System

Abstract

A system is provided.SOLUTION: A system, comprising: means for initializing a text generation model; means for reading a medical paper; means for passing text data of the read medical paper to a summary generation model to generate a summary; means for providing the generated summary to a user; means for classifying each part of the read medical paper by the text AI model using a plurality of verification points received from the user to obtain a classified result; and means for providing the obtained classified result to the user.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] The process of reviewing and verifying medical research is crucial for maintaining the integrity of scientific research and improving the quality of healthcare delivery. However, pharmacists and physicians currently lack the time and resources to quickly and efficiently review the vast amount of information. Students and patients also face challenges in accurately interpreting specialized knowledge. Therefore, there is a need for a more efficient and reliable method for verifying the content of medical research. [Means for solving the problem]

[0005] This invention proposes a system that uses a generative AI model to accurately grasp the content of medical papers and support the process of determining their validity. Specifically, the system includes a means for initializing the generative AI model, a means for reading medical papers, a means for passing the text data of the read medical paper to a summary generation model and generating a summary, a means for providing the generated summary to a user, a means for classifying each part of the read medical paper using a text classification model based on multiple verification points received from the user, and obtaining the classification results, and a means for providing the obtained classification results to the user. This system enables pharmacists, doctors, medical students, and patients seeking informed treatment to efficiently review papers and determine their validity.

[0006] A "generative AI model" is an artificial intelligence model that has been trained to perform natural language processing tasks using machine learning algorithms.

[0007] "Means for initialization" refers to the process by which the generative AI model is loaded and made available for use when the system starts.

[0008] A "medical paper" is an academic paper that describes the results and findings of scientific research related to medicine or pharmacology.

[0009] "Text data" refers to information in the form of text, and in this case refers to the contents of medical papers.

[0010] A "summary generation model" is a type of generative AI model that extracts important information from long texts and outputs them as short summaries.

[0011] A "summary" is a short sentence that succinctly summarizes the main points and conclusions of the original text.

[0012] "User" means any person or entity that wishes to use the system to verify the contents of medical treatises.

[0013] "Verification points" refer to specific viewpoints or criteria that are considered important in evaluating medical papers.

[0014] A "text classification model" is a generative AI model trained to classify input text into specific categories.

[0015] A "classification result" is the rating or category that a text classification model outputs based on the input text. [Brief explanation of the drawings]

[0016] [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. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12]FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0017] 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.

[0018] First, the terms used in the following description will be explained.

[0019] 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, a 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), and an APU (Accelerated Processing Unit).

[0020] 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.

[0021] 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.

[0022] 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.

[0023] 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."

[0024] [First embodiment]

[0025] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0026] 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.

[0027] 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).

[0028] 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.

[0029] 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.

[0030] 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.

[0031] 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.

[0032] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 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.

[0034] 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.

[0035] 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.

[0036] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. 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."

[0037] This invention is a system that uses a generative AI model to accurately grasp the content of medical papers and support the process of determining their validity. Specific embodiments of this system are described below.

[0038] The server initializes the generated AI model

[0039] First, the server initializes the generative AI model. This generative AI model includes a summary generation model for performing text summarization and a text classification model for classifying text. By initializing these models, the server prepares to process medical papers.

[0040] User uploads medical paper

[0041] Next, the user uploads a medical paper. The user inputs the file path of the medical paper and sends it to the server. This file path is used when the server reads the text of the medical paper.

[0042] The server loads medical papers

[0043] The server reads the medical paper uploaded by the user. The server opens the paper file and reads its contents as text data. This text data is then passed to the summary generation model.

[0044] Server generates summary

[0045] The server passes the text data of medical papers read into a summary generation model, which then generates a summary. This summary generation model extracts important information from long sentences and outputs a short summary. The generated summary is then provided to the user.

[0046] User registers verification point

[0047] Users register validation criteria for evaluating medical papers, including the validity of the research, statistical accuracy, and reproducibility of results, among other aspects that the user wishes to evaluate.

[0048] The server analyzes the paper

[0049] Based on the verification points received from the user, the server analyzes each part of the medical paper using a text classification model. The server uses each verification point and the text data of the paper to obtain a classification result. This classification model is trained to classify the input text into a specific category.

