelligence to automatically generate semiconductor circuit designs that avoid patents. Based on these designs, the server develops and manufactures unique, high-performance, low-cost semiconductors.

[0332] 2. Program processing explanation

[0333] The server uses the following hardware and software:

[0334] Hardware: Factory robots (e.g., general-purpose robotic arms), semiconductor manufacturing equipment (e.g., lithography equipment)

[0335] Software: Python® 3.x, JSON library, artificial intelligence models (e.g., TensorFlow)

[0336] The server first collects patent information and stores it in JSON format. It then uses an artificial intelligence model to generate a semiconductor circuit design that avoids patents. Based on this design, it runs an algorithm to optimize the manufacturing process and determines the optimal manufacturing process. Finally, factory robots manufacture semiconductors according to this optimized process.

[0337] 3. Examples and prompts

[0338] Specific examples

[0339] For example, the patent information collected by the server includes existing semiconductor circuit blueprints and patent documents. Based on this information, the AI ​​model generates a new semiconductor circuit blueprint. Based on this blueprint, each step of the manufacturing process (e.g., layer deposition, photolithography, etching, doping) is optimized.

[0340] Prompt Sentence Examples

[0341] "You are asked to develop an application that will be installed on a factory robot. This application will optimize the semiconductor manufacturing process based on automatically generated semiconductor circuit design drawings, and efficiently manufacture high-performance, low-cost semiconductors."

[0342] Thus, specific embodiments for carrying out the invention have been shown.

[0343] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0344] Step 1:

[0345] The server collects a huge number of publicly available semiconductor circuit design drawings and patent documents. It stores this data in JSON format. The input is publicly available semiconductor circuit design drawings and patent documents, and the output is patent information data stored in JSON format. Specifically, the server collects data using web scraping and APIs and stores it in local storage.

[0346] Step 2:

[0347] The server pre-learns the patent information based on the collected patent information. The input is the patent information data saved in step 1, and the output is the trained artificial intelligence model. Specifically, it uses a machine learning library such as TensorFlow to train the model using patent information as learning data.

[0348] Step 3:

[0349] The server uses a trained AI model to automatically generate semiconductor circuit design drawings that circumvent patents. The input is the trained AI model and design requirements, and the output is an automatically generated semiconductor circuit design drawing. Specifically, the server inputs a prompt statement into the generating AI model to generate the design drawing.

[0350] Step 4:

[0351] The server optimizes the manufacturing process based on automatically generated semiconductor circuit design drawings. The input is the automatically generated semiconductor circuit design drawing, and the output is the optimized manufacturing process steps. Specifically, it uses algorithms to optimize each step of the manufacturing process (e.g., layer deposition, photolithography, etching, doping).

[0352] Step 5:

[0353] The server manufactures semiconductors according to an optimized manufacturing process. The inputs are the steps in the optimized manufacturing process, and the output is the manufactured high-performance, low-cost semiconductors. Specifically, it controls factory robots and operates manufacturing equipment to manufacture semiconductors.

[0354] In this manner, the specific operations performed in each processing step have been described.

[0355] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0356] "Example 1"

[0357] One embodiment of the present invention provides a system that includes a means for pre-learning patent information, a means for automatically generating semiconductor circuit design diagrams that circumvent patents, and a means for developing and manufacturing unique, high-performance, low-cost semiconductors, as well as an emotion engine that recognizes user emotions. This emotion engine recognizes emotions from the user's voice, facial expressions, text input, etc., and controls the operation of the semiconductor based on the recognized emotion. For example, if the user expresses anger, the emotion engine analyzes that information and adjusts the operation of the semiconductor. This enables the semiconductor to perform optimal operations in response to the user's emotions.

[0358] "Example 2"

[0359] Furthermore, as another embodiment of the present invention, a specific technique is provided for an emotion engine to recognize a user's emotion. Specifically, a voice analysis technique that recognizes an emotion from a user's voice, an image analysis technique that recognizes an emotion from a user's facial expression, a natural language processing technique that recognizes an emotion from a user's text input, etc. are used. By combining these techniques, emotion recognition with higher accuracy becomes possible.

[0360] "Example 3"

[0361] Furthermore, as another embodiment of the present invention, a specific technique for controlling the operation of a semiconductor based on the emotion recognized by an emotion engine is provided. Specifically, control such as adjusting the clock frequency of the semiconductor or giving priority to the execution of a specific process is performed according to the type and intensity of the emotion recognized by the emotion engine. This enables optimal control of the semiconductor operation according to the user's emotion.

[0362] The processing flow of each embodiment will be described below.

[0363] "Example 1"

[0364] Step 1: Receive input from the user (voice, facial expression, text, etc.).

[0365] Step 2: The emotion engine recognizes emotions from the user's input.

[0366] Step 3: Control the operation of the semiconductor based on the recognized emotion.

[0367] "Example 2"

[0368] Step 1: Perform voice analysis to recognize emotions from the user's voice.

[0369] Step 2: Perform image analysis to recognize emotions from the user's facial expressions.

[0370] Step 3: Perform natural language processing to recognize emotions from the user's text input.

[0371] "Example 3"

[0372] Step 1: Determine the type and intensity of the emotion recognized by the emotion engine.

[0373] Step 2: Adjust the clock frequency of the semiconductor based on the result of the determination.

[0374] Step 3: Based on the judgment result, a specific process is executed with priority.

[0375] Example 1

[0376] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0377] In conventional semiconductor circuit design, it was difficult to develop high-performance, low-cost semiconductors while avoiding patent infringement. Furthermore, there was no technology to optimize semiconductor operation according to user emotions, making it difficult to improve the user experience.

[0378] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0379] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for recognizing user emotions, and means for controlling the operation of the semiconductor based on the recognized emotions. This makes it possible to develop high-performance, low-cost semiconductors while avoiding patent infringement, and also enables the semiconductors to perform optimal operations according to the user's emotions.

[0380] "Patent information" refers to semiconductor circuit design drawings and related technical documents registered in patent databases.

[0381] "Pre-learning" refers to the process by which artificial intelligence extracts features based on patent information and learns from them.

[0382] A "semiconductor circuit design drawing" refers to a drawing or design data that shows the circuit structure of a semiconductor.

[0383] "Automatic generation" refers to the process by which artificial intelligence automatically creates new semiconductor circuit designs based on patent information.

[0384] "High performance and low cost" refers to high performance and low manufacturing costs.

[0385] "Emotion recognition" is a technology that analyzes and recognizes emotions from a user's voice, facial expressions, text input, etc.

[0386] "Behavioral control" refers to the process of adjusting the behavior of a semiconductor based on perceived emotions.

[0387] This invention is a system that learns patent information in advance, automatically generates semiconductor circuit designs that avoid patents, and controls the operation of semiconductors by recognizing the user's emotions. This system works in cooperation with a server, a terminal, and a user.

[0388] The server accesses a patent database to retrieve semiconductor circuit design drawings. This database contains various semiconductor circuit design drawings that have been published in patents. The server inputs the retrieved design drawings into an AI model using a machine learning framework such as TensorFlow or PyTorch to extract the design features. The AI ​​model learns based on the extracted features and pre-learns patent information.

[0389] The server then uses the trained AI model to generate new semiconductor circuit design drawings. In this process, the AI ​​takes existing patent information into account when generating new designs. The resulting designs reduce the risk of patent infringement while enabling high-performance, low-cost semiconductors.

[0390] The device inputs the user's voice, facial expressions, text input, etc. into the emotion engine. This emotion engine is built using OpenCV and Microsoft (registered trademark) Azure (registered trademark) emotion recognition APIs. The emotion engine recognizes the user's emotions from the input data and analyzes that information.

[0391] The server controls the semiconductor's operation based on the user's emotional information obtained from the emotion engine. For example, if the user expresses anger, the server analyzes that information and adjusts the semiconductor's operation. This enables the semiconductor to perform optimal operations according to the user's emotions.

[0392] As a specific example, the server retrieves the semiconductor circuit design drawing for "Patent No. 123456" from a patent database and inputs it into an AI model for learning. Next, it uses the trained AI model to generate a new design drawing and confirms that it is designed to avoid patent No. 123456. The device inputs the user's voice input, "I'm very tired today," into the emotion engine, which recognizes that the user is feeling "tired." Based on this emotion information, the server switches the semiconductor's operation to energy-saving mode.

[0393] Examples of prompts include, "Build an AI model that learns from semiconductor circuit designs retrieved from a patent database and generates new designs that avoid patents. Also, describe a system that recognizes user emotions and controls the operation of semiconductors based on those emotions."

[0394] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0395] Step 1:

[0396] The server accesses the patent database and retrieves the semiconductor circuit design. Specifically, the server uses an API to download the design data for "Patent Number 123456." The input is the patent number, and the output is the design data. This design data is used to train the subsequent AI model.

[0397] Step 2:

[0398] The server inputs the acquired blueprint data into an AI model using machine learning frameworks such as TensorFlow and PyTorch. The AI ​​model extracts the features of the blueprint and performs learning. The input is the blueprint data, and the output is the trained model. In this process, image processing technology is used to analyze the blueprint and extract features.

[0399] Step 3:

[0400] The server uses the trained AI model to generate new semiconductor circuit designs. Specifically, the AI ​​model combines features from existing designs to generate new designs. The input is the trained model, and the output is the new design. This design is designed to avoid patent infringement.

[0401] Step 4:

[0402] The server checks the generated design against a patent database to determine whether it avoids patent infringement. The input is the new design, and the output is information about whether or not there is patent infringement. This process uses natural language processing technology to analyze the patent documents.

[0403] Step 5:

[0404] The device inputs the user's voice, facial expressions, and text input into the emotion engine. Specifically, it collects voice and facial expression data using a microphone and camera. The input is the user's voice and facial expression data, and the output is emotional information. The emotion engine is built using OpenCV and Microsoft Azure's emotion recognition API.

[0405] Step 6:

[0406] The emotion engine recognizes the user's emotions from input data and analyzes that information. The input is voice and facial expression data, and the output is the recognized emotional information. For example, if a user inputs "I'm very tired today," the emotion engine will recognize that the user is feeling "tired."

[0407] Step 7:

[0408] The server controls the operation of the semiconductor based on the emotional information obtained from the emotion engine. The input is emotional information, and the output is the adjusted operation of the semiconductor. Specifically, if the user is tired, the server switches the operation of the semiconductor to energy-saving mode. For example, it makes adjustments such as lowering the CPU clock speed.

[0409] (Application example 1)

[0410] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0411] Conventional semiconductor design systems have difficulty learning patent information in advance and automatically generating designs that avoid patent infringement. Furthermore, they lack the ability to recognize user emotions and optimize semiconductor operation based on those emotions, resulting in reduced work efficiency and user satisfaction. This has created a need for efficient development and manufacturing of high-performance, low-cost semiconductors while avoiding the risk of patent infringement.

[0412] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0413] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means including an emotion engine that recognizes user emotions, and means for controlling the operation of the semiconductor based on the user emotions. This makes it possible to efficiently develop and manufacture high-performance, low-cost semiconductors that perform optimal operations in response to user emotions while avoiding the risk of patent infringement.

[0414] "Means for pre-learning patent information" refers to a means for obtaining semiconductor circuit design drawings from publicly available patent databases and learning their features using artificial intelligence.

[0415] The "means for automatically generating semiconductor circuit design drawings that avoid patents" refers to a means for automatically generating new semiconductor circuit design drawings that avoid the risk of patent infringement using artificial intelligence based on learned patent information.

[0416] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to means for efficiently developing and manufacturing high-performance, low-cost semiconductors based on design drawings that avoid patents.

[0417] The "means including an emotion engine for recognizing a user's emotion" refers to a means including an engine for analyzing and recognizing an emotion from the user's voice, facial expression, text input, and the like.

[0418] The "means for controlling the operation of a semiconductor based on the user's emotions" refers to a means for generating a control signal for optimizing the operation of a semiconductor based on the recognized user's emotions, and adjusting the operation of the semiconductor.

[0419] A system for implementing this invention includes a means for pre-learning patent information, a means for automatically generating semiconductor circuit design drawings that circumvent patents, a means for developing and manufacturing unique high-performance, low-cost semiconductors, a means including an emotion engine that recognizes user emotions, and a means for controlling the operation of the semiconductor based on the user emotions.

[0420] Program processing explanation

[0421] 1. Hardware:

[0422] Robot body (e.g. industrial robot)

[0423] Emotion recognition camera (e.g. depth camera)

[0424] Voice recognition microphone (e.g. high-sensitivity microphone)

[0425] 2. Software:

[0426] Patent information learning AI (e.g., custom models using TensorFlow)

[0427] Emotion recognition engine (e.g., emotion analysis API)

[0428] Semiconductor design automatic generation algorithm (e.g., implemented in Python)

[0429] 3. Data processing and calculation:

[0430] The server retrieves semiconductor circuit design drawings from a patent database and inputs them into an AI model to extract features.

[0431] Data obtained from the emotion recognition camera and voice recognition microphone is input into the emotion recognition engine to analyze the user's emotions.

[0432] Based on the emotion data, a control signal is generated to optimize the operation of the semiconductor.

[0433] Specific examples

[0434] For example, if a worker in a factory feels stressed, the emotion recognition engine can analyze that emotion and instruct the robot to adjust its working speed, thereby reducing worker stress and providing an efficient working environment.

[0435] Prompt Sentence Examples

[0436] An example of a prompt to be input to the generative AI model is as follows:

[0437] "Learn from semiconductor circuit designs obtained from patent databases and generate new designs that avoid patents. Also, recognize user emotions and generate control signals that optimize the operation of semiconductors based on those emotions."

[0438] In this way, the system will be able to utilize patent information to efficiently develop and manufacture high-performance, low-cost semiconductors that perform optimal operations in response to user emotions.

[0439] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0440] Step 1:

[0441] The server retrieves semiconductor circuit design drawings from a patent database. The input is the patent database, and the output is the retrieved semiconductor circuit design drawings. These design drawings are then input into the patent information learning AI.

[0442] Step 2:

[0443] The server extracts features from the acquired semiconductor circuit design using patent information learning AI. The input is the acquired semiconductor circuit design, and the output is the extracted feature data. This feature data is used to generate designs to circumvent patents.

[0444] Step 3:

[0445] The server automatically generates patent-evading semiconductor circuit design drawings based on the extracted feature data. The input is the feature data, and the output is a new semiconductor circuit design drawing. This design drawing is used to develop and manufacture unique, high-performance, low-cost semiconductors.

[0446] Step 4:

[0447] The device acquires the user's emotional data using an emotion recognition camera and a voice recognition microphone. The input is the user's voice and facial expressions, and the output is emotional data. This emotional data is input into the emotion recognition engine.

[0448] Step 5:

[0449] The server analyzes the acquired emotion data using an emotion recognition engine. The input is the emotion data, and the output is the analyzed emotion information. This emotion information is used to control the operation of the semiconductor.

[0450] Step 6:

[0451] The server generates a control signal for optimizing the operation of the semiconductor based on the analyzed emotion information. The input is the analyzed emotion information, and the output is the control signal. The control signal is used to adjust the operation of the semiconductor.

[0452] Step 7:

[0453] The terminal uses the generated control signal to adjust the operation of the semiconductor. The input is the control signal, and the output is the adjusted operation of the semiconductor. This allows optimal operation according to the user's emotions to be achieved.

[0454] Example 2

[0455] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0456] In conventional semiconductor circuit design, avoiding patent infringement is extremely difficult, requiring a great deal of time and cost. Furthermore, there is a lack of technology to accurately recognize user emotions, making it difficult to improve the user experience. To solve these problems, we need technology that effectively utilizes patent information to automatically generate designs that avoid patent infringement, and technology that can accurately recognize user emotions.

[0457] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0458] In this invention, the server includes a means for prior learning of patent information, a means for automatically generating semiconductor circuit design drawings that avoid patents, and a means for verifying whether the generated design drawings infringe on existing patents, thereby enabling new semiconductor circuit design drawings to be efficiently generated while avoiding patent infringement.

[0459] Furthermore, the terminal includes a means for recognizing user emotions, a means for acquiring and analyzing voice data, a means for acquiring and analyzing facial expression data, a means for acquiring and analyzing text data, and a means for integrating emotion recognition results, thereby enabling highly accurate recognition of user emotions and improving the user experience.

[0460] "Patent information" refers to technical information contained in publicly available patent documents and patent databases.

[0461] "Pre-training" is the process of training a model with a specific dataset to improve its performance on a specific task.

[0462] A "semiconductor circuit design diagram" is a drawing or diagram that shows the structure and function of a semiconductor device.

[0463] "Auto-generation" refers to the process of using artificial intelligence and algorithms to create new designs and data without human intervention.

[0464] "Verification" refers to the process of verifying whether the generated design drawings infringe on existing patents.

[0465] "User emotion" refers to the psychological state that the user feels, such as joy, sadness, anger, or surprise.

[0466] "Voice data" refers to data that is a digital recording of a user's speech or voice.

[0467] "Analysis" is the process of examining data in detail to understand its meaning and structure.

[0468] "Facial expression data" refers to image data that captures the facial expression of a user.

[0469] "Text data" refers to character information entered by the user.

[0470] "Emotion recognition results" refer to the analysis results of emotions obtained from voice, facial expressions, and text.

[0471] "Integration" refers to the process of combining multiple data and analysis results into one.

[0472] This invention is a system that pre-learns patent information, automatically generates semiconductor circuit designs that avoid patents, and recognizes user emotions with high accuracy. Specific embodiments of this system are described below.

[0473] Automatic generation of semiconductor circuit design diagrams

[0474] To pre-train the patent information, the server retrieves relevant patent information from a patent information database. The patent information database contains a huge number of publicly available semiconductor circuit design diagrams. The server analyzes this patent information using natural language processing technology (e.g., the BERT model) and extracts important technical elements.

[0475] Next, the server uses a generative AI model (e.g., TensorFlow, PyTorch) to automatically generate new semiconductor circuit design diagrams. The generative AI model generates design patterns that avoid patents based on patent information it has learned in advance. The generated design diagrams are compared with a patent information database to ensure they do not infringe on existing patents.

[0476] Examples:

[0477] Hardware used: High-performance server (e.g., server with GPU)

[0478] Software used: TensorFlow, PyTorch

[0479] Example prompt: "Generate a new semiconductor circuit design based on the information in patent number 123456."

[0480] Recognizing user emotions with an emotion engine

[0481] To recognize the user's emotions, the device uses a combination of voice analysis technology, image analysis technology, and natural language processing technology. Specifically, the device captures the user's voice data with a microphone and analyzes emotions using voice analysis software (e.g., Google Cloud Speech-to-Text). It also uses a camera to capture the user's facial expressions and recognizes emotions using image analysis software (e.g., OpenCV). It also analyzes text data entered by the user with natural language processing software (e.g., the BERT model) to recognize emotions.

[0482] These analysis results are integrated by an integrated algorithm within the device to determine the final emotion, enabling highly accurate recognition of the user's emotions and appropriate responses.

[0483] Examples:

[0484] Hardware used: smartphone, tablet, PC

[0485] Software used: Google Cloud Speech-to-Text, OpenCV, BERT model

[0486] Example prompt: "Recognize emotions from the user's voice, facial expressions, and text input."

[0487] This system makes it possible to efficiently generate new semiconductor circuit designs while avoiding patent infringement, and also improves the user experience by recognizing user emotions with high accuracy.

[0488] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0489] Automatic generation of semiconductor circuit design diagrams

[0490] Step 1: Obtain patent information

[0491] The server retrieves relevant patent information from a patent information database. As input, it sends a query containing patent numbers and keywords to the database. As output, it returns the relevant patent information.

[0492] What happens: The server sends the query "Get information about patent number 123456" to the database.

[0493] Step 2: Analyzing patent information

[0494] The server analyzes the acquired patent information using natural language processing technology. The patent document is given as input, and important technical elements are extracted as output.

[0495] Specific operation: The server uses the BERT model and instructs the server to "extract the technical elements of patent number 123456."

[0496] Step 3: Generate a new blueprint

[0497] The server uses a generative AI model based on the analysis results to generate a new semiconductor circuit design. The extracted technical elements are given as input, and the new design is generated as output.

[0498] Specific operation: The server inputs a prompt to the generative AI model: "Generate a new design that avoids patent number 123456."

[0499] Step 4: Check the design for patent infringement

[0500] The server checks the generated design against a patent information database to determine whether the design infringes any existing patents. The generated design is given as input, and the server returns whether or not there is any patent infringement as output.

[0501] Specific operation: The server requests a database check to see if the generated design drawing infringes patent number 123456.

[0502] Recognizing user emotions with an emotion engine

[0503] Step 1: Acquire and analyze audio data

[0504] The device captures the user's voice data with a microphone and analyzes their emotions using voice analysis software. The voice data is given as input, and the emotion analysis results are obtained as output.

[0505] Specific operation: The device instructs the voice analysis software to "capture the user's voice and analyze their emotions."

[0506] Step 2: Acquire and analyze facial expression data

[0507] The device uses a camera to capture the user's facial expressions and image analysis software to recognize emotions. Facial expression data is given as input, and emotion analysis results are obtained as output.

[0508] Specific operation: The device instructs the image analysis software to "capture the user's facial expression and analyze their emotions."

[0509] Step 3: Acquire and analyze text data

[0510] The device analyzes the text data entered by the user using natural language processing software to recognize emotions. The text data is given as input, and the emotion analysis results are obtained as output.

[0511] What it does: The device instructs the natural language processing software to "analyze the user's text input and recognize emotions."

[0512] Step 4: Integrating emotion recognition results

[0513] The device integrates the emotion recognition results obtained from voice, facial expressions, and text to determine the final emotion. Each analysis result is given as input, and the integrated emotion recognition result is obtained as output.

[0514] Specific operation: The device instructs the integration algorithm to "integrate the emotion recognition results from voice, facial expression, and text to determine the final emotion."

[0515] (Application example 2)

[0516] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0517] In conventional semiconductor circuit design, it was difficult to generate new design drawings while avoiding patent infringement. Furthermore, to improve work efficiency in factories, it was necessary to recognize the emotions of workers and provide appropriate feedback, but there was a lack of effective means to achieve this.

[0518] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0519] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for recognizing worker emotions, means for providing feedback to optimize work efficiency, and means for developing and manufacturing unique high-performance, low-cost semiconductors. This makes it possible to generate new semiconductor circuit design drawings while avoiding patent infringement, and further to optimize work efficiency by recognizing worker emotions.

[0520] "Patent information" refers to information about patents that has been made public by public institutions such as the Patent Office.

[0521] "Pre-training" is the process of pre-training an artificial intelligence model using a specific dataset.

[0522] A "semiconductor circuit design diagram" is a drawing that shows the structure and operation of a semiconductor device.

[0523] "Auto-generation" is the process of using artificial intelligence and algorithms to generate new designs and data without human intervention.

[0524] "Workers" are people who work in factories or manufacturing sites.

[0525] "Emotion recognition" is a technology that identifies a person's emotional state by analyzing data such as voice, facial expressions, and text.

[0526] "Feedback" is information or instructions that a system or device provides to a user.

[0527] "High performance and low cost" means high performance and low manufacturing and operating costs.

[0528] "Development and manufacturing" refers to the process of designing new products and technologies and actually producing them.

[0529] A "system" is an integrated structure in which multiple elements and functions are combined to achieve a specific purpose.

[0530] A system for implementing this invention includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for recognizing worker emotions, means for providing feedback to optimize work efficiency, and means for developing and manufacturing unique high-performance, low-cost semiconductors.

[0531] Hardware and software used

[0532] Hardware: Factory robots, cameras, microphones, displays

[0533] Software: TensorFlow (AI model learning and inference), OpenCV (image analysis), NLTK (natural language processing), PyAudio (audio analysis)

[0534] Data processing and calculation

[0535] Collecting and learning patent information

[0536] The server collects patent information from patent databases and trains the AI ​​model with the following prompts:

[0537] "Generate new semiconductor circuit design diagrams based on patent information."

[0538] Semiconductor circuit design generation

[0539] The server uses the trained AI model to generate new design drawings based on patent information, and then checks whether the generated design drawings infringe on patents.

[0540] emotion recognition

[0541] The device uses a camera and microphone to collect facial expressions and voice data from the worker. It uses OpenCV to analyze facial expressions and NLTK to analyze voice and text input. An emotion engine integrates these data and recognizes the worker's emotions.

[0542] Providing Feedback

[0543] The server then displays feedback on the display to optimize work efficiency based on the recognized emotions. For example, if the server detects fatigue in the worker's facial expression, it will display a message urging the worker to take a break.

[0544] In this way, the system can generate new semiconductor circuit designs while avoiding patent infringement, and can also recognize the emotions of workers to optimize work efficiency.

[0545] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0546] Step 1:

[0547] The server collects patent information from patent databases. It uses the URL and API key of the patent database as input and obtains a dataset of patent information as output. Specifically, the server accesses the patent database and extracts patent information based on specific conditions.

[0548] Step 2:

[0549] The server pre-trains an AI model based on the collected patent information. It uses a patent information dataset as input and obtains a trained AI model as output. Specifically, the server uses TensorFlow to supply the patent information to the AI ​​model and optimizes the model's parameters.

[0550] Step 3:

[0551] The server generates a new semiconductor circuit design using the trained AI model. It uses the prompt "Please generate a new semiconductor circuit design based on patent information" as input and obtains the new semiconductor circuit design as output. Specifically, the server inputs the prompt into the generating AI model and obtains the generated design.

[0552] Step 4:

[0553] The server checks whether the generated semiconductor circuit design drawing does not infringe any patents. It uses the generated design drawing as input and obtains a result indicating whether or not there is patent infringement as output. Specifically, the server compares patent information with the generated design drawing to assess the risk of patent infringement.

[0554] Step 5:

[0555] The terminal uses a camera and microphone to collect the worker's facial expressions and voice. It uses real-time data from the camera and microphone as input and obtains the collected facial and voice data as output. Specifically, the terminal activates the camera and microphone and records the worker's facial expressions and voice.

[0556] Step 6:

[0557] The device analyzes the collected facial and voice data to recognize the worker's emotions. It uses facial and voice data as input and obtains the recognized emotion as output. Specifically, the device analyzes facial expressions using OpenCV and voice data using NLTK.

[0558] Step 7:

[0559] The server provides feedback to optimize work efficiency based on the recognized emotions. The server uses the recognized emotion data as input and obtains a feedback message as output. Specifically, the server analyzes the emotion data, generates an appropriate feedback message, and displays it on a display.

[0560] Step 8:

[0561] The user checks the feedback message displayed on the display and adjusts their behavior as necessary. The feedback message is used as input and the output is improved work efficiency. Specifically, the user checks the display and performs an action such as taking a break.

[0562] Example 3

[0563] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0564] Conventional semiconductor manufacturing systems have had the difficulty of producing high-performance semiconductors at low cost, and have also had the problem of being unable to control operation in response to user emotions. Furthermore, semiconductor circuits designed without taking patent information into consideration have a high risk of patent infringement, making commercialization difficult.

[0565] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0566] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for controlling the operation of the semiconductor based on emotions recognized by the emotion engine, and means for transmitting the generated semiconductor circuit design drawings to semiconductor manufacturing equipment. This makes it possible to manufacture high-performance, low-cost semiconductors and also realizes optimal operation control according to the user's emotions.

[0567] "Means for prior study of patent information" refers to a means for studying the vast number of publicly available semiconductor circuit design drawings and related patent documents, and acquiring knowledge to avoid patent infringement.

[0568] The "means for automatically generating semiconductor circuit design drawings that circumvent patents" refers to a means for automatically generating new semiconductor circuit design drawings that pose a low risk of patent infringement, taking patent information into consideration, using artificial intelligence.

[0569] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to the development and manufacturing of high-performance semiconductors at low manufacturing costs, based on automatically generated semiconductor circuit design drawings and through processes such as lithography and etching.

[0570] "Means for controlling the operation of a semiconductor based on emotions recognized by an emotion engine" refers to means for the emotion engine to recognize the user's emotions and adjust the clock frequency and processing priority of the semiconductor based on that emotion data.

[0571] The "means for transmitting the generated semiconductor circuit design drawing to the semiconductor manufacturing device" refers to a means for converting the semiconductor circuit design drawing automatically generated by the server into an appropriate format and transmitting it to the semiconductor manufacturing device.

[0572] This invention is a system that pre-learns patent information, automatically generates semiconductor circuit designs that avoid patents, and develops and manufactures unique high-performance, low-cost semiconductors. Furthermore, it has the function of controlling the operation of semiconductors based on emotions recognized by an emotion engine.

[0573] System configuration

[0574] The system consists of the following main components:

[0575] 1. Server: Pre-learns patent information, automatically generates semiconductor circuit design diagrams, and processes data from the emotion engine.

[0576] 2. Terminal: Provides an interface for users to input emotions.

[0577] 3. Semiconductor manufacturing equipment: Manufactures semiconductors based on the design drawings sent from the server.

[0578] Hardware and software used

[0579] Generative AI models: Use deep learning frameworks such as TensorFlow and PyTorch.

[0580] Emotion engine: Uses IBM Watson's Natural Language Understanding API.

[0581] Semiconductor manufacturing equipment: Uses ASML's EUV lithography equipment, etc.

[0582] Program processing

[0583] Preliminary study of patent information

[0584] The server studies a vast number of publicly available semiconductor circuit designs and related patent documents, thereby acquiring knowledge to avoid patent infringement.

[0585] Automatic generation of semiconductor circuit design diagrams

[0586] The server uses a generative AI model to automatically generate semiconductor circuit design drawings that are novel and pose a low risk of patent infringement, taking patent information into account.

[0587] Semiconductor manufacturing

[0588] The server converts the generated semiconductor circuit design drawings into an appropriate format and sends them to semiconductor manufacturing equipment, which then uses processes such as lithography and etching to produce high-performance semiconductors at low manufacturing costs.

[0589] Emotion-based motion control

[0590] The user operates the device to input emotions into the emotion engine, which analyzes the data and recognizes the type and intensity of the emotion. The server receives the data from the emotion engine and executes the semiconductor's operation control algorithm, which adjusts the clock frequency and prioritizes the execution of specific processes.

[0591] Specific examples

[0592] Example 1: When a user is feeling stressed, the emotion engine recognizes the "stress" and the server reduces the clock frequency of the semiconductor to reduce the load on the system.

[0593] Example 2: When a user is concentrating, the emotion engine recognizes this and the server prioritizes specific processes, thereby improving the user's work efficiency.

[0594] Prompt Sentence Examples

[0595] "Please explain how the system adjusts its behavior if the user is stressed."

[0596] "Please explain in detail how the emotion engine controls the semiconductor's behavior based on the emotions it recognizes."

[0597] In this way, the system manufactures high-performance, low-cost semiconductors and realizes optimal operation control according to the user's emotions. The flow of the specific processing in the third embodiment will be described with reference to FIG.

[0598] Step 1:

[0599] The server pre-learns patent information.

[0600] Input: A huge number of publicly available semiconductor circuit designs and related patent documents.

[0601] Data processing: The server uses a deep learning framework (e.g., TensorFlow, PyTorch) to learn from this data and acquire knowledge to avoid patent infringement.

[0602] Output: A knowledge base that takes into account patent information.

[0603] What it does: The server loads the dataset, trains the neural network, and generates a model for patent circumvention.

[0604] Step 2:

[0605] The server automatically generates semiconductor circuit design drawings that circumvent patents.

[0606] Input: Knowledge base considering patent information.

[0607] Data processing: The server uses a generative AI model to automatically generate new semiconductor circuit design drawings that have a low risk of patent infringement.

[0608] Output: Patent-evading semiconductor circuit design.

[0609] Specific operation: The server references the knowledge base and uses a generative AI model to output the optimal circuit design.

[0610] Step 3:

[0611] The server transmits the generated semiconductor circuit design drawing to the semiconductor manufacturing equipment.

[0612] Input: Patent-circuit-evading semiconductor circuit design.

[0613] Data processing: The server converts the design drawing file into an appropriate format and sends it to the control system of the manufacturing equipment.

[0614] Output: Blueprint sent to semiconductor manufacturing equipment.

[0615] Specific operation: The server converts the design drawing into a different format and sends it to the manufacturing device via the network.

[0616] Step 4:

[0617] Semiconductor manufacturing equipment manufactures semiconductors based on design drawings.

[0618] Input: Semiconductor circuit design diagram sent from the server.

[0619] Data processing: Manufacturing equipment produces semiconductors through processes such as lithography, etching, and doping.

[0620] Output: Finished semiconductor.

[0621] Specific operation: The manufacturing equipment reads the blueprint and executes each process in sequence to manufacture semiconductors.

[0622] Step 5:

[0623] The user operates the terminal to input emotions into the emotion engine.

[0624] Input: User emotion data (text and / or voice).

[0625] Data processing: The emotion engine analyzes the data to recognize the type and intensity of emotions.

[0626] Output: Recognized emotion data.

[0627] Specific operation: The user inputs emotions into the device, and the emotion engine analyzes the data.

[0628] Step 6:

[0629] The server receives data from the emotion engine and executes the semiconductor's motion control algorithms.

[0630] Input: Recognized emotion data from the emotion engine.

[0631] Data processing: The server analyzes the emotion data, adjusts the clock frequency, and prioritizes the execution of specific processes.

[0632] Output: Optimized semiconductor operating instructions.

[0633] Specific operation: The server executes the operation control algorithm based on the emotion data and adjusts the operation of the semiconductor.

[0634] In this way, the system will produce high-performance, low-cost semiconductors and achieve optimal operation control according to the user's emotions.

[0635] (Application example 3)

[0636] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0637] With conventional semiconductor manufacturing technology, it is difficult to manufacture high-performance semiconductors at low cost, and there is a high risk of patent infringement. Furthermore, because robots' movements are not optimized according to the emotions of factory workers, work efficiency and worker stress reduction are not fully achieved. To solve these issues, a system is needed that can manufacture high-performance, low-cost semiconductors while avoiding patent infringement, and that can optimize robot movements based on emotion recognition data.

[0638] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0639] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for acquiring emotion recognition data, means for controlling the operation of the semiconductor based on the emotion recognition data, and means for optimizing the operation of the robot in accordance with the emotions of workers in the factory. This enables optimal robot operation in accordance with the emotions of workers while avoiding the risk of patent infringement using high-performance, low-cost semiconductors.

[0640] "Means for prior learning of patent information" refers to a means for studying the vast number of publicly available semiconductor circuit design drawings and patent documents to acquire knowledge for avoiding patent infringement.

[0641] "Means for automatically generating semiconductor circuit design drawings to avoid patent infringement" refers to means for automatically generating semiconductor circuit design drawings to avoid patent infringement using artificial intelligence.

[0642] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to means for developing and manufacturing semiconductors that are high-performance yet can be manufactured at low cost based on automatically generated design drawings.

[0643] "Means for acquiring emotion recognition data" refers to a method for recognizing emotions from the facial expressions and voices of workers using sensors such as cameras and microphones, and acquiring that data.

[0644] The "means for controlling the operation of a semiconductor based on emotion recognition data" refers to a means for adjusting the clock frequency and processing priority of a semiconductor based on the acquired emotion recognition data.

[0645] The "means for optimizing the operation of a robot in accordance with the emotions of workers in a factory" refers to a means for optimizing the operation of a robot in a factory in accordance with the emotions of workers based on emotion recognition data.

[0646] A system for implementing this invention includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for acquiring emotion recognition data, means for controlling the operation of semiconductors based on the emotion recognition data, and means for optimizing the operation of robots according to the emotions of workers in a factory.

[0647] Program processing explanation

[0648] 1. Methods for prior learning of patent information:

[0649] The server collects a huge number of publicly available semiconductor circuit designs and patent documents, and uses machine learning algorithms (e.g., TensorFlow) to learn from these data, thereby acquiring knowledge to avoid patent infringement.

[0650] 2. A method for automatically generating semiconductor circuit designs that circumvent patents:

[0651] Based on the learned patent information, the server uses a generative AI model to automatically generate semiconductor circuit designs that avoid patent infringement. This generative AI model receives prompt text as input and outputs designs that avoid patent infringement.

[0652] 3. A means to develop and manufacture proprietary high-performance, low-cost semiconductors:

[0653] The server inputs the automatically generated design drawings into the semiconductor manufacturing equipment, which then produces high-performance, low-cost semiconductors based on the design drawings.

[0654] 4. How to obtain emotion recognition data:

[0655] Terminals in the factory (including sensors such as cameras and microphones) capture workers' facial expressions and voices in real time and generate emotional data using emotion recognition algorithms (e.g., OpenCV).

[0656] 5. Means for controlling semiconductor operation based on emotion recognition data:

[0657] Based on the acquired emotion recognition data, the server adjusts the clock frequency and processing priority of the semiconductor equipped with the emotion engine.

[0658] 6. How to optimize robot behavior based on the emotions of factory workers:

[0659] The server then uses emotion recognition data to optimize the behavior of robots in the factory. For example, if a worker is recognized as tired, the robot will adjust its behavior to automatically take over the task.

[0660] Specific examples

[0661] For example, if the emotion engine recognizes that a worker in a factory is tired, the robot will adjust its behavior to automatically take on more work to reduce the worker's burden. Also, if a worker is feeling stressed, the robot will temporarily slow down its behavior to allow the worker to relax.

[0662] Prompt Sentence Examples

[0663] Examples of prompts to be input to a generative AI model include:

[0664] "If the emotion engine recognizes that a worker in a factory is tired, how does the robot adjust its behavior?"

[0665] Using this prompt, the generative AI model can suggest specific ways to adjust behavior.

[0666] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0667] Step 1:

[0668] The server collects a huge number of publicly available semiconductor circuit designs and patent documents, and uses machine learning algorithms (e.g., TensorFlow) to learn from these data. The input is publicly available semiconductor circuit designs and patent documents, and the output is knowledge for avoiding patent infringement. This knowledge is used to generate designs for avoiding patent infringement.

[0669] Step 2:

[0670] Based on the learned patent information, the server uses a generative AI model to automatically generate semiconductor circuit design drawings that circumvent patents. The input is knowledge for patent circumvention and a prompt, and the output is a semiconductor circuit design drawing that circumvents patents. Specifically, the prompt is input, and the generative AI model outputs the design drawing.