[0050] The server provides the classification results

[0051] The server provides the user with the classification results for each verification point, allowing the user to check the evaluation results for each aspect of the medical paper.

[0052] Specific examples

[0053] A specific example is given below.

[0054] 1. User uploads paper file

[0055] Example: A user uploads a medical article called "path / to / medical_article.txt" to the server.

[0056] 2. The server generates the summary

[0057] Example: The server reads the content of a paper and uses a summary generation model to generate the following summary:

[0058] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, showed the expected effects."

[0059] 3. User registers verification points

[0060] Example: A user registers verification points such as "research validity," "statistical accuracy," and "reproducibility of results."

[0061] 4. The server analyzes the paper

[0062] Example: The server runs a text classification model on each validation point and obtains the following classification results:

[0063] Study validity: "Positive"

[0064] Statistical accuracy: "Neutral"

[0065] Reproducibility of results: "Mixed"

[0066] In this way, by using this system, medical and pharmaceutical experts can efficiently and accurately evaluate and verify a huge amount of medical research papers.

[0067] The processing flow will be explained below.

[0068] Step 1:

[0069] The server initializes the generative AI model, specifically, loads the summary generation model and text classification model, and deploys them to storage devices to efficiently perform natural language processing tasks.

[0070] Step 2:

[0071] The user inputs the file path of the medical paper and sends it to the server. The file path is input via the interface of the terminal used by the user.

[0072] Step 3:

[0073] The server receives the file path and opens the specified file. The server reads the contents of the paper file as text data. This text data includes the full text of the medical paper.

[0074] Step 4:

[0075] The server passes the read text data to a summary generation model. The model extracts the main points of the text and generates a summary. The generated summary is short and concise, describing the main points of the paper.

[0076] Step 5:

[0077] The server provides the generated abstract to the user, who can then view the abstract through a terminal interface to quickly understand the content of the paper.

[0078] Step 6:

[0079] Users input the validation points used to evaluate papers and submit them to the server. These validation points include the validity of the research, statistical accuracy, and reproducibility of results.

[0080] Step 7:

[0081] The server uses the verification points received from the user and the text data of the medical paper to analyze the paper using a text classification model. For each verification point, the corresponding part is classified and a classification result is obtained.

[0082] Step 8:

[0083] The server provides the classification results to the user, who can then view the evaluation results for each verification point via the terminal interface, thereby confirming the analysis results for each aspect of the medical paper.

[0084] Step 9:

[0085] Users evaluate the validity and reliability of medical papers based on the provided abstracts and classification results. The server records this evaluation process and stores it in a database for future reference and analysis.

[0086] Example 1

[0087] Next, a description will be given of 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."

[0088] Efficiently evaluating and verifying large volumes of pharmaceutical research papers is a major challenge for medical and pharmaceutical experts. In particular, there is a need to quickly and accurately evaluate the validity, statistical accuracy, and reproducibility of results of research papers. However, conventional methods require a huge amount of time and effort to analyze each paper in detail, making them inefficient. Furthermore, results can vary depending on the evaluator's subjectivity, making it difficult to perform consistent evaluations.

[0089] 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.

[0090] In this invention, the server includes means for initializing the generative AI model, means for a user to upload a medical paper, means for reading the medical paper, means for passing the text data of the read medical paper to a summary generation model to generate a summary, means for providing the generated summary to the user, means for registering multiple verification points received from the user, means for classifying each part of the read medical paper using the multiple verification points received from the user with a text classification model to obtain the classification results, and means for providing the obtained classification results to the user, thereby enabling efficient summary generation of medical papers and automatic and consistent evaluation from multiple perspectives.

[0091] A "generative AI model" is a trained model that uses artificial intelligence to summarize and classify text.

[0092] The "server" is the computer system that serves as the core of the system, initializing the generative AI model, processing medical papers, and interfacing with users.

[0093] A "user" is a person or organization that uses the system to upload medical papers and receive their summaries and evaluation results.

[0094] A "medical paper" is a piece of writing that summarizes research results and findings related to medicine and pharmacy.

[0095] A "summary generation model" is a type of generative AI model used to extract important information from medical papers and generate short summaries.

[0096] A "text classification model" is a type of generative AI model used to classify parts of medical papers into specific categories.