[0671] Step 3:

[0672] The server inputs the automatically generated blueprints into semiconductor manufacturing equipment, which then manufactures high-performance, low-cost semiconductors based on the blueprints. The input is the automatically generated blueprints, and the output is the manufactured semiconductors. Specifically, the server sends the blueprints to the manufacturing equipment and starts the manufacturing process.

[0673] Step 4:

[0674] The terminal (including sensors such as cameras and microphones) captures the facial expressions and voices of factory workers in real time and generates emotion data using an emotion recognition algorithm (e.g., OpenCV). The input is the facial expressions and voices of the workers, and the output is emotion data. Specifically, the sensors collect data and the emotion recognition algorithm analyzes the emotions.

[0675] Step 5:

[0676] The server adjusts the clock frequency and processing priority of the semiconductor equipped with the emotion engine based on the acquired emotion recognition data. The input is emotion recognition data, and the output is the adjusted semiconductor operating parameters. Specifically, it analyzes the emotion data and dynamically changes the semiconductor operating parameters.

[0677] Step 6:

[0678] The server optimizes the robots' behavior in the factory based on the emotion recognition data. The input is the emotion recognition data, and the output is the optimized robot's behavior instructions. Specifically, the emotion data is analyzed and the robot's behavior is adjusted. For example, if a worker is recognized as tired, the robot adjusts its behavior so that it automatically takes over the work.

[0679] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0680] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0681] Another example of generative AI is Gemini (registered trademark) (Internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0682] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0683] [Second embodiment]

[0684] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0685] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0686] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0687] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0688] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0689] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0690] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0691] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0692] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0693] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0694] In the smart glasses 214, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0695] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0696] "Example 1"

[0697] In one embodiment of the present invention, a means for pre-learning patent information is to use an AI that learns from semiconductor circuit design drawings obtained from a patent database. This AI extracts and learns the features of semiconductor circuit design drawings published in patents.

[0698] "Example 2"

[0699] Furthermore, as a means of automatically generating semiconductor circuit designs that avoid patents, the system has the ability to automatically generate new semiconductor circuit designs that do not infringe patents based on the patent information it has learned. Specifically, the AI ​​generates new designs based on patent information and checks whether the designs infringe existing patents.

[0700] "Example 3"

[0701] In addition, as a means of developing and manufacturing unique high-performance, low-cost semiconductors, semiconductors are manufactured based on automatically generated semiconductor circuit design drawings. Specifically, the automatically generated design drawings are input into semiconductor manufacturing equipment, which then manufactures semiconductors based on the design drawings. This makes it possible to build and operate high-performance AI without using expensive semiconductors.

[0702] The processing flow of each embodiment will be described below.

[0703] "Example 1"

[0704] Step 1: Obtain patent information for semiconductor circuit design drawings from a patent database.

[0705] Step 2: The acquired patent information is input into the AI, which then learns from the patent information. Through this learning, the AI ​​extracts the characteristics of the semiconductor circuit design.

[0706] "Example 2"

[0707] Step 1: Based on the learned patent information, the artificial intelligence automatically generates new semiconductor circuit design drawings.

[0708] Step 2: Check whether the automatically generated semiconductor circuit design diagram infringes on any existing patents. If it is confirmed that there is no infringement, the design diagram is adopted.

[0709] "Example 3"

[0710] Step 1: The selected semiconductor circuit design is input into the semiconductor manufacturing equipment.

[0711] Step 2: Semiconductor manufacturing equipment produces semiconductors based on the blueprint. This manufacturing process makes it possible to build and operate high-performance AI without using expensive semiconductors.

[0712] Example 1

[0713] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0714] In conventional semiconductor circuit design, searching and analyzing patent information is often done manually, which is time-consuming and labor-intensive. It is also difficult to automatically generate design drawings that avoid patents, making it difficult to efficiently develop and manufacture high-performance, low-cost semiconductors. To solve these issues, a system that can efficiently learn patent information and automatically generate design drawings that avoid patents is needed.

[0715] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0716] In this invention, the server includes means for pre-learning patent information, means for acquiring semiconductor circuit design drawings from a patent database, means for extracting the acquired design drawings as image data, means for inputting the image data into an artificial intelligence model to extract and learn features of the design drawings, means for receiving prompts from a user and searching for related patent information, and means for providing the search results to the user. This enables efficient learning and analysis of patent information, and realizes the automatic generation of design drawings that avoid patents.

[0717] "Patent information" refers to the technical information and invention content described in patent documents.

[0718] "Pre-training" refers to the process of training a model with relevant data before it performs a specific task.

[0719] "Patent database" refers to an information system that collects, stores, and provides patent documents in a searchable format.

[0720] A "semiconductor circuit design drawing" refers to a drawing that shows the circuit structure and layout of a semiconductor device.

[0721] "Image data" refers to data that represents visual information in digital form.

[0722] An "artificial intelligence model" refers to an algorithm or system built using techniques such as machine learning and deep learning to perform a specific task.

[0723] "Feature extraction" refers to the process of identifying important patterns or attributes from data and converting them into a numerical representation.

[0724] "Prompt sentence" refers to text containing instructions or questions that a user enters into a system.

[0725] "Search Results" refers to the collection of relevant information returned by the system based on a user's query.

[0726] The present invention relates to a system for efficiently learning patent information and automatically generating semiconductor circuit design diagrams that avoid patents. Specific embodiments of this system will be described below.

[0727] The server generates a program for pre-learning patent information. This program retrieves semiconductor circuit design drawings from a patent database and trains the AI ​​model on them. Patent databases used include Google Patents and Espacenet.

[0728] The server accesses a patent database to retrieve patent information, including semiconductor circuit design drawings. Since the retrieved patent documents are usually provided in PDF format, the server uses a PDF analysis library to extract the design drawings as image data.

[0729] The server then preprocesses the extracted image data and inputs it into an artificial intelligence model, which includes image resizing and noise reduction. The preprocessed image data is then input into an artificial intelligence model built using a deep learning framework such as TensorFlow or PyTorch.

[0730] The AI ​​model uses a convolutional neural network (CNN) to extract and learn from the features of the blueprint. Specifically, it recognizes the patterns and circuit structures of the blueprint and saves them as feature vectors. The server stores these feature vectors in a database for later use in searches.

[0731] The user inputs a prompt sentence using a terminal, such as "Please tell me patent information on the latest semiconductor circuit designs." The server analyzes this prompt sentence and extracts related keywords and feature vectors.

[0732] The server searches the database for relevant patent information based on the extracted keywords and feature vectors. The search results are provided to the user via their terminal. Specifically, a list of related patent documents and thumbnail images of blueprints are displayed.

[0733] Examples:

[0734] If a user wants to search for patent information on new semiconductor circuit designs, they can input the following prompt into the generative AI model:

[0735] Example prompt sentence:

[0736] "Please tell me the patent information on the latest semiconductor circuit designs."

[0737] When this prompt is input into the generative AI model, the server retrieves the relevant semiconductor circuit design from the patent database, and the AI ​​model analyzes it to extract features, ultimately providing the user with relevant patent information.

[0738] The above is a specific implementation method of the system according to one embodiment of the present invention.

[0739] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0740] Step 1:

[0741] The server accesses a patent database and obtains patent information including semiconductor circuit design drawings.

[0742] Input: API endpoint of patent database, search query (e.g., keywords or classification code)

[0743] Data processing: Send API requests and retrieve patent documents

[0744] Output: Acquired patent document (PDF format, etc.)

[0745] Specific operation: The server uses the patent database API to search for patent documents based on the specified keywords and classification codes, and downloads the relevant patent documents.

[0746] Step 2:

[0747] The server extracts semiconductor circuit design drawings as image data from the acquired patent documents.

[0748] Input: Acquired patent document (PDF format, etc.)

[0749] Data processing: Extracting blueprints as image data using a PDF parsing library

[0750] Output: Extracted image data

[0751] Specific operation: The server uses a PDF parsing library to extract semiconductor circuit design diagrams from patent documents in image format.

[0752] Step 3:

[0753] The server pre-processes the extracted image data and inputs it into an artificial intelligence model.

[0754] Input: Extracted image data

[0755] Data processing: Perform preprocessing such as image resizing and noise removal

[0756] Output: Preprocessed image data

[0757] Specific operation: The server uses an image processing library (e.g., OpenCV) to resize and denoise the image data and convert it into a format suitable for the artificial intelligence model.

[0758] Step 4:

[0759] The server inputs the preprocessed image data into an artificial intelligence model, which extracts and learns the features of the blueprint.

[0760] Input: Preprocessed image data

[0761] Data processing: Feature extraction using artificial intelligence models (e.g., TensorFlow or PyTorch)

[0762] Output: Feature vector

[0763] Specific operation: The server uses a convolutional neural network (CNN) to extract the patterns and structures of the blueprints from the image data as feature vectors and stores them in a database.

[0764] Step 5:

[0765] The user uses the terminal to enter a prompt sentence.

[0766] Input: Prompt statement (e.g., "Please tell me patent information on the latest semiconductor circuit designs.")

[0767] Data processing: Analyze the prompt sentence and extract relevant keywords and feature vectors

[0768] Output: Parsed keywords and feature vectors

[0769] Specific operation: The user inputs a prompt sentence through the terminal interface and sends it to the server.

[0770] Step 6:

[0771] The server searches the database for relevant patent information based on the analyzed keywords and feature vectors.

[0772] Input: Parsed keywords and feature vectors

[0773] Data processing: generating database queries and performing searches

[0774] Output: Search results (list of relevant patent documents and thumbnail images of blueprints)

[0775] Specific operations: The server generates a database query to search for relevant patent information based on the stored feature vectors.

[0776] Step 7:

[0777] The server provides the search results to the user.

[0778] Input: Search results (list of relevant patent documents and thumbnail images of blueprints)

[0779] Data processing: converting search results into a format that can be displayed to the user

[0780] Output: Search results served to the user

[0781] Specific operation: The server sends the search results to the user's terminal and displays them in a format that the user can view.

[0782] (Application example 1)

[0783] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0784] In conventional semiconductor circuit design, it was difficult to utilize patent information to select optimal designs and adjust manufacturing processes in real time. Furthermore, there was a lack of efficient means to develop and manufacture high-performance, low-cost electronic components while avoiding patents. This led to problems such as insufficient optimization of manufacturing processes and quality control, resulting in a decline in product competitiveness.

[0785] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0786] In this invention, the server includes a means for pre-learning patent information, a means for automatically generating patent-avoiding electronic component design drawings, a means for developing and manufacturing unique high-performance, low-cost electronic components, and a means for selecting an optimal design based on patent information and adjusting the manufacturing process in real time. This makes it possible to utilize patent information to select an optimal design and adjust the manufacturing process in real time. Furthermore, it is possible to efficiently develop and manufacture high-performance, low-cost electronic components while avoiding patent infringement, thereby optimizing the manufacturing process and improving quality control, thereby increasing the competitiveness of the product.

[0787] "Patent information" refers to data about patents that is made public by public institutions such as the Patent Office.

[0788] "Pre-training" is the process of training an artificial intelligence model with relevant data before it performs a specific task.

[0789] An "electronic component design drawing" is a drawing or diagram that shows the structure and function of an electronic component.

[0790] "Auto-generation" is the process of using artificial intelligence or algorithms to generate data and blueprints without manual intervention.

[0791] "High performance and low cost" refers to high performance and low manufacturing and operating costs.

[0792] "Development and manufacturing" is the process of designing new products and technologies and actually producing them.

[0793] An "optimal design" is the most efficient and effective design under specific conditions and constraints.

[0794] The "manufacturing process" is the series of steps that go from raw materials to creating a final product.

[0795] "Real-time" refers to immediate processing or response without delay.

[0796] "Adjustment" refers to changing or modifying a system or process to suit specific purposes or conditions.

[0797] The following system configuration and processing procedure will be described as an embodiment of the present invention.

[0798] The server includes a means for learning patent information in advance, a means for automatically generating electronic component design drawings that avoid patents, a means for developing and manufacturing unique high-performance, low-cost electronic components, and a means for selecting the optimal design based on patent information and adjusting the manufacturing process in real time.

[0799] System Configuration

[0800] 1. Hardware:

[0801] Server: A server equipped with a high-performance CPU and GPU.

[0802] Camera: High-resolution camera mounted on factory robot.

[0803] Factory robots: Robots used to carry out manufacturing processes.

[0804] 2. Software:

[0805] TensorFlow: Used for training and inference of AI models.

[0806] OpenCV: Used for image processing.

[0807] Database: A database for storing patent information.

[0808] Processing steps

[0809] 1. Preliminary study of patent information:

[0810] The server uses TensorFlow to pre-train the electronic component design drawings retrieved from the patent database. During this process, the AI ​​model extracts and learns the features of the design drawings.

[0811] 2. Automatic generation of electronic component design drawings:

[0812] The server uses a trained AI model to automatically generate patent-evading electronic component designs, which are optimized based on patent information.

[0813] 3. Selecting the best design and adjusting the manufacturing process:

[0814] Images captured by cameras mounted on factory robots are processed using OpenCV and sent to a server, which uses AI models to select the optimal design and adjust the manufacturing process in real time.

[0815] Specific examples

[0816] For example, an AI model can be created that selects the optimal circuit design based on electronic component blueprints retrieved from patent databases. This model can then be installed on factory robots and used to adjust the manufacturing process in real time.

[0817] Example prompt sentence:

[0818] "Create an AI model that selects the optimal circuit design based on electronic component blueprints retrieved from patent databases. This model can be installed on factory robots and adjust the manufacturing process in real time."

[0819] In this way, patent information can be used to select optimal designs and adjust manufacturing processes in real time, enabling efficient development and manufacturing of high-performance, low-cost electronic components while avoiding patents, improving manufacturing process optimization and quality control, and enhancing product competitiveness.

[0820] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0821] Step 1:

[0822] The server retrieves electronic component blueprints from a patent database. As input, it requires a query of the patent database, and as output, it obtains the retrieved blueprint data, which is used in the subsequent pre-training process.

[0823] Step 2:

[0824] The server pre-trains the acquired electronic component blueprints using TensorFlow. Blueprint data is required as input, and a trained AI model is obtained as output. In this process, the features of the blueprints are extracted and trained into the AI ​​model.

[0825] Step 3:

[0826] The server uses a trained AI model to automatically generate patent-avoidance electronic component design drawings. Patent information and the trained model are required as input, and an optimized design drawing is obtained as output. This design drawing is optimized based on the patent information.

[0827] Step 4:

[0828] The factory robots are equipped with cameras that capture images of the manufacturing process. As input, real-time images of the manufacturing process are required, and as output, captured image data is obtained. This image data is used for subsequent processing.

[0829] Step 5:

[0830] The server processes the captured image data using OpenCV and inputs it into the AI ​​model. Image data is required as input, and image features are extracted as output. This feature data is used to select the optimal design.

[0831] Step 6:

[0832] The server uses an AI model to select the optimal design and adjust the manufacturing process in real time. Image feature data and a trained model are required as input, and the output is adjusted manufacturing instructions, which are then sent to factory robots to optimize the manufacturing process.

[0833] Step 7:

[0834] The factory robot adjusts the manufacturing process based on instructions from the server and produces products according to the optimal design. Instructions from the server are required as input, and an optimized product is obtained as output. This process achieves optimization and quality control of the manufacturing process.

[0835] Example 2

[0836] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0837] In conventional semiconductor circuit design, it has been difficult to generate new design drawings while avoiding patent infringement. There is a need for a method that can properly analyze patent information and automatically generate design drawings that avoid patent infringement.

[0838] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0839] In this invention, the server includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for generating new design drawings using a generative AI model, means for confirming whether the generated design drawings avoid patent infringement, and means for regenerating the design drawings if patent infringement is confirmed. This makes it possible to automatically generate new semiconductor circuit design drawings that avoid patent infringement based on patent information.

[0840] "Patent information" is detailed data about existing patents, including claims, invention details, drawings, etc.

[0841] "Pre-training" is the process of training a model with a specific dataset to improve its predictive accuracy for a specific task.

[0842] A "semiconductor circuit design diagram" is a drawing that shows the structure and operation of a semiconductor device, and describes in detail the layout and connections of the circuit.

[0843] "Auto-generation" is the process of using artificial intelligence and algorithms to generate new designs and data without human intervention.

[0844] A "generative AI model" is a type of artificial intelligence that has the ability to generate new data and designs for specific tasks.

[0845] A "prompt" is text that is input to a generative AI model and contains instructions or conditions for the model to perform a specific task.

[0846] "Patent infringement" is the act of using a product or technology without permission within the scope of an existing patent, which can lead to legal issues.

[0847] "Regeneration" refers to running the generation process again when a design or data that has already been generated does not meet certain conditions.

[0848] This invention relates to a system for automatically generating new semiconductor circuit design drawings that avoid patent infringement based on patent information. Specific embodiments of this system will be described below.

[0849] Hardware and software used

[0850] Hardware:

[0851] High-performance servers (e.g., servers equipped with GPUs)

[0852] software:

[0853] Deep learning frameworks (e.g. TensorFlow, PyTorch)

[0854] Generative AI models (e.g., GPT-4)

[0855] System Operation

[0856] The server first accesses a patent information database to retrieve existing patent information, which contains detailed data on semiconductor circuit designs. Based on the retrieved patent information, the server then generates a prompt to be input into the generative AI model.

[0857] Example prompt sentence:

[0858] "Please generate a new semiconductor circuit design that does not infringe patent number 123456."

[0859] The server inputs this prompt into a generative AI model to generate a new semiconductor circuit design. The generated design is then compared with patent information to check for patent infringement. This check is performed using a patent text analysis algorithm. Specifically, it checks whether each element of the design is included in the scope of the patent claims.

[0860] If a patent infringement is confirmed, the server inputs the prompt sentence into the generative AI model again to generate a new design. This process is repeated until a patent-free design is generated.

[0861] Specific examples

[0862] For example, to generate a design drawing that does not infringe patent number 123456, the server inputs the following prompt statement into the generative AI model:

[0863] Example prompt sentence:

[0864] "Please generate a new semiconductor circuit design that does not infringe patent number 123456."

[0865] The generative AI model generates a new design based on this prompt. The generated design is compared with patent information to check for patent infringement. If an infringement is confirmed, the server inputs the prompt into the generative AI model again to generate a new design. This process is repeated until a design that does not infringe on a patent is generated.

[0866] The above is a specific embodiment for carrying out the present invention. This system makes it possible to automatically generate new semiconductor circuit design diagrams that avoid patent infringement based on patent information.

[0867] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0868] Step 1:

[0869] Obtaining patent information

[0870] The server accesses a patent information database to retrieve existing patent information. Patent numbers and specific keywords are given as input. The server searches the database based on this and retrieves relevant patent information. The output is patent information including claims, invention details, drawings, etc. Specifically, the server executes SQL queries to extract the required information from the database.

[0871] Step 2:

[0872] Generate prompt statement

[0873] Based on the acquired patent information, the server generates a prompt sentence to be input into the generative AI model. Patent information is given as input. The server analyzes the patent information and generates a prompt sentence that includes conditions for avoiding patent infringement. The output is a prompt sentence to be input into the generative AI model. Specifically, the server extracts the key points of the patent information and constructs a prompt sentence based on them.

[0874] Step 3:

[0875] Generating new blueprints

[0876] The server inputs a prompt statement into the generative AI model to generate a new semiconductor circuit design. The prompt statement is given as input. The generative AI model analyzes the prompt statement and generates a new design. The output is a new semiconductor circuit design. Specifically, the generative AI model generates the design using a deep learning algorithm.

[0877] Step 4:

[0878] Confirmation of patent infringement

[0879] The server compares the generated design with patent information to check for patent infringement. A new design and patent information are given as input. The server uses a patent text analysis algorithm to check whether each element of the design is included in the scope of the patent claims. The output is a result of whether there is patent infringement. Specifically, the server compares the design with patent information and evaluates the risk of patent infringement.

[0880] Step 5:

[0881] Regenerate blueprints (if necessary)

[0882] If patent infringement is confirmed, the server inputs the prompt sentence into the generative AI model again to generate a new design. The patent infringement result and the prompt sentence are given as input. The server adjusts the prompt sentence and inputs it into the generative AI model again. The output is a new semiconductor circuit design. Specifically, the server modifies the prompt sentence and runs the generation process again. This process is repeated until a design without patent infringement is generated.

[0883] (Application example 2)

[0884] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0885] In conventional semiconductor circuit design, it was extremely difficult to design in a way that avoided patent infringement, and the manual task of checking patent information required a great deal of effort. Furthermore, there was a lack of technology to automatically generate design drawings that avoided patent infringement, which placed a heavy burden on designers. Furthermore, there were insufficient means to check whether the generated design drawings infringed on existing patents, which increased the risk of patent infringement.

[0886] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for verifying whether the generated design drawings infringe existing patents, and means for exporting the design drawings. This enables the automatic generation of new semiconductor circuit design drawings based on patent information, enabling efficient design to avoid patent infringement. Furthermore, it is possible to automatically verify whether the generated design drawings infringe existing patents, thereby reducing the risk of patent infringement.

[0887] "Patent information" refers to information about published patent documents and patent applications, which indicates the technical content and scope of rights of the patents.

[0888] "Pre-training" refers to the process by which an artificial intelligence or machine learning model is trained in advance using a specific dataset.

[0889] A "semiconductor circuit blueprint" is a drawing or blueprint that shows the structure and operation of a semiconductor device, and plays an important role in the manufacturing process.

[0890] "Auto-generation" refers to the process of using artificial intelligence and algorithms to automatically generate blueprints and data without human intervention.

[0891] "Verification" refers to the process of verifying whether the generated design drawings infringe on existing patents.

[0892] "Export" refers to the process of outputting the generated design drawings in a format such as PDF or CAD file.

[0893] "Artificial intelligence" refers to systems and programs that mimic human intelligence using technologies such as machine learning and deep learning.

[0894] A system for implementing this invention includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for verifying whether the generated design drawings infringe existing patents, and means for exporting the design drawings.

[0895] System configuration

[0896] The system consists of the following main components:

[0897] 1. Server: Pre-learns patent information and performs processing to generate and verify design drawings.

[0898] 2. Terminal: Provides an interface for users to input design requirements and review and export the generated design drawings.

[0899] 3. Database: Storage for patent information and generated design drawings.

[0900] Hardware and software used

[0901] Hardware: High-performance servers, user devices (smartphones, tablets, PCs, etc.)

[0902] software:

[0903] Generative AI model: Uses OpenAI's API to generate blueprints based on patent information.

[0904] Patent Database API: An API for obtaining patent information.

[0905] Design Tool API: API for exporting generated design drawings in PDF and CAD formats.

[0906] Data processing and calculation

[0907] The server uses a patent database API to obtain the latest patent information and pre-trains the generative AI model. When the user inputs design requirements from their device, the server uses the generative AI model to generate a semiconductor circuit design that avoids patents. The generated design is compared with the patent database to check for patent infringement. Once the check is complete, the design is exported in PDF or CAD format using the design tool API.

[0908] Specific examples

[0909] For example, if a user wants to design a "low-power, high-performance semiconductor circuit," the following prompt sentence is used:

[0910] Prompt Sentence Examples

[0911] Patent information: [Patent information details]

[0912] User requirement: low-power, high-performance semiconductor circuit

[0913] Generate patent-free semiconductor circuit designs.

[0914] Based on this prompt, the server uses a generative AI model to generate a new semiconductor circuit design that bypasses the patent, checks for patent infringement, and then exports the design, allowing the user to efficiently obtain a design that bypasses the patent.

[0915] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0916] Step 1:

[0917] The server uses a patent database API to retrieve the latest patent information. The input is patent information from the patent database API, and the output is a dataset of the retrieved patent information. This dataset contains detailed information about patent documents and patent applications.

[0918] Step 2:

[0919] The server pre-trains the generative AI model using the acquired patent information. The input is a dataset of patent information, and the output is a generative AI model that has learned from the patent information. This process involves data processing based on the patent information and model training.

[0920] Step 3:

[0921] A user inputs design requirements using a terminal. The input is the user's design requirements (e.g., "a low-power, high-performance semiconductor circuit"), and the output is the transmission of the design requirements to the server. The terminal provides an interface that transmits the requirements entered by the user to the server.

[0922] Step 4:

[0923] The server receives the user's design requirements and uses a generative AI model to generate a semiconductor circuit design that avoids patents. The input is the user's design requirements and a pre-trained generative AI model, and the output is the generated semiconductor circuit design. In this process, the generative AI model generates a design based on a prompt.

[0924] Step 5:

[0925] The server checks whether the generated design drawings infringe on existing patents. The input is a dataset of the generated design drawings and patent information, and the output is a result of whether or not there is patent infringement. In this process, the generated design drawings are matched with patent information to assess the risk of patent infringement.

[0926] Step 6:

[0927] The server exports the verified design drawings using the design tool API. The input is the verified design drawings, and the output is a PDF or CAD file. This process outputs the design drawings in a format that can be used by the user.

[0928] Step 7:

[0929] The user downloads and checks the exported blueprints using a terminal. The input is a blueprint file in PDF or CAD format, and the output is the blueprint saved on the user's terminal. The terminal provides an interface for the user to check and use the blueprints.

[0930] Example 3

[0931] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0932] Conventional semiconductor manufacturing processes have the problem of high costs for producing high-performance semiconductors. It is also difficult to design new semiconductor circuits while avoiding patent infringement. Furthermore, the process from creating design drawings to manufacturing is complex, requiring efficient operation.

[0933] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0934] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for transmitting the generated semiconductor circuit design drawings to a manufacturing device, means for the manufacturing device to manufacture semiconductors based on the design drawings, and means for developing and manufacturing high-performance, low-cost semiconductors. This makes it possible to efficiently manufacture high-performance, low-cost semiconductors and design new semiconductor circuits while avoiding patent infringement.

[0935] "Patent information" refers to technical information contained in publicly available patent documents and patent databases.

[0936] "Pre-training" refers to the process of training a model with relevant data before it performs a specific task.

[0937] A "semiconductor circuit design diagram" is a drawing that shows the circuit configuration and layout of a semiconductor device.

[0938] "Automated generation" refers to the process of using artificial intelligence and algorithms to generate blueprints and data without human intervention.

[0939] "Manufacturing equipment" refers to machines and facilities used to manufacture semiconductor devices.

[0940] A "communication protocol" refers to the rules and procedures for data communication.

[0941] "Lithography technology" is a technology that uses light to form fine patterns on silicon wafers.

[0942] "High performance and low cost" means high performance and low manufacturing and operating costs.

[0943] "Artificial intelligence" refers to the technology that enables computer systems to learn and reason by imitating human intelligence.

[0944] MODE FOR CARRYING OUT THE INVENTION

[0945] This invention relates to a system that learns patent information in advance, automatically generates semiconductor circuit design drawings that avoid patents, and manufactures high-performance, low-cost semiconductors based on those design drawings. Specific embodiments of this system are described below.

[0946] 1. System Configuration

[0947] This system consists of three main components: a server, a terminal, and a user. The server includes a means for pre-learning patent information, a means for automatically generating semiconductor circuit design drawings that circumvent patents, and a means for transmitting the generated design drawings to manufacturing equipment. The terminal provides an interface for users to input prompt sentences. The user is responsible for operating the system and inputting the necessary prompt sentences.

[0948] 2. Hardware and Software Used

[0949] The server is a computer system equipped with a high-performance processor and large memory capacity. Specifically, it is equipped with a GPU (graphics processing unit) for running the generative AI model. The generative AI model is implemented using machine learning frameworks such as TensorFlow and PyTorch. Publicly available patent databases are used for pre-training patent information. The generated design drawings are sent to the manufacturing equipment using secure communication protocols (e.g., HTTPS or SFTP).

[0950] Manufacturing equipment is used to manufacture semiconductors using lithography technology. Specifically, EUV (extreme ultraviolet) lithography equipment is used. This equipment is used to form fine circuit patterns on silicon wafers.

[0951] 3. Program Processing

[0952] The server inputs the prompt received from the user into a generative AI model to generate an optimal semiconductor circuit design. The generated design is then sent from the server to the manufacturing equipment, which then manufactures semiconductors based on the design.

[0953] 4. Specific Examples

[0954] If a user wants to build a high-performance AI system, they input a prompt sentence like the following into the generative AI model:

[0955] Example prompt sentence:

[0956] "Design a high-performance, low-cost AI processor."

[0957] When this prompt is entered, the server generates an optimal semiconductor circuit design using a generative AI model. The generated design is then sent to the manufacturing equipment using a secure communication protocol, and the manufacturing equipment manufactures semiconductors based on the design. This enables users to build and operate high-performance AI systems without using expensive semiconductors. The flow of the specific processing in Example 3 will be explained using Figure 15.

[0958] Program processing flow

[0959] Step 1:

[0960] The user enters a prompt statement.

[0961] Input: The user accesses the system's interface and enters a prompt statement into a text box.

[0962] Specific operation: The user enters, for example, "Please design a high-performance, low-cost AI processor" and clicks the submit button.

[0963] Output: The prompt text is sent to the server.

[0964] Step 2:

[0965] The server receives the prompt sentence and inputs it into the generative AI model.

[0966] Input: The prompt the server received from the user.

[0967] Specific operation: The server sends the prompt sentence to the generative AI model as an API request.

[0968] Data processing: Convert the prompt sentence into an appropriate format and input it into the generative AI model.

[0969] Output: The generative AI model generates an optimal semiconductor circuit design and sends it back to the server.

[0970] Step 3:

[0971] The server transmits the generated semiconductor circuit design drawing to the manufacturing device.

[0972] Input: Semiconductor circuit design diagrams returned from a generative AI model.

[0973] Specific operation: The server encrypts the generated design drawing and sends it to the manufacturing device using a secure communication protocol (e.g., HTTPS or SFTP).

[0974] Data processing: Encrypt the blueprint data and convert it into a format suitable for the communication protocol.

[0975] Output: Encrypted blueprint sent to fabricator.

[0976] Step 4:

[0977] Manufacturing equipment produces semiconductors based on blueprints.

[0978] Input: Encrypted semiconductor circuit blueprints sent from the server.

[0979] How it works: The manufacturing equipment receives the design drawings, first decrypts them, then analyzes them, prepares silicon wafers, and forms circuits on them using lithography technology.

[0980] Data processing: Analysis of design data and circuit formation using lithography technology.

[0981] Output: Manufactured semiconductor device.

[0982] Step 5:

[0983] Users receive the manufactured semiconductors and build and operate AI systems.

[0984] Input: Manufactured semiconductor devices.

[0985] What happens next: The user receives the manufactured semiconductors and incorporates them into an AI system, for example by attaching the AI ​​processor to a motherboard and assembling the entire system. Then, the user installs the AI ​​software and begins training and inference of machine learning models.

[0986] Data processing: Embedding semiconductor devices and installing AI software.

[0987] Output: A high-performance AI system is built and operational.

[0988] (Application example 3)

[0989] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0990] Conventional semiconductor manufacturing processes require expensive materials and complex processes to produce high-performance semiconductors, resulting in high costs. Furthermore, there was a lack of means to efficiently manufacture semiconductors while avoiding patent infringement. Furthermore, there was also the issue of insufficient optimization of the manufacturing process, resulting in low manufacturing efficiency.

[0991] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0992] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for optimizing the manufacturing process based on the automatically generated semiconductor circuit design drawings, and means for manufacturing semiconductors according to the optimized manufacturing process, thereby enabling the efficient manufacturing of high-performance, low-cost semiconductors.

[0993] "Means for prior learning of patent information" refers to a means for studying the vast number of publicly available semiconductor circuit design drawings and patent documents, and accumulating knowledge to avoid patent infringement.

[0994] The "means for automatically generating semiconductor circuit design drawings that avoid patents" refers to a means for automatically generating new semiconductor circuit design drawings that avoid patent infringement using artificial intelligence.

[0995] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to the development and manufacturing of high-performance semiconductors at low manufacturing costs based on semiconductor circuit design drawings that avoid patents.

[0996] "Means for optimizing the manufacturing process based on automatically generated semiconductor circuit design drawings" refers to means for optimizing each step of the manufacturing process based on automatically generated semiconductor circuit design drawings, and for efficiently manufacturing semiconductors.

[0997] The "means for manufacturing a semiconductor according to an optimized manufacturing process" refers to a means for actually manufacturing a semiconductor according to an optimized manufacturing process.

[0998] A system for implementing this invention includes a means for prior learning of patent information, a means for automatically generating semiconductor circuit design drawings that circumvent patents, a means for developing and manufacturing unique high-performance, low-cost semiconductors, a means for optimizing the manufacturing process based on the automatically generated semiconductor circuit design drawings, and a means for manufacturing semiconductors in accordance with the optimized manufacturing process.

[0999] 1. Program Generation

[1000] The server collects and studies a vast number of publicly available semiconductor circuit designs and patent documents to learn about patent information. It then uses artificial intelligence to automatically generate semiconductor circuit designs that avoid patents. Based on these designs, the server develops and manufactures unique, high-performance, low-cost semiconductors.

[1001] 2. Program processing explanation

[1002] The server uses the following hardware and software:

[1003] Hardware: Factory robots (e.g., general-purpose robotic arms), semiconductor manufacturing equipment (e.g., lithography equipment)

[1004] Software: Python 3.x, JSON library, artificial intelligence model (e.g. TensorFlow)

[1005] The server first collects patent information and stores it in JSON format. It then uses an artificial intelligence model to generate a semiconductor circuit design that avoids patents. Based on this design, it runs an algorithm to optimize the manufacturing process and determines the optimal manufacturing process. Finally, factory robots manufacture semiconductors according to this optimized process.

[1006] 3. Examples and prompts

[1007] Specific examples

[1008] For example, the patent information collected by the server includes existing semiconductor circuit blueprints and patent documents. Based on this information, the AI ​​model generates a new semiconductor circuit blueprint. Based on this blueprint, each step of the manufacturing process (e.g., layer deposition, photolithography, etching, doping) is optimized.

[1009] Prompt Sentence Examples

[1010] "You are asked to develop an application that will be installed on a factory robot. This application will optimize the semiconductor manufacturing process based on automatically generated semiconductor circuit design drawings, and efficiently manufacture high-performance, low-cost semiconductors."

[1011] Thus, specific embodiments for carrying out the invention have been shown.

[1012] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1013] Step 1:

[1014] The server collects a huge number of publicly available semiconductor circuit design drawings and patent documents. It stores this data in JSON format. The input is publicly available semiconductor circuit design drawings and patent documents, and the output is patent information data stored in JSON format. Specifically, the server collects data using web scraping and APIs and stores it in local storage.

[1015] Step 2:

[1016] The server pre-learns the patent information based on the collected patent information. The input is the patent information data saved in step 1, and the output is the trained artificial intelligence model. Specifically, it uses a machine learning library such as TensorFlow to train the model using patent information as learning data.

[1017] Step 3:

[1018] The server uses a trained AI model to automatically generate semiconductor circuit design drawings that circumvent patents. The input is the trained AI model and design requirements, and the output is an automatically generated semiconductor circuit design drawing. Specifically, the server inputs a prompt statement into the generating AI model to generate the design drawing.

[1019] Step 4:

[1020] The server optimizes the manufacturing process based on automatically generated semiconductor circuit design drawings. The input is the automatically generated semiconductor circuit design drawing, and the output is the optimized manufacturing process steps. Specifically, it uses algorithms to optimize each step of the manufacturing process (e.g., layer deposition, photolithography, etching, doping).

[1021] Step 5:

[1022] The server manufactures semiconductors according to an optimized manufacturing process. The inputs are the steps in the optimized manufacturing process, and the output is the manufactured high-performance, low-cost semiconductors. Specifically, it controls factory robots and operates manufacturing equipment to manufacture semiconductors.

[1023] In this manner, the specific operations performed in each processing step have been described.

[1024] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1025] "Example 1"

[1026] One embodiment of the present invention provides a system that includes a means for pre-learning patent information, a means for automatically generating semiconductor circuit design diagrams that circumvent patents, and a means for developing and manufacturing unique, high-performance, low-cost semiconductors, as well as an emotion engine that recognizes user emotions. This emotion engine recognizes emotions from the user's voice, facial expressions, text input, etc., and controls the operation of the semiconductor based on the recognized emotion. For example, if the user expresses anger, the emotion engine analyzes that information and adjusts the operation of the semiconductor. This enables the semiconductor to perform optimal operations in response to the user's emotions.

[1027] "Example 2"

[1028] Furthermore, as another embodiment of the present invention, a specific technique is provided for an emotion engine to recognize a user's emotion. Specifically, a voice analysis technique that recognizes an emotion from a user's voice, an image analysis technique that recognizes an emotion from a user's facial expression, a natural language processing technique that recognizes an emotion from a user's text input, etc. are used. By combining these techniques, emotion recognition with higher accuracy becomes possible.

[1029] "Example 3"

[1030] Furthermore, as another embodiment of the present invention, a specific technique for controlling the operation of a semiconductor based on the emotion recognized by an emotion engine is provided. Specifically, control such as adjusting the clock frequency of the semiconductor or giving priority to the execution of a specific process is performed according to the type and intensity of the emotion recognized by the emotion engine. This enables optimal control of the semiconductor operation according to the user's emotion.

[1031] The processing flow of each embodiment will be described below.

[1032] "Example 1"

[1033] Step 1: Receive input from the user (voice, facial expression, text, etc.).

[1034] Step 2: The emotion engine recognizes emotions from the user's input.

[1035] Step 3: Control the operation of the semiconductor based on the recognized emotion.

[1036] "Example 2"

[1037] Step 1: Perform voice analysis to recognize emotions from the user's voice.

[1038] Step 2: Perform image analysis to recognize emotions from the user's facial expressions.

[1039] Step 3: Perform natural language processing to recognize emotions from the user's text input.

[1040] "Example 3"

[1041] Step 1: Determine the type and intensity of the emotion recognized by the emotion engine.

[1042] Step 2: Adjust the clock frequency of the semiconductor based on the result of the determination.