[0097] "Verification points" are items that serve as viewpoints or criteria when a user evaluates a medical paper.

[0098] "Classification results" are evaluation results for each part of a medical paper generated by a text classification model.

[0099] "Upload" is the act of a user sending data from their own terminal to a server.

[0100] A "web interface" is a screen that uses a web browser to allow a user to access and operate the system.

[0101] This invention is a system that uses a generative AI model to efficiently summarize the contents of medical papers and assists in the process of determining their validity based on each verification point. Specific embodiments are described below.

[0102] 1. Hardware and Software Configuration

[0103] The server plays a central role in this system. The server hardware includes a powerful CPU, sufficient memory, and large storage capacity. The main software used includes Python, the Transformers library, and a suitable database management system.

[0104] 2. Initializing the Generative AI Model

[0105] When the system is initialized, the server loads generative AI models from disk. Specifically, it loads a "summary generation model" that generates summaries of medical papers and a "text classification model" that classifies text. These models have been trained in advance using large amounts of text data, and the server initializes their parameters and loads them into memory.

[0106] 3. User upload of medical papers

[0107] Users upload medical papers through the system's web interface. They use a file selection dialog to select the medical paper file they want to evaluate and click the upload button. The terminal sends the selected file path to the server, which stores it in a temporary directory.

[0108] 4. Reading medical papers

[0109] The server uses the file path sent by the user to open the uploaded medical paper, reads the file contents as text data, and stores it in memory. This text data is used in the subsequent summary generation process.

[0110] 5. Summary Generation

[0111] The server passes the text data of the loaded medical paper to a summary generation model to generate a summary. This summary generation model extracts important information from the medical paper and outputs a short summary. The generated summary is provided to users so that they can easily understand the outline of the paper.

[0112] 6. Registering Verification Points

[0113] Users register multiple validation points to evaluate the validity of a paper. These include aspects such as "research validity," "statistical accuracy," and "reproducibility of results." These validation points are entered on a web interface and sent to the server, which stores them in a database.

[0114] 7. Analysis of the paper

[0115] The server applies a text classification model using the verification points received from the user and the text data loaded, generating a classification result for each verification point and storing it in an internal data structure.

[0116] 8. Providing classification results

[0117] The server takes the generated classification results, converts them into a user-friendly format, and provides the results to the user through a web interface.

[0118] Specific examples

[0119] Initializing a generative AI model

[0120] Example: The server starts and loads a "summary generation model" and a "text classification model" using common natural language processing models.

[0121] User-uploaded medical papers

[0122] Example: User uploads "path / to / medical_article.txt" through the web interface.

[0123] Reading medical papers

[0124] Example: The server opens "medical_article.txt" from the specified path and reads its contents as text data.

[0125] Summary Generation

[0126] Example: The server generates the following summary using the summary generation model:

[0127] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, generally showed the expected effects."

[0128] Registering verification points

[0129] Example: A user enters verification points such as "study validity," "statistical accuracy," and "result reproducibility" into a web form.

[0130] Analysis of the paper

[0131] Example: The server applies a text classification model to each validation point and produces the following classification results:

[0132] Study validity: "Positive"

[0133] Statistical accuracy: "Neutral"

[0134] Reproducibility of results: "Mixed"

[0135] Providing classification results

[0136] Example: Displaying the analysis results in a table format in a web interface.

[0137] This system allows medical and pharmaceutical experts to efficiently evaluate and verify vast amounts of medical research.

[0138] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0139] Step 1: Initializing the generative AI model

[0140] Specific behavior:

[0141] When the system starts up, the server loads the generative AI model from disk, loads a "summary generation model" and a "text classification model" for analyzing medical papers from around the world, and initializes them.

[0142] Input and Output: The input is the generative AI model file on disk, and the output is the initialized generative AI model (in memory).

[0143] Step 2: User uploads medical papers

[0144] Specific behavior:

[0145] The user opens the system's web interface, selects a medical paper file using the file selection dialog, and clicks the upload button. The terminal sends the selected file path to the server, which stores it in a temporary directory.

[0146] Input and Output: The input is the file path of the medical paper selected by the user, and the output is the paper file stored in the server's storage.