[1043] Step 3: Based on the judgment result, a specific process is executed with priority.

[1044] Example 1

[1045] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1046] In conventional semiconductor circuit design, it was difficult to develop high-performance, low-cost semiconductors while avoiding patent infringement. Furthermore, there was no technology to optimize semiconductor operation according to user emotions, making it difficult to improve the user experience.

[1047] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1048] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for recognizing user emotions, and means for controlling the operation of the semiconductor based on the recognized emotions. This makes it possible to develop high-performance, low-cost semiconductors while avoiding patent infringement, and also enables the semiconductors to perform optimal operations according to the user's emotions.

[1049] "Patent information" refers to semiconductor circuit design drawings and related technical documents registered in patent databases.

[1050] "Pre-learning" refers to the process by which artificial intelligence extracts features based on patent information and learns from them.

[1051] A "semiconductor circuit design drawing" refers to a drawing or design data that shows the circuit structure of a semiconductor.

[1052] "Automatic generation" refers to the process by which artificial intelligence automatically creates new semiconductor circuit designs based on patent information.

[1053] "High performance and low cost" refers to high performance and low manufacturing costs.

[1054] "Emotion recognition" is a technology that analyzes and recognizes emotions from a user's voice, facial expressions, text input, etc.

[1055] "Behavioral control" refers to the process of adjusting the behavior of a semiconductor based on perceived emotions.

[1056] This invention is a system that learns patent information in advance, automatically generates semiconductor circuit designs that avoid patents, and controls the operation of semiconductors by recognizing the user's emotions. This system works in cooperation with a server, a terminal, and a user.

[1057] The server accesses a patent database to retrieve semiconductor circuit design drawings. This database contains various semiconductor circuit design drawings that have been published in patents. The server inputs the retrieved design drawings into an AI model using a machine learning framework such as TensorFlow or PyTorch to extract the design features. The AI ​​model learns based on the extracted features and pre-learns patent information.

[1058] The server then uses the trained AI model to generate new semiconductor circuit design drawings. In this process, the AI ​​takes existing patent information into account when generating new designs. The resulting designs reduce the risk of patent infringement while enabling high-performance, low-cost semiconductors.

[1059] The device inputs the user's voice, facial expressions, text input, etc. into the emotion engine. This emotion engine is built using OpenCV and Microsoft Azure's emotion recognition API. The emotion engine recognizes the user's emotions from the input data and analyzes that information.

[1060] The server controls the semiconductor's operation based on the user's emotional information obtained from the emotion engine. For example, if the user expresses anger, the server analyzes that information and adjusts the semiconductor's operation. This enables the semiconductor to perform optimal operations according to the user's emotions.

[1061] As a specific example, the server retrieves the semiconductor circuit design drawing for "Patent No. 123456" from a patent database and inputs it into an AI model for learning. Next, it uses the trained AI model to generate a new design drawing and confirms that it is designed to avoid patent No. 123456. The device inputs the user's voice input, "I'm very tired today," into the emotion engine, which recognizes that the user is feeling "tired." Based on this emotion information, the server switches the semiconductor's operation to energy-saving mode.

[1062] Examples of prompts include, "Build an AI model that learns from semiconductor circuit designs retrieved from a patent database and generates new designs that avoid patents. Also, describe a system that recognizes user emotions and controls the operation of semiconductors based on those emotions."

[1063] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1064] Step 1:

[1065] The server accesses the patent database and retrieves the semiconductor circuit design. Specifically, the server uses an API to download the design data for "Patent Number 123456." The input is the patent number, and the output is the design data. This design data is used to train the subsequent AI model.

[1066] Step 2:

[1067] The server inputs the acquired blueprint data into an AI model using machine learning frameworks such as TensorFlow and PyTorch. The AI ​​model extracts the features of the blueprint and performs learning. The input is the blueprint data, and the output is the trained model. In this process, image processing technology is used to analyze the blueprint and extract features.

[1068] Step 3:

[1069] The server uses the trained AI model to generate new semiconductor circuit designs. Specifically, the AI ​​model combines features from existing designs to generate new designs. The input is the trained model, and the output is the new design. This design is designed to avoid patent infringement.

[1070] Step 4:

[1071] The server checks the generated design against a patent database to determine whether it avoids patent infringement. The input is the new design, and the output is information about whether or not there is patent infringement. This process uses natural language processing technology to analyze the patent documents.

[1072] Step 5:

[1073] The device inputs the user's voice, facial expressions, and text input into the emotion engine. Specifically, it collects voice and facial expression data using a microphone and camera. The input is the user's voice and facial expression data, and the output is emotional information. The emotion engine is built using OpenCV and Microsoft Azure's emotion recognition API.

[1074] Step 6:

[1075] The emotion engine recognizes the user's emotions from input data and analyzes that information. The input is voice and facial expression data, and the output is the recognized emotional information. For example, if a user inputs "I'm very tired today," the emotion engine will recognize that the user is feeling "tired."

[1076] Step 7:

[1077] The server controls the operation of the semiconductor based on the emotional information obtained from the emotion engine. The input is emotional information, and the output is the adjusted operation of the semiconductor. Specifically, if the user is tired, the server switches the operation of the semiconductor to energy-saving mode. For example, it makes adjustments such as lowering the CPU clock speed.

[1078] (Application example 1)

[1079] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1080] Conventional semiconductor design systems have difficulty learning patent information in advance and automatically generating designs that avoid patent infringement. Furthermore, they lack the ability to recognize user emotions and optimize semiconductor operation based on those emotions, resulting in reduced work efficiency and user satisfaction. This has created a need for efficient development and manufacturing of high-performance, low-cost semiconductors while avoiding the risk of patent infringement.

[1081] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1082] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means including an emotion engine that recognizes user emotions, and means for controlling the operation of the semiconductor based on the user emotions. This makes it possible to efficiently develop and manufacture high-performance, low-cost semiconductors that perform optimal operations in response to user emotions while avoiding the risk of patent infringement.

[1083] "Means for pre-learning patent information" refers to a means for obtaining semiconductor circuit design drawings from publicly available patent databases and learning their features using artificial intelligence.

[1084] The "means for automatically generating semiconductor circuit design drawings that avoid patents" refers to a means for automatically generating new semiconductor circuit design drawings that avoid the risk of patent infringement using artificial intelligence based on learned patent information.

[1085] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to means for efficiently developing and manufacturing high-performance, low-cost semiconductors based on design drawings that avoid patents.

[1086] The "means including an emotion engine for recognizing a user's emotion" refers to a means including an engine for analyzing and recognizing an emotion from the user's voice, facial expression, text input, and the like.

[1087] The "means for controlling the operation of a semiconductor based on the user's emotions" refers to a means for generating a control signal for optimizing the operation of a semiconductor based on the recognized user's emotions, and adjusting the operation of the semiconductor.

[1088] A system for implementing this invention includes a means for pre-learning patent information, a means for automatically generating semiconductor circuit design drawings that circumvent patents, a means for developing and manufacturing unique high-performance, low-cost semiconductors, a means including an emotion engine that recognizes user emotions, and a means for controlling the operation of the semiconductor based on the user emotions.

[1089] Program processing explanation

[1090] 1. Hardware:

[1091] Robot body (e.g. industrial robot)

[1092] Emotion recognition camera (e.g. depth camera)

[1093] Voice recognition microphone (e.g. high-sensitivity microphone)

[1094] 2. Software:

[1095] Patent information learning AI (e.g., custom models using TensorFlow)

[1096] Emotion recognition engine (e.g., emotion analysis API)

[1097] Semiconductor design automatic generation algorithm (e.g., implemented in Python)

[1098] 3. Data processing and calculation:

[1099] The server retrieves semiconductor circuit design drawings from a patent database and inputs them into an AI model to extract features.

[1100] Data obtained from the emotion recognition camera and voice recognition microphone is input into the emotion recognition engine to analyze the user's emotions.

[1101] Based on the emotion data, a control signal is generated to optimize the operation of the semiconductor.

[1102] Specific examples

[1103] For example, if a worker in a factory feels stressed, the emotion recognition engine can analyze that emotion and instruct the robot to adjust its working speed, thereby reducing worker stress and providing an efficient working environment.

[1104] Prompt Sentence Examples

[1105] An example of a prompt to be input to the generative AI model is as follows:

[1106] "Learn from semiconductor circuit designs obtained from patent databases and generate new designs that avoid patents. Also, recognize user emotions and generate control signals that optimize the operation of semiconductors based on those emotions."

[1107] In this way, the system will be able to utilize patent information to efficiently develop and manufacture high-performance, low-cost semiconductors that perform optimal operations in response to user emotions.

[1108] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1109] Step 1:

[1110] The server retrieves semiconductor circuit design drawings from a patent database. The input is the patent database, and the output is the retrieved semiconductor circuit design drawings. These design drawings are then input into the patent information learning AI.

[1111] Step 2:

[1112] The server extracts features from the acquired semiconductor circuit design using patent information learning AI. The input is the acquired semiconductor circuit design, and the output is the extracted feature data. This feature data is used to generate designs to circumvent patents.

[1113] Step 3:

[1114] The server automatically generates patent-evading semiconductor circuit design drawings based on the extracted feature data. The input is the feature data, and the output is a new semiconductor circuit design drawing. This design drawing is used to develop and manufacture unique, high-performance, low-cost semiconductors.

[1115] Step 4:

[1116] The device acquires the user's emotional data using an emotion recognition camera and a voice recognition microphone. The input is the user's voice and facial expressions, and the output is emotional data. This emotional data is input into the emotion recognition engine.

[1117] Step 5:

[1118] The server analyzes the acquired emotion data using an emotion recognition engine. The input is the emotion data, and the output is the analyzed emotion information. This emotion information is used to control the operation of the semiconductor.

[1119] Step 6:

[1120] The server generates a control signal for optimizing the operation of the semiconductor based on the analyzed emotion information. The input is the analyzed emotion information, and the output is the control signal. The control signal is used to adjust the operation of the semiconductor.

[1121] Step 7:

[1122] The terminal uses the generated control signal to adjust the operation of the semiconductor. The input is the control signal, and the output is the adjusted operation of the semiconductor. This allows optimal operation according to the user's emotions to be achieved.

[1123] Example 2

[1124] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1125] In conventional semiconductor circuit design, avoiding patent infringement is extremely difficult, requiring a great deal of time and cost. Furthermore, there is a lack of technology to accurately recognize user emotions, making it difficult to improve the user experience. To solve these problems, we need technology that effectively utilizes patent information to automatically generate designs that avoid patent infringement, and technology that can accurately recognize user emotions.

[1126] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1127] In this invention, the server includes a means for prior learning of patent information, a means for automatically generating semiconductor circuit design drawings that avoid patents, and a means for verifying whether the generated design drawings infringe on existing patents, thereby enabling new semiconductor circuit design drawings to be efficiently generated while avoiding patent infringement.

[1128] Furthermore, the terminal includes a means for recognizing user emotions, a means for acquiring and analyzing voice data, a means for acquiring and analyzing facial expression data, a means for acquiring and analyzing text data, and a means for integrating emotion recognition results, thereby enabling highly accurate recognition of user emotions and improving the user experience.

[1129] "Patent information" refers to technical information contained in publicly available patent documents and patent databases.

[1130] "Pre-training" is the process of training a model with a specific dataset to improve its performance on a specific task.

[1131] A "semiconductor circuit design diagram" is a drawing or diagram that shows the structure and function of a semiconductor device.

[1132] "Auto-generation" refers to the process of using artificial intelligence and algorithms to create new designs and data without human intervention.

[1133] "Verification" refers to the process of verifying whether the generated design drawings infringe on existing patents.

[1134] "User emotion" refers to the psychological state that the user feels, such as joy, sadness, anger, or surprise.

[1135] "Voice data" refers to data that is a digital recording of a user's speech or voice.

[1136] "Analysis" is the process of examining data in detail to understand its meaning and structure.

[1137] "Facial expression data" refers to image data that captures the facial expression of a user.

[1138] "Text data" refers to character information entered by the user.

[1139] "Emotion recognition results" refer to the analysis results of emotions obtained from voice, facial expressions, and text.

[1140] "Integration" refers to the process of combining multiple data and analysis results into one.

[1141] This invention is a system that pre-learns patent information, automatically generates semiconductor circuit designs that avoid patents, and recognizes user emotions with high accuracy. Specific embodiments of this system are described below.

[1142] Automatic generation of semiconductor circuit design diagrams

[1143] To pre-train the patent information, the server retrieves relevant patent information from a patent information database. The patent information database contains a huge number of publicly available semiconductor circuit design diagrams. The server analyzes this patent information using natural language processing technology (e.g., the BERT model) and extracts important technical elements.

[1144] Next, the server uses a generative AI model (e.g., TensorFlow, PyTorch) to automatically generate new semiconductor circuit design diagrams. The generative AI model generates design patterns that avoid patents based on patent information it has learned in advance. The generated design diagrams are compared with a patent information database to ensure they do not infringe on existing patents.

[1145] Examples:

[1146] Hardware used: High-performance server (e.g., server with GPU)

[1147] Software used: TensorFlow, PyTorch

[1148] Example prompt: "Generate a new semiconductor circuit design based on the information in patent number 123456."

[1149] Recognizing user emotions with an emotion engine

[1150] To recognize the user's emotions, the device uses a combination of voice analysis technology, image analysis technology, and natural language processing technology. Specifically, the device captures the user's voice data with a microphone and analyzes emotions using voice analysis software (e.g., Google Cloud Speech-to-Text). It also uses a camera to capture the user's facial expressions and recognizes emotions using image analysis software (e.g., OpenCV). It also analyzes text data entered by the user with natural language processing software (e.g., the BERT model) to recognize emotions.

[1151] These analysis results are integrated by an integrated algorithm within the device to determine the final emotion, enabling highly accurate recognition of the user's emotions and appropriate responses.

[1152] Examples:

[1153] Hardware used: smartphone, tablet, PC

[1154] Software used: Google Cloud Speech-to-Text, OpenCV, BERT model

[1155] Example prompt: "Recognize emotions from the user's voice, facial expressions, and text input."

[1156] This system makes it possible to efficiently generate new semiconductor circuit designs while avoiding patent infringement, and also improves the user experience by recognizing user emotions with high accuracy.

[1157] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1158] Automatic generation of semiconductor circuit design diagrams

[1159] Step 1: Obtain patent information

[1160] The server retrieves relevant patent information from a patent information database. As input, it sends a query containing patent numbers and keywords to the database. As output, it returns the relevant patent information.

[1161] What happens: The server sends the query "Get information about patent number 123456" to the database.

[1162] Step 2: Analyzing patent information

[1163] The server analyzes the acquired patent information using natural language processing technology. The patent document is given as input, and important technical elements are extracted as output.

[1164] Specific operation: The server uses the BERT model and instructs the server to "extract the technical elements of patent number 123456."

[1165] Step 3: Generate a new blueprint

[1166] The server uses a generative AI model based on the analysis results to generate a new semiconductor circuit design. The extracted technical elements are given as input, and the new design is generated as output.

[1167] Specific operation: The server inputs a prompt to the generative AI model: "Generate a new design that avoids patent number 123456."

[1168] Step 4: Check the design for patent infringement

[1169] The server checks the generated design against a patent information database to determine whether the design infringes any existing patents. The generated design is given as input, and the server returns whether or not there is any patent infringement as output.

[1170] Specific operation: The server requests a database check to see if the generated design drawing infringes patent number 123456.

[1171] Recognizing user emotions with an emotion engine

[1172] Step 1: Acquire and analyze audio data

[1173] The device captures the user's voice data with a microphone and analyzes their emotions using voice analysis software. The voice data is given as input, and the emotion analysis results are obtained as output.

[1174] Specific operation: The device instructs the voice analysis software to "capture the user's voice and analyze their emotions."

[1175] Step 2: Acquire and analyze facial expression data

[1176] The device uses a camera to capture the user's facial expressions and image analysis software to recognize emotions. Facial expression data is given as input, and emotion analysis results are obtained as output.

[1177] Specific operation: The device instructs the image analysis software to "capture the user's facial expression and analyze their emotions."

[1178] Step 3: Acquire and analyze text data

[1179] The device analyzes the text data entered by the user using natural language processing software to recognize emotions. The text data is given as input, and the emotion analysis results are obtained as output.

[1180] What it does: The device instructs the natural language processing software to "analyze the user's text input and recognize emotions."

[1181] Step 4: Integrating emotion recognition results

[1182] The device integrates the emotion recognition results obtained from voice, facial expressions, and text to determine the final emotion. Each analysis result is given as input, and the integrated emotion recognition result is obtained as output.

[1183] Specific operation: The device instructs the integration algorithm to "integrate the emotion recognition results from voice, facial expression, and text to determine the final emotion."

[1184] (Application example 2)

[1185] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1186] In conventional semiconductor circuit design, it was difficult to generate new design drawings while avoiding patent infringement. Furthermore, to improve work efficiency in factories, it was necessary to recognize the emotions of workers and provide appropriate feedback, but there was a lack of effective means to achieve this.

[1187] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1188] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for recognizing worker emotions, means for providing feedback to optimize work efficiency, and means for developing and manufacturing unique high-performance, low-cost semiconductors. This makes it possible to generate new semiconductor circuit design drawings while avoiding patent infringement, and further to optimize work efficiency by recognizing worker emotions.

[1189] "Patent information" refers to information about patents that has been made public by public institutions such as the Patent Office.

[1190] "Pre-training" is the process of pre-training an artificial intelligence model using a specific dataset.

[1191] A "semiconductor circuit design diagram" is a drawing that shows the structure and operation of a semiconductor device.

[1192] "Auto-generation" is the process of using artificial intelligence and algorithms to generate new designs and data without human intervention.

[1193] "Workers" are people who work in factories or manufacturing sites.

[1194] "Emotion recognition" is a technology that identifies a person's emotional state by analyzing data such as voice, facial expressions, and text.

[1195] "Feedback" is information or instructions that a system or device provides to a user.

[1196] "High performance and low cost" means high performance and low manufacturing and operating costs.

[1197] "Development and manufacturing" refers to the process of designing new products and technologies and actually producing them.

[1198] A "system" is an integrated structure in which multiple elements and functions are combined to achieve a specific purpose.

[1199] A system for implementing this invention includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for recognizing worker emotions, means for providing feedback to optimize work efficiency, and means for developing and manufacturing unique high-performance, low-cost semiconductors.

[1200] Hardware and software used

[1201] Hardware: Factory robots, cameras, microphones, displays

[1202] Software: TensorFlow (AI model learning and inference), OpenCV (image analysis), NLTK (natural language processing), PyAudio (audio analysis)

[1203] Data processing and calculation

[1204] Collecting and learning patent information

[1205] The server collects patent information from patent databases and trains the AI ​​model with the following prompts:

[1206] "Generate new semiconductor circuit design diagrams based on patent information."

[1207] Semiconductor circuit design generation

[1208] The server uses the trained AI model to generate new design drawings based on patent information, and then checks whether the generated design drawings infringe on patents.

[1209] emotion recognition

[1210] The device uses a camera and microphone to collect facial expressions and voice data from the worker. It uses OpenCV to analyze facial expressions and NLTK to analyze voice and text input. An emotion engine integrates these data and recognizes the worker's emotions.

[1211] Providing Feedback

[1212] The server then displays feedback on the display to optimize work efficiency based on the recognized emotions. For example, if the server detects fatigue in the worker's facial expression, it will display a message urging the worker to take a break.

[1213] In this way, the system can generate new semiconductor circuit designs while avoiding patent infringement, and can also recognize the emotions of workers to optimize work efficiency.

[1214] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1215] Step 1:

[1216] The server collects patent information from patent databases. It uses the URL and API key of the patent database as input and obtains a dataset of patent information as output. Specifically, the server accesses the patent database and extracts patent information based on specific conditions.

[1217] Step 2:

[1218] The server pre-trains an AI model based on the collected patent information. It uses a patent information dataset as input and obtains a trained AI model as output. Specifically, the server uses TensorFlow to supply the patent information to the AI ​​model and optimizes the model's parameters.

[1219] Step 3:

[1220] The server generates a new semiconductor circuit design using the trained AI model. It uses the prompt "Please generate a new semiconductor circuit design based on patent information" as input and obtains the new semiconductor circuit design as output. Specifically, the server inputs the prompt into the generating AI model and obtains the generated design.

[1221] Step 4:

[1222] The server checks whether the generated semiconductor circuit design drawing does not infringe any patents. It uses the generated design drawing as input and obtains a result indicating whether or not there is patent infringement as output. Specifically, the server compares patent information with the generated design drawing to assess the risk of patent infringement.

[1223] Step 5:

[1224] The terminal uses a camera and microphone to collect the worker's facial expressions and voice. It uses real-time data from the camera and microphone as input and obtains the collected facial and voice data as output. Specifically, the terminal activates the camera and microphone and records the worker's facial expressions and voice.

[1225] Step 6:

[1226] The device analyzes the collected facial and voice data to recognize the worker's emotions. It uses facial and voice data as input and obtains the recognized emotion as output. Specifically, the device analyzes facial expressions using OpenCV and voice data using NLTK.

[1227] Step 7:

[1228] The server provides feedback to optimize work efficiency based on the recognized emotions. The server uses the recognized emotion data as input and obtains a feedback message as output. Specifically, the server analyzes the emotion data, generates an appropriate feedback message, and displays it on a display.

[1229] Step 8:

[1230] The user checks the feedback message displayed on the display and adjusts their behavior as necessary. The feedback message is used as input and the output is improved work efficiency. Specifically, the user checks the display and performs an action such as taking a break.

[1231] Example 3

[1232] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1233] Conventional semiconductor manufacturing systems have had the difficulty of producing high-performance semiconductors at low cost, and have also had the problem of being unable to control operation in response to user emotions. Furthermore, semiconductor circuits designed without taking patent information into consideration have a high risk of patent infringement, making commercialization difficult.

[1234] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1235] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for controlling the operation of the semiconductor based on emotions recognized by the emotion engine, and means for transmitting the generated semiconductor circuit design drawings to semiconductor manufacturing equipment. This makes it possible to manufacture high-performance, low-cost semiconductors and also realizes optimal operation control according to the user's emotions.

[1236] "Means for prior study of patent information" refers to a means for studying the vast number of publicly available semiconductor circuit design drawings and related patent documents, and acquiring knowledge to avoid patent infringement.

[1237] The "means for automatically generating semiconductor circuit design drawings that circumvent patents" refers to a means for automatically generating new semiconductor circuit design drawings that pose a low risk of patent infringement, taking patent information into consideration, using artificial intelligence.

[1238] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to the development and manufacturing of high-performance semiconductors at low manufacturing costs, based on automatically generated semiconductor circuit design drawings and through processes such as lithography and etching.

[1239] "Means for controlling the operation of a semiconductor based on emotions recognized by an emotion engine" refers to means for the emotion engine to recognize the user's emotions and adjust the clock frequency and processing priority of the semiconductor based on that emotion data.

[1240] The "means for transmitting the generated semiconductor circuit design drawing to the semiconductor manufacturing device" refers to a means for converting the semiconductor circuit design drawing automatically generated by the server into an appropriate format and transmitting it to the semiconductor manufacturing device.

[1241] This invention is a system that pre-learns patent information, automatically generates semiconductor circuit designs that avoid patents, and develops and manufactures unique high-performance, low-cost semiconductors. Furthermore, it has the function of controlling the operation of semiconductors based on emotions recognized by an emotion engine.

[1242] System configuration

[1243] The system consists of the following main components:

[1244] 1. Server: Pre-learns patent information, automatically generates semiconductor circuit design diagrams, and processes data from the emotion engine.

[1245] 2. Terminal: Provides an interface for users to input emotions.

[1246] 3. Semiconductor manufacturing equipment: Manufactures semiconductors based on the design drawings sent from the server.

[1247] Hardware and software used

[1248] Generative AI models: Use deep learning frameworks such as TensorFlow and PyTorch.

[1249] Emotion engine: Uses IBM Watson's Natural Language Understanding API.

[1250] Semiconductor manufacturing equipment: Uses ASML's EUV lithography equipment, etc.

[1251] Program processing

[1252] Preliminary study of patent information

[1253] The server studies a vast number of publicly available semiconductor circuit designs and related patent documents, thereby acquiring knowledge to avoid patent infringement.

[1254] Automatic generation of semiconductor circuit design diagrams

[1255] The server uses a generative AI model to automatically generate semiconductor circuit design drawings that are novel and pose a low risk of patent infringement, taking patent information into account.

[1256] Semiconductor manufacturing

[1257] The server converts the generated semiconductor circuit design drawings into an appropriate format and sends them to semiconductor manufacturing equipment, which then uses processes such as lithography and etching to produce high-performance semiconductors at low manufacturing costs.

[1258] Emotion-based motion control

[1259] The user operates the device to input emotions into the emotion engine, which analyzes the data and recognizes the type and intensity of the emotion. The server receives the data from the emotion engine and executes the semiconductor's operation control algorithm, which adjusts the clock frequency and prioritizes the execution of specific processes.

[1260] Specific examples

[1261] Example 1: When a user is feeling stressed, the emotion engine recognizes the "stress" and the server reduces the clock frequency of the semiconductor to reduce the load on the system.

[1262] Example 2: When a user is concentrating, the emotion engine recognizes this and the server prioritizes specific processes, thereby improving the user's work efficiency.

[1263] Prompt Sentence Examples

[1264] "Please explain how the system adjusts its behavior if the user is stressed."

[1265] "Please explain in detail how the emotion engine controls the semiconductor's behavior based on the emotions it recognizes."

[1266] In this way, the system manufactures high-performance, low-cost semiconductors and realizes optimal operation control according to the user's emotions. The flow of the specific processing in the third embodiment will be described with reference to FIG.

[1267] Step 1:

[1268] The server pre-learns patent information.

[1269] Input: A huge number of publicly available semiconductor circuit designs and related patent documents.

[1270] Data processing: The server uses a deep learning framework (e.g., TensorFlow, PyTorch) to learn from this data and acquire knowledge to avoid patent infringement.

[1271] Output: A knowledge base that takes into account patent information.

[1272] What it does: The server loads the dataset, trains the neural network, and generates a model for patent circumvention.

[1273] Step 2:

[1274] The server automatically generates semiconductor circuit design drawings that circumvent patents.

[1275] Input: Knowledge base considering patent information.

[1276] Data processing: The server uses a generative AI model to automatically generate new semiconductor circuit design drawings that have a low risk of patent infringement.

[1277] Output: Patent-evading semiconductor circuit design.

[1278] Specific operation: The server references the knowledge base and uses a generative AI model to output the optimal circuit design.

[1279] Step 3:

[1280] The server transmits the generated semiconductor circuit design drawing to the semiconductor manufacturing equipment.

[1281] Input: Patent-circuit-evading semiconductor circuit design.

[1282] Data processing: The server converts the design drawing file into an appropriate format and sends it to the control system of the manufacturing equipment.

[1283] Output: Blueprint sent to semiconductor manufacturing equipment.

[1284] Specific operation: The server converts the design drawing into a different format and sends it to the manufacturing device via the network.

[1285] Step 4:

[1286] Semiconductor manufacturing equipment manufactures semiconductors based on design drawings.

[1287] Input: Semiconductor circuit design diagram sent from the server.

[1288] Data processing: Manufacturing equipment produces semiconductors through processes such as lithography, etching, and doping.

[1289] Output: Finished semiconductor.

[1290] Specific operation: The manufacturing equipment reads the blueprint and executes each process in sequence to manufacture semiconductors.

[1291] Step 5:

[1292] The user operates the terminal to input emotions into the emotion engine.

[1293] Input: User emotion data (text and / or voice).

[1294] Data processing: The emotion engine analyzes the data to recognize the type and intensity of emotions.

[1295] Output: Recognized emotion data.

[1296] Specific operation: The user inputs emotions into the device, and the emotion engine analyzes the data.

[1297] Step 6:

[1298] The server receives data from the emotion engine and executes the semiconductor's motion control algorithms.

[1299] Input: Recognized emotion data from the emotion engine.

[1300] Data processing: The server analyzes the emotion data, adjusts the clock frequency, and prioritizes the execution of specific processes.

[1301] Output: Optimized semiconductor operating instructions.

[1302] Specific operation: The server executes the operation control algorithm based on the emotion data and adjusts the operation of the semiconductor.

[1303] In this way, the system will produce high-performance, low-cost semiconductors and achieve optimal operation control according to the user's emotions.

[1304] (Application example 3)

[1305] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1306] With conventional semiconductor manufacturing technology, it is difficult to manufacture high-performance semiconductors at low cost, and there is a high risk of patent infringement. Furthermore, because robots' movements are not optimized according to the emotions of factory workers, work efficiency and worker stress reduction are not fully achieved. To solve these issues, a system is needed that can manufacture high-performance, low-cost semiconductors while avoiding patent infringement, and that can optimize robot movements based on emotion recognition data.

[1307] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1308] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for acquiring emotion recognition data, means for controlling the operation of the semiconductor based on the emotion recognition data, and means for optimizing the operation of the robot in accordance with the emotions of workers in the factory. This enables optimal robot operation in accordance with the emotions of workers while avoiding the risk of patent infringement using high-performance, low-cost semiconductors.

[1309] "Means for prior learning of patent information" refers to a means for studying the vast number of publicly available semiconductor circuit design drawings and patent documents to acquire knowledge for avoiding patent infringement.

[1310] "Means for automatically generating semiconductor circuit design drawings to avoid patent infringement" refers to means for automatically generating semiconductor circuit design drawings to avoid patent infringement using artificial intelligence.

[1311] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to means for developing and manufacturing semiconductors that are high-performance yet can be manufactured at low cost based on automatically generated design drawings.

[1312] "Means for acquiring emotion recognition data" refers to a method for recognizing emotions from the facial expressions and voices of workers using sensors such as cameras and microphones, and acquiring that data.

[1313] The "means for controlling the operation of a semiconductor based on emotion recognition data" refers to a means for adjusting the clock frequency and processing priority of a semiconductor based on the acquired emotion recognition data.

[1314] The "means for optimizing the operation of a robot in accordance with the emotions of workers in a factory" refers to a means for optimizing the operation of a robot in a factory in accordance with the emotions of workers based on emotion recognition data.

[1315] A system for implementing this invention includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for acquiring emotion recognition data, means for controlling the operation of semiconductors based on the emotion recognition data, and means for optimizing the operation of robots according to the emotions of workers in a factory.

[1316] Program processing explanation

[1317] 1. Methods for prior learning of patent information:

[1318] The server collects a huge number of publicly available semiconductor circuit designs and patent documents, and uses machine learning algorithms (e.g., TensorFlow) to learn from these data, thereby acquiring knowledge to avoid patent infringement.

[1319] 2. A method for automatically generating semiconductor circuit designs that circumvent patents:

[1320] Based on the learned patent information, the server uses a generative AI model to automatically generate semiconductor circuit designs that avoid patent infringement. This generative AI model receives prompt text as input and outputs designs that avoid patent infringement.

[1321] 3. A means to develop and manufacture proprietary high-performance, low-cost semiconductors:

[1322] The server inputs the automatically generated design drawings into the semiconductor manufacturing equipment, which then produces high-performance, low-cost semiconductors based on the design drawings.

[1323] 4. How to obtain emotion recognition data:

[1324] Terminals in the factory (including sensors such as cameras and microphones) capture workers' facial expressions and voices in real time and generate emotional data using emotion recognition algorithms (e.g., OpenCV).

[1325] 5. Means for controlling semiconductor operation based on emotion recognition data:

[1326] Based on the acquired emotion recognition data, the server adjusts the clock frequency and processing priority of the semiconductor equipped with the emotion engine.

[1327] 6. How to optimize robot behavior based on the emotions of factory workers:

[1328] The server then uses emotion recognition data to optimize the behavior of robots in the factory. For example, if a worker is recognized as tired, the robot will adjust its behavior to automatically take over the task.

[1329] Specific examples

[1330] For example, if the emotion engine recognizes that a worker in a factory is tired, the robot will adjust its behavior to automatically take on more work to reduce the worker's burden. Also, if a worker is feeling stressed, the robot will temporarily slow down its behavior to allow the worker to relax.

[1331] Prompt Sentence Examples

[1332] Examples of prompts to be input to a generative AI model include:

[1333] "If the emotion engine recognizes that a worker in a factory is tired, how does the robot adjust its behavior?"

[1334] Using this prompt, the generative AI model can suggest specific ways to adjust behavior.

[1335] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1336] Step 1:

[1337] The server collects a huge number of publicly available semiconductor circuit designs and patent documents, and uses machine learning algorithms (e.g., TensorFlow) to learn from these data. The input is publicly available semiconductor circuit designs and patent documents, and the output is knowledge for avoiding patent infringement. This knowledge is used to generate designs for avoiding patent infringement.

[1338] Step 2:

[1339] Based on the learned patent information, the server uses a generative AI model to automatically generate semiconductor circuit design drawings that circumvent patents. The input is knowledge for patent circumvention and a prompt, and the output is a semiconductor circuit design drawing that circumvents patents. Specifically, the prompt is input, and the generative AI model outputs the design drawing.

[1340] Step 3:

[1341] The server inputs the automatically generated blueprints into semiconductor manufacturing equipment, which then manufactures high-performance, low-cost semiconductors based on the blueprints. The input is the automatically generated blueprints, and the output is the manufactured semiconductors. Specifically, the server sends the blueprints to the manufacturing equipment and starts the manufacturing process.

[1342] Step 4:

[1343] The terminal (including sensors such as cameras and microphones) captures the facial expressions and voices of factory workers in real time and generates emotion data using an emotion recognition algorithm (e.g., OpenCV). The input is the facial expressions and voices of the workers, and the output is emotion data. Specifically, the sensors collect data and the emotion recognition algorithm analyzes the emotions.

[1344] Step 5:

[1345] The server adjusts the clock frequency and processing priority of the semiconductor equipped with the emotion engine based on the acquired emotion recognition data. The input is emotion recognition data, and the output is the adjusted semiconductor operating parameters. Specifically, it analyzes the emotion data and dynamically changes the semiconductor operating parameters.

[1346] Step 6:

[1347] The server optimizes the robots' behavior in the factory based on the emotion recognition data. The input is the emotion recognition data, and the output is the optimized robot's behavior instructions. Specifically, the emotion data is analyzed and the robot's behavior is adjusted. For example, if a worker is recognized as tired, the robot adjusts its behavior so that it automatically takes over the work.

[1348] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1349] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1350] Another example of generative AI is Gemini (internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1351] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1352] [Third embodiment]

[1353] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1354] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1355] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1356] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1357] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1358] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1359] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1360] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1361] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1362] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1363] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1364] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1365] "Example 1"

[1366] In one embodiment of the present invention, a means for pre-learning patent information is to use an AI that learns from semiconductor circuit design drawings obtained from a patent database. This AI extracts and learns the features of semiconductor circuit design drawings published in patents.

[1367] "Example 2"

[1368] Furthermore, as a means of automatically generating semiconductor circuit designs that avoid patents, the system has the ability to automatically generate new semiconductor circuit designs that do not infringe patents based on the patent information it has learned. Specifically, the AI ​​generates new designs based on patent information and checks whether the designs infringe existing patents.

[1369] "Example 3"

[1370] In addition, as a means of developing and manufacturing unique high-performance, low-cost semiconductors, semiconductors are manufactured based on automatically generated semiconductor circuit design drawings. Specifically, the automatically generated design drawings are input into semiconductor manufacturing equipment, which then manufactures semiconductors based on the design drawings. This makes it possible to build and operate high-performance AI without using expensive semiconductors.

[1371] The processing flow of each embodiment will be described below.

[1372] "Example 1"

[1373] Step 1: Obtain patent information for semiconductor circuit design drawings from a patent database.

[1374] Step 2: The acquired patent information is input into the AI, which then learns from the patent information. Through this learning, the AI ​​extracts the characteristics of the semiconductor circuit design.

[1375] "Example 2"

[1376] Step 1: Based on the learned patent information, the artificial intelligence automatically generates new semiconductor circuit design drawings.

[1377] Step 2: Check whether the automatically generated semiconductor circuit design diagram infringes on any existing patents. If it is confirmed that there is no infringement, the design diagram is adopted.

[1378] "Example 3"

[1379] Step 1: The selected semiconductor circuit design is input into the semiconductor manufacturing equipment.

[1380] Step 2: Semiconductor manufacturing equipment produces semiconductors based on the blueprint. This manufacturing process makes it possible to build and operate high-performance AI without using expensive semiconductors.

[1381] Example 1

[1382] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1383] In conventional semiconductor circuit design, searching and analyzing patent information is often done manually, which is time-consuming and labor-intensive. It is also difficult to automatically generate design drawings that avoid patents, making it difficult to efficiently develop and manufacture high-performance, low-cost semiconductors. To solve these issues, a system that can efficiently learn patent information and automatically generate design drawings that avoid patents is needed.

[1384] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1385] In this invention, the server includes means for pre-learning patent information, means for acquiring semiconductor circuit design drawings from a patent database, means for extracting the acquired design drawings as image data, means for inputting the image data into an artificial intelligence model to extract and learn features of the design drawings, means for receiving prompts from a user and searching for related patent information, and means for providing the search results to the user. This enables efficient learning and analysis of patent information, and realizes the automatic generation of design drawings that avoid patents.

[1386] "Patent information" refers to the technical information and invention content described in patent documents.

[1387] "Pre-training" refers to the process of training a model with relevant data before it performs a specific task.

[1388] "Patent database" refers to an information system that collects, stores, and provides patent documents in a searchable format.

[1389] A "semiconductor circuit design drawing" refers to a drawing that shows the circuit structure and layout of a semiconductor device.

[1390] "Image data" refers to data that represents visual information in digital form.

[1391] An "artificial intelligence model" refers to an algorithm or system built using techniques such as machine learning and deep learning to perform a specific task.

[1392] "Feature extraction" refers to the process of identifying important patterns or attributes from data and converting them into a numerical representation.

[1393] "Prompt sentence" refers to text containing instructions or questions that a user enters into a system.