[0147] Step 3: Reading medical papers

[0148] Specific behavior:

[0149] The server uses the file path of the uploaded medical paper to open the paper file, read the contents of the file as text data, and save it in memory. It then records the first few lines of the read text data in a log to confirm that it was read successfully.

[0150] Input and output: The input is a paper file stored in the server's storage, and the output is text data stored in memory.

[0151] Step 4: Summary generation

[0152] Specific behavior:

[0153] The server passes the text data of medical papers stored in memory to a summary generation model, which processes the text data to generate a short summary, which is then sent to a web interface for presentation to the user.

[0154] Input and Output: The input is the text data stored in memory, and the output is the generated summary sentence.

[0155] Step 5: Register your verification points

[0156] Specific behavior:

[0157] The user enters the verification points in a form on the web interface and sends them to the server, which stores them in a database.

[0158] Input and Output: The input is the verification point entered by the user and the output is the verification point stored in the database.

[0159] Step 6: Analyzing the paper

[0160] Specific behavior:

[0161] The server applies a text classification model to the medical text data stored in memory and the validation points received from the user, generating a classification result for each validation point and storing it in an internal data structure.

[0162] Input and Output: The input is the text data stored in memory and the validation points stored in the database, and the output is the classification results stored in the internal data structure.

[0163] Step 7: Providing classification results

[0164] Specific behavior:

[0165] The server retrieves the classification results stored in an internal data structure, converts them into a user-friendly format, and sends the converted analysis results to a web interface for display to the user.

[0166] Input and Output: The input is the classification results stored in an internal data structure, and the output is the analysis results displayed in a web interface.

[0167] (Application example 1)

[0168] Next, a description will be given of Application 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."

[0169] Traditional logistics centers are required to manage a huge number of shipping lists and inventory reports, but they lack the means to process them efficiently.In addition, verifying and accurately evaluating the contents of the reports requires a lot of time and effort, and there is a risk of human error.

[0170] 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.

[0171] In this invention, the server includes means for initializing a generative AI model, means for reading a document, means for passing the text data of the read document to a summary generation model and generating a summary, means for providing the generated summary to a user, means for classifying each part of the read document using a text classification model using multiple verification points received from the user and obtaining the classification results, and means for providing the obtained classification results to the user. This enables efficient and accurate processing and evaluation of documents such as shipping lists and inventory reports, even in logistics centers.

[0172] A "generative AI model" is an algorithmic model designed to use artificial intelligence to perform tasks such as text generation, summarization, and classification.

[0173] A "document" is a collection of information, including text data, recorded on paper or in digital form.

[0174] "Means of reading" refers to the process of inputting a specified document file into the system and capturing its contents as digital data.

[0175] A "summary generation model" is an algorithmic model that extracts important information from long pieces of text and summarizes it concisely.

[0176] The "means for generating a summary" refers to a process for generating a summary from the input text data using a summary generation model.

[0177] "Means for providing" refers to the process for displaying or communicating the generated information or classification results to the user.

[0178] "Verification points" refer to specific perspectives or criteria that users evaluate.

[0179] A "text classification model" is an algorithmic model for classifying input text data into specific categories.

[0180] The "means for obtaining classification results" refers to the process of using a text classification model to determine the category of text data for each particular validation point and obtain a result.

[0181] A "smart device" refers to an electronic device that has internet connectivity, computing power, and can run applications.

[0182] This invention relates to a system that uses generative AI models to efficiently manage shipping lists and inventory reports within a logistics center and accurately evaluate their contents. This system allows managers and staff to acquire, analyze, and evaluate information using smart devices, thereby improving work efficiency and accuracy.

[0183] The server first initializes the generative AI models, which include a summary generation model and a text classification model. The summary generation model is used to extract important information from long texts and provide them as short summaries, while the text classification model is used to classify texts into specific categories.

[0184] Users upload logistics documents such as shipping lists and inventory reports, which are then loaded onto the server, where they are converted into text format and prepared for analysis.

[0185] The server then passes the read text data to a summary generation model to generate a summary. This summary is automatically generated in a short sentence format, allowing users to quickly grasp the content. The generated summary is then provided to the user via their smart device.

[0186] Users can set multiple evaluation points for shipping and inventory management. These verification points include evaluation criteria such as "shipping accuracy," "inventory accuracy," and "report validity."