[1394] "Search Results" refers to the collection of relevant information returned by the system based on a user's query.

[1395] The present invention relates to a system for efficiently learning patent information and automatically generating semiconductor circuit design diagrams that avoid patents. Specific embodiments of this system will be described below.

[1396] The server generates a program for pre-learning patent information. This program retrieves semiconductor circuit design drawings from a patent database and trains the AI ​​model on them. Patent databases used include Google Patents and Espacenet.

[1397] The server accesses a patent database to retrieve patent information, including semiconductor circuit design drawings. Since the retrieved patent documents are usually provided in PDF format, the server uses a PDF analysis library to extract the design drawings as image data.

[1398] The server then preprocesses the extracted image data and inputs it into an artificial intelligence model, which includes image resizing and noise reduction. The preprocessed image data is then input into an artificial intelligence model built using a deep learning framework such as TensorFlow or PyTorch.

[1399] The AI ​​model uses a convolutional neural network (CNN) to extract and learn from the features of the blueprint. Specifically, it recognizes the patterns and circuit structures of the blueprint and saves them as feature vectors. The server stores these feature vectors in a database for later use in searches.

[1400] The user inputs a prompt sentence using a terminal, such as "Please tell me patent information on the latest semiconductor circuit designs." The server analyzes this prompt sentence and extracts related keywords and feature vectors.

[1401] The server searches the database for relevant patent information based on the extracted keywords and feature vectors. The search results are provided to the user via their terminal. Specifically, a list of related patent documents and thumbnail images of blueprints are displayed.

[1402] Examples:

[1403] If a user wants to search for patent information on new semiconductor circuit designs, they can input the following prompt into the generative AI model:

[1404] Example prompt sentence:

[1405] "Please tell me the patent information on the latest semiconductor circuit designs."

[1406] When this prompt is input into the generative AI model, the server retrieves the relevant semiconductor circuit design from the patent database, and the AI ​​model analyzes it to extract features, ultimately providing the user with relevant patent information.

[1407] The above is a specific implementation method of the system according to one embodiment of the present invention.

[1408] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1409] Step 1:

[1410] The server accesses a patent database and obtains patent information including semiconductor circuit design drawings.

[1411] Input: API endpoint of patent database, search query (e.g., keywords or classification code)

[1412] Data processing: Send API requests and retrieve patent documents

[1413] Output: Acquired patent document (PDF format, etc.)

[1414] Specific operation: The server uses the patent database API to search for patent documents based on the specified keywords and classification codes, and downloads the relevant patent documents.

[1415] Step 2:

[1416] The server extracts semiconductor circuit design drawings as image data from the acquired patent documents.

[1417] Input: Acquired patent document (PDF format, etc.)

[1418] Data processing: Extracting blueprints as image data using a PDF parsing library

[1419] Output: Extracted image data

[1420] Specific operation: The server uses a PDF parsing library to extract semiconductor circuit design diagrams from patent documents in image format.

[1421] Step 3:

[1422] The server pre-processes the extracted image data and inputs it into an artificial intelligence model.

[1423] Input: Extracted image data

[1424] Data processing: Perform preprocessing such as image resizing and noise removal

[1425] Output: Preprocessed image data

[1426] Specific operation: The server uses an image processing library (e.g., OpenCV) to resize and denoise the image data and convert it into a format suitable for the artificial intelligence model.

[1427] Step 4:

[1428] The server inputs the preprocessed image data into an artificial intelligence model, which extracts and learns the features of the blueprint.

[1429] Input: Preprocessed image data

[1430] Data processing: Feature extraction using artificial intelligence models (e.g., TensorFlow or PyTorch)

[1431] Output: Feature vector

[1432] Specific operation: The server uses a convolutional neural network (CNN) to extract the patterns and structures of the blueprints from the image data as feature vectors and stores them in a database.

[1433] Step 5:

[1434] The user uses the terminal to enter a prompt sentence.

[1435] Input: Prompt statement (e.g., "Please tell me patent information on the latest semiconductor circuit designs.")

[1436] Data processing: Analyze the prompt sentence and extract relevant keywords and feature vectors

[1437] Output: Parsed keywords and feature vectors

[1438] Specific operation: The user inputs a prompt sentence through the terminal interface and sends it to the server.

[1439] Step 6:

[1440] The server searches the database for relevant patent information based on the analyzed keywords and feature vectors.

[1441] Input: Parsed keywords and feature vectors

[1442] Data processing: generating database queries and performing searches

[1443] Output: Search results (list of relevant patent documents and thumbnail images of blueprints)

[1444] Specific operations: The server generates a database query to search for relevant patent information based on the stored feature vectors.

[1445] Step 7:

[1446] The server provides the search results to the user.

[1447] Input: Search results (list of relevant patent documents and thumbnail images of blueprints)

[1448] Data processing: converting search results into a format that can be displayed to the user

[1449] Output: Search results served to the user

[1450] Specific operation: The server sends the search results to the user's terminal and displays them in a format that the user can view.

[1451] (Application example 1)

[1452] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1453] In conventional semiconductor circuit design, it was difficult to utilize patent information to select optimal designs and adjust manufacturing processes in real time. Furthermore, there was a lack of efficient means to develop and manufacture high-performance, low-cost electronic components while avoiding patents. This led to problems such as insufficient optimization of manufacturing processes and quality control, resulting in a decline in product competitiveness.

[1454] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1455] In this invention, the server includes a means for pre-learning patent information, a means for automatically generating patent-avoiding electronic component design drawings, a means for developing and manufacturing unique high-performance, low-cost electronic components, and a means for selecting an optimal design based on patent information and adjusting the manufacturing process in real time. This makes it possible to utilize patent information to select an optimal design and adjust the manufacturing process in real time. Furthermore, it is possible to efficiently develop and manufacture high-performance, low-cost electronic components while avoiding patent infringement, thereby optimizing the manufacturing process and improving quality control, thereby increasing the competitiveness of the product.

[1456] "Patent information" refers to data about patents that is made public by public institutions such as the Patent Office.

[1457] "Pre-training" is the process of training an artificial intelligence model with relevant data before it performs a specific task.

[1458] An "electronic component design drawing" is a drawing or diagram that shows the structure and function of an electronic component.

[1459] "Auto-generation" is the process of using artificial intelligence or algorithms to generate data and blueprints without manual intervention.

[1460] "High performance and low cost" refers to high performance and low manufacturing and operating costs.

[1461] "Development and manufacturing" is the process of designing new products and technologies and actually producing them.

[1462] An "optimal design" is the most efficient and effective design under specific conditions and constraints.

[1463] The "manufacturing process" is the series of steps that go from raw materials to creating a final product.

[1464] "Real-time" refers to immediate processing or response without delay.

[1465] "Adjustment" refers to changing or modifying a system or process to suit specific purposes or conditions.

[1466] The following system configuration and processing procedure will be described as an embodiment of the present invention.

[1467] The server includes a means for learning patent information in advance, a means for automatically generating electronic component design drawings that avoid patents, a means for developing and manufacturing unique high-performance, low-cost electronic components, and a means for selecting the optimal design based on patent information and adjusting the manufacturing process in real time.

[1468] System Configuration

[1469] 1. Hardware:

[1470] Server: A server equipped with a high-performance CPU and GPU.

[1471] Camera: High-resolution camera mounted on factory robot.

[1472] Factory robots: Robots used to carry out manufacturing processes.

[1473] 2. Software:

[1474] TensorFlow: Used for training and inference of AI models.

[1475] OpenCV: Used for image processing.

[1476] Database: A database for storing patent information.

[1477] Processing steps

[1478] 1. Preliminary study of patent information:

[1479] The server uses TensorFlow to pre-train the electronic component design drawings retrieved from the patent database. During this process, the AI ​​model extracts and learns the features of the design drawings.

[1480] 2. Automatic generation of electronic component design drawings:

[1481] The server uses a trained AI model to automatically generate patent-evading electronic component designs, which are optimized based on patent information.

[1482] 3. Selecting the best design and adjusting the manufacturing process:

[1483] Images captured by cameras mounted on factory robots are processed using OpenCV and sent to a server, which uses AI models to select the optimal design and adjust the manufacturing process in real time.

[1484] Specific examples

[1485] For example, an AI model can be created that selects the optimal circuit design based on electronic component blueprints retrieved from patent databases. This model can then be installed on factory robots and used to adjust the manufacturing process in real time.

[1486] Example prompt sentence:

[1487] "Create an AI model that selects the optimal circuit design based on electronic component blueprints retrieved from patent databases. This model can be installed on factory robots and adjust the manufacturing process in real time."

[1488] In this way, patent information can be used to select optimal designs and adjust manufacturing processes in real time, enabling efficient development and manufacturing of high-performance, low-cost electronic components while avoiding patents, improving manufacturing process optimization and quality control, and enhancing product competitiveness.

[1489] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1490] Step 1:

[1491] The server retrieves electronic component blueprints from a patent database. As input, it requires a query of the patent database, and as output, it obtains the retrieved blueprint data, which is used in the subsequent pre-training process.

[1492] Step 2:

[1493] The server pre-trains the acquired electronic component blueprints using TensorFlow. Blueprint data is required as input, and a trained AI model is obtained as output. In this process, the features of the blueprints are extracted and trained into the AI ​​model.

[1494] Step 3:

[1495] The server uses a trained AI model to automatically generate patent-avoidance electronic component design drawings. Patent information and the trained model are required as input, and an optimized design drawing is obtained as output. This design drawing is optimized based on the patent information.

[1496] Step 4:

[1497] The factory robots are equipped with cameras that capture images of the manufacturing process. As input, real-time images of the manufacturing process are required, and as output, captured image data is obtained. This image data is used for subsequent processing.

[1498] Step 5:

[1499] The server processes the captured image data using OpenCV and inputs it into the AI ​​model. Image data is required as input, and image features are extracted as output. This feature data is used to select the optimal design.

[1500] Step 6:

[1501] The server uses an AI model to select the optimal design and adjust the manufacturing process in real time. Image feature data and a trained model are required as input, and the output is adjusted manufacturing instructions, which are then sent to factory robots to optimize the manufacturing process.

[1502] Step 7:

[1503] The factory robot adjusts the manufacturing process based on instructions from the server and produces products according to the optimal design. Instructions from the server are required as input, and an optimized product is obtained as output. This process achieves optimization and quality control of the manufacturing process.

[1504] Example 2

[1505] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1506] In conventional semiconductor circuit design, it has been difficult to generate new design drawings while avoiding patent infringement. There is a need for a method that can properly analyze patent information and automatically generate design drawings that avoid patent infringement.

[1507] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1508] In this invention, the server includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for generating new design drawings using a generative AI model, means for confirming whether the generated design drawings avoid patent infringement, and means for regenerating the design drawings if patent infringement is confirmed. This makes it possible to automatically generate new semiconductor circuit design drawings that avoid patent infringement based on patent information.

[1509] "Patent information" is detailed data about existing patents, including claims, invention details, drawings, etc.

[1510] "Pre-training" is the process of training a model with a specific dataset to improve its predictive accuracy for a specific task.

[1511] A "semiconductor circuit design diagram" is a drawing that shows the structure and operation of a semiconductor device, and describes in detail the layout and connections of the circuit.

[1512] "Auto-generation" is the process of using artificial intelligence and algorithms to generate new designs and data without human intervention.

[1513] A "generative AI model" is a type of artificial intelligence that has the ability to generate new data and designs for specific tasks.

[1514] A "prompt" is text that is input to a generative AI model and contains instructions or conditions for the model to perform a specific task.

[1515] "Patent infringement" is the act of using a product or technology without permission within the scope of an existing patent, which can lead to legal issues.

[1516] "Regeneration" refers to running the generation process again when a design or data that has already been generated does not meet certain conditions.

[1517] This invention relates to a system for automatically generating new semiconductor circuit design drawings that avoid patent infringement based on patent information. Specific embodiments of this system will be described below.

[1518] Hardware and software used

[1519] Hardware:

[1520] High-performance servers (e.g., servers equipped with GPUs)

[1521] software:

[1522] Deep learning frameworks (e.g. TensorFlow, PyTorch)

[1523] Generative AI models (e.g., GPT-4)

[1524] System Operation

[1525] The server first accesses a patent information database to retrieve existing patent information, which contains detailed data on semiconductor circuit designs. Based on the retrieved patent information, the server then generates a prompt to be input into the generative AI model.

[1526] Example prompt sentence:

[1527] "Please generate a new semiconductor circuit design that does not infringe patent number 123456."

[1528] The server inputs this prompt into a generative AI model to generate a new semiconductor circuit design. The generated design is then compared with patent information to check for patent infringement. This check is performed using a patent text analysis algorithm. Specifically, it checks whether each element of the design is included in the scope of the patent claims.

[1529] If a patent infringement is confirmed, the server inputs the prompt sentence into the generative AI model again to generate a new design. This process is repeated until a patent-free design is generated.

[1530] Specific examples

[1531] For example, to generate a design drawing that does not infringe patent number 123456, the server inputs the following prompt statement into the generative AI model:

[1532] Example prompt sentence:

[1533] "Please generate a new semiconductor circuit design that does not infringe patent number 123456."

[1534] The generative AI model generates a new design based on this prompt. The generated design is compared with patent information to check for patent infringement. If an infringement is confirmed, the server inputs the prompt into the generative AI model again to generate a new design. This process is repeated until a design that does not infringe on a patent is generated.

[1535] The above is a specific embodiment for carrying out the present invention. This system makes it possible to automatically generate new semiconductor circuit design diagrams that avoid patent infringement based on patent information.

[1536] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1537] Step 1:

[1538] Obtaining patent information

[1539] The server accesses a patent information database to retrieve existing patent information. Patent numbers and specific keywords are given as input. The server searches the database based on this and retrieves relevant patent information. The output is patent information including claims, invention details, drawings, etc. Specifically, the server executes SQL queries to extract the required information from the database.

[1540] Step 2:

[1541] Generate prompt statement

[1542] Based on the acquired patent information, the server generates a prompt sentence to be input into the generative AI model. Patent information is given as input. The server analyzes the patent information and generates a prompt sentence that includes conditions for avoiding patent infringement. The output is a prompt sentence to be input into the generative AI model. Specifically, the server extracts the key points of the patent information and constructs a prompt sentence based on them.

[1543] Step 3:

[1544] Generating new blueprints

[1545] The server inputs a prompt statement into the generative AI model to generate a new semiconductor circuit design. The prompt statement is given as input. The generative AI model analyzes the prompt statement and generates a new design. The output is a new semiconductor circuit design. Specifically, the generative AI model generates the design using a deep learning algorithm.

[1546] Step 4:

[1547] Confirmation of patent infringement

[1548] The server compares the generated design with patent information to check for patent infringement. A new design and patent information are given as input. The server uses a patent text analysis algorithm to check whether each element of the design is included in the scope of the patent claims. The output is a result of whether there is patent infringement. Specifically, the server compares the design with patent information and evaluates the risk of patent infringement.

[1549] Step 5:

[1550] Regenerate blueprints (if necessary)

[1551] If patent infringement is confirmed, the server inputs the prompt sentence into the generative AI model again to generate a new design. The patent infringement result and the prompt sentence are given as input. The server adjusts the prompt sentence and inputs it into the generative AI model again. The output is a new semiconductor circuit design. Specifically, the server modifies the prompt sentence and runs the generation process again. This process is repeated until a design without patent infringement is generated.

[1552] (Application example 2)

[1553] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1554] In conventional semiconductor circuit design, it was extremely difficult to design in a way that avoided patent infringement, and the manual task of checking patent information required a great deal of effort. Furthermore, there was a lack of technology to automatically generate design drawings that avoided patent infringement, which placed a heavy burden on designers. Furthermore, there were insufficient means to check whether the generated design drawings infringed on existing patents, which increased the risk of patent infringement.

[1555] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for verifying whether the generated design drawings infringe existing patents, and means for exporting the design drawings. This enables the automatic generation of new semiconductor circuit design drawings based on patent information, enabling efficient design to avoid patent infringement. Furthermore, it is possible to automatically verify whether the generated design drawings infringe existing patents, thereby reducing the risk of patent infringement.

[1556] "Patent information" refers to information about published patent documents and patent applications, which indicates the technical content and scope of rights of the patents.

[1557] "Pre-training" refers to the process by which an artificial intelligence or machine learning model is trained in advance using a specific dataset.

[1558] A "semiconductor circuit blueprint" is a drawing or blueprint that shows the structure and operation of a semiconductor device, and plays an important role in the manufacturing process.

[1559] "Auto-generation" refers to the process of using artificial intelligence and algorithms to automatically generate blueprints and data without human intervention.

[1560] "Verification" refers to the process of verifying whether the generated design drawings infringe on existing patents.

[1561] "Export" refers to the process of outputting the generated design drawings in a format such as PDF or CAD file.

[1562] "Artificial intelligence" refers to systems and programs that mimic human intelligence using technologies such as machine learning and deep learning.

[1563] A system for implementing this invention includes means for pre-learning patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for verifying whether the generated design drawings infringe existing patents, and means for exporting the design drawings.

[1564] System configuration

[1565] The system consists of the following main components:

[1566] 1. Server: Pre-learns patent information and performs processing to generate and verify design drawings.

[1567] 2. Terminal: Provides an interface for users to input design requirements and review and export the generated design drawings.

[1568] 3. Database: Storage for patent information and generated design drawings.

[1569] Hardware and software used

[1570] Hardware: High-performance servers, user devices (smartphones, tablets, PCs, etc.)

[1571] software:

[1572] Generative AI model: Uses OpenAI's API to generate blueprints based on patent information.

[1573] Patent Database API: An API for obtaining patent information.

[1574] Design Tool API: API for exporting generated design drawings in PDF and CAD formats.

[1575] Data processing and calculation

[1576] The server uses a patent database API to obtain the latest patent information and pre-trains the generative AI model. When the user inputs design requirements from their device, the server uses the generative AI model to generate a semiconductor circuit design that avoids patents. The generated design is compared with the patent database to check for patent infringement. Once the check is complete, the design is exported in PDF or CAD format using the design tool API.

[1577] Specific examples

[1578] For example, if a user wants to design a "low-power, high-performance semiconductor circuit," the following prompt sentence is used:

[1579] Prompt Sentence Examples

[1580] Patent information: [Patent information details]

[1581] User requirement: low-power, high-performance semiconductor circuit

[1582] Generate patent-free semiconductor circuit designs.

[1583] Based on this prompt, the server uses a generative AI model to generate a new semiconductor circuit design that bypasses the patent, checks for patent infringement, and then exports the design, allowing the user to efficiently obtain a design that bypasses the patent.

[1584] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1585] Step 1:

[1586] The server uses a patent database API to retrieve the latest patent information. The input is patent information from the patent database API, and the output is a dataset of the retrieved patent information. This dataset contains detailed information about patent documents and patent applications.

[1587] Step 2:

[1588] The server pre-trains the generative AI model using the acquired patent information. The input is a dataset of patent information, and the output is a generative AI model that has learned from the patent information. This process involves data processing based on the patent information and model training.

[1589] Step 3:

[1590] A user inputs design requirements using a terminal. The input is the user's design requirements (e.g., "a low-power, high-performance semiconductor circuit"), and the output is the transmission of the design requirements to the server. The terminal provides an interface that transmits the requirements entered by the user to the server.