[0187] The server analyzes the loaded text data using a text classification model based on the verification points received from the user. As a result of the analysis, a classification result is obtained for each verification point. This classification result is provided to the user via their smart device, and the user can check the evaluation results for each verification point.

[0188] The hardware used includes servers, smartphones, smart glasses, or logistics robots, while the software used is primarily Python programs and the Hugging Face transformers library.

[0189] For example, let's take a concrete example of how the system is used within a logistics center. When a user uploads a shipping report called "path / to / shipping_report.txt" to the system, the server reads the contents of the report and generates a summary. The summary is provided in the form of, "This shipping report describes the shipping status within a specific period, and shows that most tasks are being carried out according to schedule." Next, when the user registers verification points such as "Shipping accuracy," "Inventory accuracy," and "Report validity," the server classifies each verification point and produces classification results such as "Shipping accuracy: Positive," "Inventory accuracy: Neutral," and "Report validity: Positive."

[0190] Example prompt sentence:

[0191] "Design a system to assist distribution centers in the process of accurately capturing and validating the contents of shipping lists and inventory reports. The system will utilize a generative AI model to generate summaries and provide classification results based on validation points. Specifically, we will use a Python program and the Hugging Face transformers library."

[0192] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0193] Step 1:

[0194] The server initializes the generative AI model. Specifically, it initializes a summary generation model for summarizing text data and a text classification model for classifying text. As input, the configuration file and initialization data for the generative AI model are passed. As output, the launched summary generation model and text classification model are obtained.

[0195] Step 2:

[0196] The user uploads a shipping list or inventory report through a terminal. Specifically, the user enters the file path specified by the user and sends it to the server. The input includes the file path information selected by the user. The output is the file path information sent to the server.

[0197] Step 3:

[0198] The server reads the uploaded document file. Specifically, it opens the specified file path and reads its contents as text data. The input includes the file path information sent by the user. The output is the read text data.

[0199] Step 4:

[0200] The server passes text data to a summary generation model to generate a summary. Specifically, the text data is input to the summary generation model, and the model generates a summary. The input includes the read text data. The output is the generated summary text.

[0201] Step 5:

[0202] The server provides the generated summary to the user. Specifically, it sends the summary text to the terminal so that the user can view it on their smart device. The input includes the generated summary text. The output is display data that the user can view on their terminal.

[0203] Step 6:

[0204] The user sets the verification points for evaluation. Specifically, the user sends the aspects they want to evaluate (e.g., "shipping accuracy," "inventory accuracy," "report validity") to the server via their terminal. The input includes the verification point information selected by the user. The output is the verification point information sent to the server.

[0205] Step 7:

[0206] The server analyzes the text data loaded by the server using a text classification model based on the verification points. Specifically, the text data is classified for each verification point using the verification point information. The input includes the loaded text data and verification point information. The output is a classification result for each verification point.

[0207] Step 8:

[0208] The classification results acquired by the server are provided to the user. Specifically, the classification results are sent to the terminal, allowing the user to check the evaluation results for each perspective on their smart device. The input includes the classification results. The output is display data that the user can check on their terminal.

[0209] 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.

[0210] This invention provides more flexible and personalized information by combining a system that uses a generative AI model to accurately grasp the content of medical papers and assist in the process of determining their validity with an emotion engine that recognizes the user's emotions. Specific embodiments of this system are described below.

[0211] The server initializes the generated AI model

[0212] First, the server initializes the generative AI model. This generative AI model includes a summary generation model for performing text summarization and a text classification model for classifying text. By initializing these models, the server prepares for the automatic processing of medical papers. At the same time, an emotion engine for recognizing user emotions is also initialized.

[0213] User uploads medical paper

[0214] Next, the user uploads the medical paper. The user inputs the file path of the medical paper and sends it to the server. The file path is input via the interface of the terminal used by the user.

[0215] The server loads medical papers

[0216] The server receives the file path and opens the specified file. The server reads the contents of the paper file as text data. This text data includes the full text of the medical paper.

[0217] Server generates summary

[0218] The server passes the text data it has read to a summary generation model, which then generates a summary. The summary generation model extracts the main points of the text and generates a summary. The generated summary is short and concise, describing the main points of the paper.

[0219] The server provides the summary

[0220] The server provides the generated abstract to the user, who can then view the abstract through a terminal interface to quickly understand the content of the paper.

[0221] User registers verification point

[0222] Users input verification criteria for evaluating medical papers and submit them to the server. These verification criteria include the validity of the research, statistical accuracy, and reproducibility of the results.

[0223] The server analyzes the paper

[0224] The server uses the verification points received from the user and the text data of the medical paper to analyze the paper using a text classification model, and classifies the corresponding part for each verification point to obtain the classification result.

[0225] The server provides the classification results

[0226] The server provides the user with the classification results for each verification point. The user can view the evaluation results for each verification point through the terminal interface, allowing them to check the analysis results for each aspect of the medical paper.

[0227] Recognize user emotions

[0228] The emotion engine recognizes the user's emotions based on their input and operation history. For example, the emotion engine can detect when the user is feeling stressed or interested. This allows the system to understand the user's current emotional state.

[0229] Emotion-based adjustment of information provision

[0230] The server can then adapt the way it presents information based on the user's perceived emotional state. For example, if the user is feeling stressed, it can provide a more concise summary or classification in an easy-to-understand format. It can also provide more detailed information or additional references if the user is interested.

[0231] Specific examples

[0232] 1. User uploads paper file

[0233] Example: A user uploads a medical article called "path / to / medical_article.txt" to the server.

[0234] 2. The server generates the summary

[0235] Example: The server reads the content of a paper and uses a summary generation model to generate the following summary:

[0236] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, showed the expected effects."

[0237] 3. User registers verification points

[0238] Example: A user registers verification points such as "research validity," "statistical accuracy," and "reproducibility of results."

[0239] 4. The server analyzes the paper

[0240] Example: The server runs a text classification model on each validation point and obtains the following classification results:

[0241] Study validity: "Positive"

[0242] Statistical accuracy: "Neutral"

[0243] Reproducibility of results: "Mixed"

[0244] 5. Emotion recognition and information provision coordination

[0245] Example: If the emotion engine detects a stressed state from the user's operation history, the server will present the information more concisely, making it easier to understand.

[0246] This system allows users to efficiently evaluate the validity and reliability of medical papers and receive personalized information based on the user's emotional state.

[0247] The processing flow will be explained below.

[0248] Step 1:

[0249] The server initializes the generative AI model and emotion engine. Specifically, it loads the summary generation model, text classification model, and emotion engine that recognizes the user's emotions into memory and makes them ready for use.

[0250] Step 2:

[0251] The user inputs the file path of the medical paper using the terminal and sends it to the server. The file path indicates the location information of the medical paper specified by the user.

[0252] Step 3:

[0253] The server opens the specified medical paper file based on the file path received from the user. The server reads the contents of the medical paper as text data and stores it in memory.

[0254] Step 4:

[0255] The server inputs the text data into a summary generation model to generate a summary, which is a short sentence that succinctly summarizes the main points of the paper.

[0256] Step 5:

[0257] The server sends the generated abstract to the user's terminal, and the user can view the abstract through the terminal interface. The user can refer to the abstract and quickly understand the content of the paper.

[0258] Step 6:

[0259] Users input the validation points used to evaluate medical papers through their terminals and send them to the server. These validation points include aspects that users want to evaluate, such as the validity of the research, statistical accuracy, and reproducibility of results.

[0260] Step 7:

[0261] The server uses the verification points received from the user and the text data of the medical paper it has read to analyze the paper using a text classification model. For each verification point, the model classifies the corresponding part and obtains the classification result.

[0262] Step 8:

[0263] The server sends the classification results for each verification point to the user's device, and the user can view the classification results through the device's interface, allowing the user to check the evaluation results of the paper for each perspective.

[0264] Step 9:

[0265] The emotion engine recognizes the user's emotional state from their input, operation history, reactions, etc. For example, if a user browses a lot of literature in a short period of time, the emotion engine can determine that the user is in a stressful state.

[0266] Step 10:

[0267] The server adjusts the way it presents information based on the user's emotional state. For example, if the user is feeling stressed, it will adjust the way it presents information by making summaries and classification results more concise and easy to understand.

[0268] Step 11:

[0269] The server then sends the adjusted information to the user's device, where the user can view the adjusted summary and classification results through the device interface. The adjusted information is optimized to fit the user's emotional state.