[1591] Step 4:

[1592] The server receives the user's design requirements and uses a generative AI model to generate a semiconductor circuit design that avoids patents. The input is the user's design requirements and a pre-trained generative AI model, and the output is the generated semiconductor circuit design. In this process, the generative AI model generates a design based on a prompt.

[1593] Step 5:

[1594] The server checks whether the generated design drawings infringe on existing patents. The input is a dataset of the generated design drawings and patent information, and the output is a result of whether or not there is patent infringement. In this process, the generated design drawings are matched with patent information to assess the risk of patent infringement.

[1595] Step 6:

[1596] The server exports the verified design drawings using the design tool API. The input is the verified design drawings, and the output is a PDF or CAD file. This process outputs the design drawings in a format that can be used by the user.

[1597] Step 7:

[1598] The user downloads and checks the exported blueprints using a terminal. The input is a blueprint file in PDF or CAD format, and the output is the blueprint saved on the user's terminal. The terminal provides an interface for the user to check and use the blueprints.

[1599] Example 3

[1600] Next, a third embodiment of the third embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1601] Conventional semiconductor manufacturing processes have the problem of high costs for producing high-performance semiconductors. It is also difficult to design new semiconductor circuits while avoiding patent infringement. Furthermore, the process from creating design drawings to manufacturing is complex, requiring efficient operation.

[1602] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1603] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for transmitting the generated semiconductor circuit design drawings to a manufacturing device, means for the manufacturing device to manufacture semiconductors based on the design drawings, and means for developing and manufacturing high-performance, low-cost semiconductors. This makes it possible to efficiently manufacture high-performance, low-cost semiconductors and design new semiconductor circuits while avoiding patent infringement.

[1604] "Patent information" refers to technical information contained in publicly available patent documents and patent databases.

[1605] "Pre-training" refers to the process of training a model with relevant data before it performs a specific task.

[1606] A "semiconductor circuit design diagram" is a drawing that shows the circuit configuration and layout of a semiconductor device.

[1607] "Automated generation" refers to the process of using artificial intelligence and algorithms to generate blueprints and data without human intervention.

[1608] "Manufacturing equipment" refers to machines and facilities used to manufacture semiconductor devices.

[1609] A "communication protocol" refers to the rules and procedures for data communication.

[1610] "Lithography technology" is a technology that uses light to form fine patterns on silicon wafers.

[1611] "High performance and low cost" means high performance and low manufacturing and operating costs.

[1612] "Artificial intelligence" refers to the technology that enables computer systems to learn and reason by imitating human intelligence.

[1613] MODE FOR CARRYING OUT THE INVENTION

[1614] This invention relates to a system that learns patent information in advance, automatically generates semiconductor circuit design drawings that avoid patents, and manufactures high-performance, low-cost semiconductors based on those design drawings. Specific embodiments of this system are described below.

[1615] 1. System Configuration

[1616] This system consists of three main components: a server, a terminal, and a user. The server includes a means for pre-learning patent information, a means for automatically generating semiconductor circuit design drawings that circumvent patents, and a means for transmitting the generated design drawings to manufacturing equipment. The terminal provides an interface for users to input prompt sentences. The user is responsible for operating the system and inputting the necessary prompt sentences.

[1617] 2. Hardware and Software Used

[1618] The server is a computer system equipped with a high-performance processor and large memory capacity. Specifically, it is equipped with a GPU (graphics processing unit) for running the generative AI model. The generative AI model is implemented using machine learning frameworks such as TensorFlow and PyTorch. Publicly available patent databases are used for pre-training patent information. The generated design drawings are sent to the manufacturing equipment using secure communication protocols (e.g., HTTPS or SFTP).

[1619] Manufacturing equipment is used to manufacture semiconductors using lithography technology. Specifically, EUV (extreme ultraviolet) lithography equipment is used. This equipment is used to form fine circuit patterns on silicon wafers.

[1620] 3. Program Processing

[1621] The server inputs the prompt received from the user into a generative AI model to generate an optimal semiconductor circuit design. The generated design is then sent from the server to the manufacturing equipment, which then manufactures semiconductors based on the design.

[1622] 4. Specific Examples

[1623] If a user wants to build a high-performance AI system, they input a prompt sentence like the following into the generative AI model:

[1624] Example prompt sentence:

[1625] "Design a high-performance, low-cost AI processor."

[1626] When this prompt is entered, the server generates an optimal semiconductor circuit design using a generative AI model. The generated design is then sent to the manufacturing equipment using a secure communication protocol, and the manufacturing equipment manufactures semiconductors based on the design. This enables users to build and operate high-performance AI systems without using expensive semiconductors. The flow of the specific processing in Example 3 will be explained using Figure 15.

[1627] Program processing flow

[1628] Step 1:

[1629] The user enters a prompt statement.

[1630] Input: The user accesses the system's interface and enters a prompt statement into a text box.

[1631] Specific operation: The user enters, for example, "Please design a high-performance, low-cost AI processor" and clicks the submit button.

[1632] Output: The prompt text is sent to the server.

[1633] Step 2:

[1634] The server receives the prompt sentence and inputs it into the generative AI model.

[1635] Input: The prompt the server received from the user.

[1636] Specific operation: The server sends the prompt sentence to the generative AI model as an API request.

[1637] Data processing: Convert the prompt sentence into an appropriate format and input it into the generative AI model.

[1638] Output: The generative AI model generates an optimal semiconductor circuit design and sends it back to the server.

[1639] Step 3:

[1640] The server transmits the generated semiconductor circuit design drawing to the manufacturing device.

[1641] Input: Semiconductor circuit design diagrams returned from a generative AI model.

[1642] Specific operation: The server encrypts the generated design drawing and sends it to the manufacturing device using a secure communication protocol (e.g., HTTPS or SFTP).

[1643] Data processing: Encrypt the blueprint data and convert it into a format suitable for the communication protocol.

[1644] Output: Encrypted blueprint sent to fabricator.

[1645] Step 4:

[1646] Manufacturing equipment produces semiconductors based on blueprints.

[1647] Input: Encrypted semiconductor circuit blueprints sent from the server.

[1648] How it works: The manufacturing equipment receives the design drawings, first decrypts them, then analyzes them, prepares silicon wafers, and forms circuits on them using lithography technology.

[1649] Data processing: Analysis of design data and circuit formation using lithography technology.

[1650] Output: Manufactured semiconductor device.

[1651] Step 5:

[1652] Users receive the manufactured semiconductors and build and operate AI systems.

[1653] Input: Manufactured semiconductor devices.

[1654] What happens next: The user receives the manufactured semiconductors and incorporates them into an AI system, for example by attaching the AI ​​processor to a motherboard and assembling the entire system. Then, the user installs the AI ​​software and begins training and inference of machine learning models.

[1655] Data processing: Embedding semiconductor devices and installing AI software.

[1656] Output: A high-performance AI system is built and operational.

[1657] (Application example 3)

[1658] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1659] Conventional semiconductor manufacturing processes require expensive materials and complex processes to produce high-performance semiconductors, resulting in high costs. Furthermore, there was a lack of means to efficiently manufacture semiconductors while avoiding patent infringement. Furthermore, there was also the issue of insufficient optimization of the manufacturing process, resulting in low manufacturing efficiency.

[1660] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1661] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that avoid patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for optimizing the manufacturing process based on the automatically generated semiconductor circuit design drawings, and means for manufacturing semiconductors according to the optimized manufacturing process, thereby enabling the efficient manufacturing of high-performance, low-cost semiconductors.

[1662] "Means for prior learning of patent information" refers to a means for studying the vast number of publicly available semiconductor circuit design drawings and patent documents, and accumulating knowledge to avoid patent infringement.

[1663] The "means for automatically generating semiconductor circuit design drawings that avoid patents" refers to a means for automatically generating new semiconductor circuit design drawings that avoid patent infringement using artificial intelligence.

[1664] "Means for developing and manufacturing unique high-performance, low-cost semiconductors" refers to the development and manufacturing of high-performance semiconductors at low manufacturing costs based on semiconductor circuit design drawings that avoid patents.

[1665] "Means for optimizing the manufacturing process based on automatically generated semiconductor circuit design drawings" refers to means for optimizing each step of the manufacturing process based on automatically generated semiconductor circuit design drawings, and for efficiently manufacturing semiconductors.

[1666] The "means for manufacturing a semiconductor according to an optimized manufacturing process" refers to a means for actually manufacturing a semiconductor according to an optimized manufacturing process.

[1667] A system for implementing this invention includes a means for prior learning of patent information, a means for automatically generating semiconductor circuit design drawings that circumvent patents, a means for developing and manufacturing unique high-performance, low-cost semiconductors, a means for optimizing the manufacturing process based on the automatically generated semiconductor circuit design drawings, and a means for manufacturing semiconductors in accordance with the optimized manufacturing process.

[1668] 1. Program Generation

[1669] The server collects and studies a vast number of publicly available semiconductor circuit designs and patent documents to learn about patent information. It then uses artificial intelligence to automatically generate semiconductor circuit designs that avoid patents. Based on these designs, the server develops and manufactures unique, high-performance, low-cost semiconductors.

[1670] 2. Program processing explanation

[1671] The server uses the following hardware and software:

[1672] Hardware: Factory robots (e.g., general-purpose robotic arms), semiconductor manufacturing equipment (e.g., lithography equipment)

[1673] Software: Python 3.x, JSON library, artificial intelligence model (e.g. TensorFlow)

[1674] The server first collects patent information and stores it in JSON format. It then uses an artificial intelligence model to generate a semiconductor circuit design that avoids patents. Based on this design, it runs an algorithm to optimize the manufacturing process and determines the optimal manufacturing process. Finally, factory robots manufacture semiconductors according to this optimized process.

[1675] 3. Examples and prompts

[1676] Specific examples

[1677] For example, the patent information collected by the server includes existing semiconductor circuit blueprints and patent documents. Based on this information, the AI ​​model generates a new semiconductor circuit blueprint. Based on this blueprint, each step of the manufacturing process (e.g., layer deposition, photolithography, etching, doping) is optimized.

[1678] Prompt Sentence Examples

[1679] "You are asked to develop an application that will be installed on a factory robot. This application will optimize the semiconductor manufacturing process based on automatically generated semiconductor circuit design drawings, and efficiently manufacture high-performance, low-cost semiconductors."

[1680] Thus, specific embodiments for carrying out the invention have been shown.

[1681] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1682] Step 1:

[1683] The server collects a huge number of publicly available semiconductor circuit design drawings and patent documents. It stores this data in JSON format. The input is publicly available semiconductor circuit design drawings and patent documents, and the output is patent information data stored in JSON format. Specifically, the server collects data using web scraping and APIs and stores it in local storage.

[1684] Step 2:

[1685] The server pre-learns the patent information based on the collected patent information. The input is the patent information data saved in step 1, and the output is the trained artificial intelligence model. Specifically, it uses a machine learning library such as TensorFlow to train the model using patent information as learning data.

[1686] Step 3:

[1687] The server uses a trained AI model to automatically generate semiconductor circuit design drawings that circumvent patents. The input is the trained AI model and design requirements, and the output is an automatically generated semiconductor circuit design drawing. Specifically, the server inputs a prompt statement into the generating AI model to generate the design drawing.

[1688] Step 4:

[1689] The server optimizes the manufacturing process based on automatically generated semiconductor circuit design drawings. The input is the automatically generated semiconductor circuit design drawing, and the output is the optimized manufacturing process steps. Specifically, it uses algorithms to optimize each step of the manufacturing process (e.g., layer deposition, photolithography, etching, doping).

[1690] Step 5:

[1691] The server manufactures semiconductors according to an optimized manufacturing process. The inputs are the steps in the optimized manufacturing process, and the output is the manufactured high-performance, low-cost semiconductors. Specifically, it controls factory robots and operates manufacturing equipment to manufacture semiconductors.

[1692] In this manner, the specific operations performed in each processing step have been described.

[1693] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1694] "Example 1"

[1695] One embodiment of the present invention provides a system that includes a means for pre-learning patent information, a means for automatically generating semiconductor circuit design diagrams that circumvent patents, and a means for developing and manufacturing unique, high-performance, low-cost semiconductors, as well as an emotion engine that recognizes user emotions. This emotion engine recognizes emotions from the user's voice, facial expressions, text input, etc., and controls the operation of the semiconductor based on the recognized emotion. For example, if the user expresses anger, the emotion engine analyzes that information and adjusts the operation of the semiconductor. This enables the semiconductor to perform optimal operations in response to the user's emotions.

[1696] "Example 2"

[1697] Furthermore, as another embodiment of the present invention, a specific technique is provided for an emotion engine to recognize a user's emotion. Specifically, a voice analysis technique that recognizes an emotion from a user's voice, an image analysis technique that recognizes an emotion from a user's facial expression, a natural language processing technique that recognizes an emotion from a user's text input, etc. are used. By combining these techniques, emotion recognition with higher accuracy becomes possible.

[1698] "Example 3"

[1699] Furthermore, as another embodiment of the present invention, a specific technique for controlling the operation of a semiconductor based on the emotion recognized by an emotion engine is provided. Specifically, control such as adjusting the clock frequency of the semiconductor or giving priority to the execution of a specific process is performed according to the type and intensity of the emotion recognized by the emotion engine. This enables optimal control of the semiconductor operation according to the user's emotion.

[1700] The processing flow of each embodiment will be described below.

[1701] "Example 1"

[1702] Step 1: Receive input from the user (voice, facial expression, text, etc.).

[1703] Step 2: The emotion engine recognizes emotions from the user's input.

[1704] Step 3: Control the operation of the semiconductor based on the recognized emotion.

[1705] "Example 2"

[1706] Step 1: Perform voice analysis to recognize emotions from the user's voice.

[1707] Step 2: Perform image analysis to recognize emotions from the user's facial expressions.

[1708] Step 3: Perform natural language processing to recognize emotions from the user's text input.

[1709] "Example 3"

[1710] Step 1: Determine the type and intensity of the emotion recognized by the emotion engine.

[1711] Step 2: Adjust the clock frequency of the semiconductor based on the result of the determination.

[1712] Step 3: Based on the judgment result, a specific process is executed with priority.

[1713] Example 1

[1714] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1715] In conventional semiconductor circuit design, it was difficult to develop high-performance, low-cost semiconductors while avoiding patent infringement. Furthermore, there was no technology to optimize semiconductor operation according to user emotions, making it difficult to improve the user experience.

[1716] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1717] In this invention, the server includes means for prior learning of patent information, means for automatically generating semiconductor circuit design drawings that circumvent patents, means for developing and manufacturing unique high-performance, low-cost semiconductors, means for recognizing user emotions, and means for controlling the operation of the semiconductor based on the recognized emotions. This makes it possible to develop high-performance, low-cost semiconductors while avoiding patent infringement, and also enables the semiconductors to perform optimal operations according to the user's emotions.

[1718] "Patent information" refers to semiconductor circuit design drawings and related technical documents registered in patent databases.

[1719] "Pre-learning" refers to the process by which artificial intelligence extracts features based on patent information and learns from them.

[1720] A "semiconductor circuit design drawing" refers to a drawing or design data that shows the circuit structure of a semiconductor.

[1721] "Automatic generation" refers to the process by which artificial intelligence automatically creates new semiconductor circuit designs based on patent information.

[1722] "High performance and low cost" refers to high performance and low manufacturing costs.

[1723] "Emotion recognition" is a technology that analyzes and recognizes emotions from a user's voice, facial expressions, text input, etc.

[1724] "Behavioral control" refers to the process of adjusting the behavior of a semiconductor based on perceived emotions.

[1725] This invention is a system that learns patent information in advance, automatically generates semiconductor circuit designs that avoid patents, and controls the operation of semiconductors by recognizing the user's emotions. This system works in cooperation with a server, a terminal, and a user.

[1726] The server accesses a patent database to retrieve semiconductor circuit design drawings. This database contains various semiconductor circuit design drawings that have been published in patents. The server inputs the retrieved design drawings into an AI model using a machine learning framework such as TensorFlow or PyTorch to extract the design features. The AI ​​model learns based on the extracted features and pre-learns patent information.

[1727] The server then uses the trained AI model to generate new semiconductor circuit design drawings. In this process, the AI ​​takes existing patent information into account when generating new designs. The resulting designs reduce the risk of patent infringement while enabling high-performance, low-cost semiconductors.

[1728] The device inputs the user's voice, facial expressions, text input, etc. into the emotion engine. This emotion engine is built using OpenCV and Microsoft Azure's emotion recognition API. The emotion engine recognizes the user's emotions from the input data and analyzes that information.

[1729] The server controls the semiconductor's operation based on the user's emotional information obtained from the emotion engine. For example, if the user expresses anger, the server analyzes that information and adjusts the semiconductor's operation. This enables the semiconductor to perform optimal operations according to the user's emotions.

[1730] As a specific example, the server retrieves the semiconductor circuit design drawing for "Patent No. 123456" from a patent database and inputs it into an AI model for learning. Next, it uses the trained AI model to generate a new design drawing and confirms that it is designed to avoid patent No. 123456. The device inputs the user's voice input, "I'm very tired today," into the emotion engine, which recognizes that the user is feeling "tired." Based on this emotion information, the server switches the semiconductor's operation to energy-saving mode.

[1731] Examples of prompts include, "Build an AI model that learns from semiconductor circuit designs retrieved from a patent database and generates new designs that avoid patents. Also, describe a system that recognizes user emotions and controls the operation of semiconductors based on those emotions."

[1732] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1733] Step 1:

[1734] The server accesses the patent database and retrieves the semiconductor circuit design. Specifically, the server uses an API to download the design data for "Patent Number 123456." The input is the patent number, and the output is the design data. This design data is used to train the subsequent AI model.

[1735] Step 2:

[1736] The server inputs the acquired blueprint data into an AI model using machine learning frameworks such as TensorFlow and PyTorch. The AI ​​model extracts the features of the blueprint and performs learning. The input is the blueprint data, and the output is the trained model. In this process, image processing technology is used to analyze the blueprint and extract features.

[1737] Step 3:

[1738] The server uses the trained AI model to generate new semiconductor circuit designs. Specifically, the AI ​​model combines features from existing designs to generate new designs. The input is the trained model, and the output is the new design. This design is designed to avoid patent infringement.

[1739] Step 4:

[1740] The server checks the generated design against a patent database to determine whether it avoids patent infringement. The input is the new design, and the output is information about...

Claims

1. A means for generating semiconductor circuit design diagrams that avoid patents using a generation AI model that has previously learned patent information; A means for verifying whether the generated semiconductor circuit design avoids patent infringement; means for regenerating said semiconductor circuit design drawing if patent infringement is confirmed; A system including:

2. The generating means generates the semiconductor circuit design diagram based on a prompt sentence corresponding to the patent to be avoided. The system of claim 1 .

3. A means for obtaining a semiconductor circuit design drawing from a patent database; A means for extracting the acquired semiconductor circuit design drawing as image data; A means for inputting the extracted image data into the generating AI model and having the generating AI model learn features of the acquired semiconductor circuit design drawing, The system of claim 1 .

Images