[0270] Step 12:

[0271] Users can efficiently evaluate the validity and reliability of medical papers based on the provided abstracts and classification results, and the server records this evaluation process and stores it in a database for future reference and analysis.

[0272] In this way, a system that combines an emotion engine can accurately grasp the content of medical papers and provide information that corresponds to the user's emotional state, thereby providing personalized support that meets the user's needs.

[0273] Example 2

[0274] Next, a description will be given of 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."

[0275] Modern medical papers contain a vast amount of information, making it extremely difficult to quickly and accurately grasp and evaluate its validity. Furthermore, because information is not provided in response to the user's emotions or state, there are concerns that this will increase the burden on researchers. There is a need to solve these problems and provide an environment in which users can more efficiently and comfortably evaluate and understand medical papers.

[0276] 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.

[0277] In this invention, the server includes means for initializing the generation AI model, means for reading electronic documents, and means for passing the text data of the read electronic documents to the summary generation model and generating summaries, thereby enabling users to quickly grasp the vast amount of information in electronic documents and providing personalized information according to the user's emotional state.

[0278] A "generative AI model" refers to an artificial intelligence model for summarizing and classifying text data.

[0279] A "summary generation model" refers to an artificial intelligence model that extracts the main points of input text data and generates a concise summary.

[0280] "Text classification model" refers to an artificial intelligence model for classifying input text data based on specific criteria.

[0281] "Electronic literature" refers to text data such as papers and articles stored in digital format.

[0282] "User" refers to a person who uses this system and operates it for the purpose of understanding and evaluating the content of electronic documents.

[0283] An "emotion engine" refers to software or a system that recognizes emotions and states based on user input and operation history.

[0284] "Verification points" refer to specific viewpoints or criteria that users set for evaluating electronic documents.

[0285] "Means for adjusting the way information is presented" refers to processes or software functions for changing the format or content of the information provided depending on the perceived user's emotions or state.

[0286] This invention provides more flexible and personalized information by combining a system that uses a generative AI model to accurately grasp the content of electronic documents and assist in the process of evaluating their validity with an emotion engine that recognizes the user's emotions. Specific embodiments of this system are described below.

[0287] First, the server initializes the generative AI model. This generative AI model includes a "summary generation model" for summarizing text and a "text classification model" for classifying text. By initializing these models, the server prepares for the automatic processing of electronic documents. At the same time, the server also initializes an "emotion engine" for recognizing user emotions. This initialization process includes loading model parameters and configuration files.

[0288] Next, the user uploads the electronic document. The user uses the terminal interface to input the file path of the electronic document and sends it to the server. Specifically, the user can specify the file path as "path / to / medical_article.txt". This input information is received by the server.

[0289] The server uses the received file path to open the specified file. Using the standard Python library, the file is opened and the contents are read as text data. This allows the full text of the electronic document to be obtained as data.

[0290] The acquired text data is passed to a summary generation model, which extracts the key points of the text and generates a summary. For example, the model can generate a summary like this: "A study evaluated the efficacy and safety of drug A under specific conditions. The results, although condition dependent, showed the expected effects." This summary is then sent from the server to the user's device.

[0291] After viewing an abstract, users input specific verification criteria for evaluating the electronic literature, such as "study validity," "statistical accuracy," and "reproducibility of results." These verification criteria are then sent to the server.

[0292] The server uses the received validation points and the text data of the electronic document to analyze the document using a text classification model. It classifies the corresponding part for each validation point and obtains a classification result. For example, it can generate classification results such as "Research validity: Positive," "Statistical accuracy: Neutral," and "Reproducibility of results: Mixed." The classification results are then sent to the user's device, where the user can view the evaluation results for each validation point.

[0293] Furthermore, the emotion engine recognizes emotions based on the user's input and operation history. For example, the emotion engine detects whether the user is feeling stressed or interested. Based on the recognized emotional state, the server adjusts the way information is provided. For example, if the user is feeling stressed, the server can provide summaries and classification results in a more concise and easy-to-understand format. If the user is interested, the server can provide detailed information and additional reference materials.

[0294] As an example of a prompt sentence, a user might upload a medical paper called "path / to / medical_article.txt" to a server, and a summary generation model might generate the following summary: "The study evaluated the efficacy and safety of drug A under specific conditions. The results, although dependent on the conditions, showed the expected effects."

[0295] This allows users to efficiently evaluate the validity and reliability of medical papers and receive personalized information according to their emotional state.

[0296] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0297] Step 1:

[0298] The server initializes the generative AI model. Specifically, the server loads and initializes the summary generation model and text classification model. It also initializes the emotion engine to recognize the user's emotions. This includes loading the model parameters and configuration files. The model file and configuration file are required as input, and an initialized model object is generated as output.

[0299] Step 2:

[0300] The user uploads an electronic document. The user specifies the file path on the terminal interface and sends it to the server. Specifically, the user enters the file path "path / to / medical_article.txt". The file path is provided as input, and the file path information is sent to the server.

[0301] Step 3:

[0302] The server reads the electronic document. Based on the received file path, the server opens the specified file and reads the contents as text data. Depending on the programming language used, for example, Python's "open()" function is used. The file path is received as input, and the full text data of the electronic document is generated as output.

[0303] Step 4:

[0304] The server generates summaries of electronic literature. The text data is passed to a summary generation model, which extracts the main points of the text and generates a summary. Specifically, a generative AI model such as BERT is used to summarize the text data. The text data of the electronic literature is required as input, and summary text is generated as output. For example, a summary might be generated that reads, "A study evaluated the efficacy and safety of drug A under specific conditions. The results, while dependent on the conditions, showed the expected effects."

[0305] Step 5:

[0306] The server provides the generated summary to the user. The summary text is sent from the server to the user's terminal, and the user views the summary on the terminal interface. The generated summary text is required as input, and the summary text is displayed on the terminal screen as output.

[0307] Step 6:

[0308] The user registers validation points. The user enters multiple validation points into the terminal interface and sends them to the server. Specifically, the user enters items such as "research validity," "statistical accuracy," and "result reproducibility." A list of validation points is provided as input, and the information is sent to the server.

[0309] Step 7:

[0310] The server analyzes the electronic document. It uses a text classification model to classify the specified part of the electronic document and obtains classification results for each verification point. The input requires the text data of the electronic document and a list of verification points, and the output generates classification results for each verification point. For example, classification results such as "Research validity: Positive," "Statistical accuracy: Neutral," and "Reproducibility of results: Mixed" can be obtained.

[0311] Step 8:

[0312] The server provides the classification results to the user. The acquired classification results are sent from the server to the user's device, and the user can view the evaluation results of each verification point on the device interface. The classification results are required as input, and are displayed on the screen as output.

[0313] Step 9:

[0314] The emotion engine recognizes the user's emotions. It analyzes emotions from the user's input and operation history, and detects emotional states such as stress and interest. Input data includes operation history such as input speed and click frequency. The user's emotional state is recognized as the output.

[0315] Step 10:

[0316] The server adjusts the way information is presented. The server adjusts the format of information presentation based on the recognized emotional state. For example, if the user is feeling stressed, the server will provide information in a concise and easy-to-understand format. The emotional state is required as input, and tailored information is provided as output.

[0317] (Application example 2)

[0318] Next, a description will be given of Application 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."

[0319] Problems often occur in the manufacturing process of modern factories, and dealing with them can take a lot of time and effort. It is also difficult to quickly and accurately grasp the large amount of information available in technical journals and specialized papers, and to implement appropriate countermeasures. Furthermore, dealing with problems often causes stress for on-site workers. In these situations, there is a growing need for a system that can efficiently summarize information and provide it while taking emotions into consideration.

[0320] 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.

[0321] In this invention, the server includes means for initializing a generative AI model, means for reading text data, means for passing the read text data to a summary generation model and generating a summary, means for providing the generated summary to a user, means for classifying each portion of the read text data using a text classification model and obtaining a classification result using multiple verification points received from the user, means for providing the obtained classification result to the user, means for initializing an emotion engine that recognizes the user's emotions, and means for adjusting a method of providing information based on the user's emotional state recognized by the emotion engine. This speeds up troubleshooting in factories and enables support that takes user emotions into consideration.

[0322] A "generative AI model" is an artificial intelligence model trained to perform tasks such as summarizing, classifying, and generating text data.

[0323] An "emotion engine" is software or hardware that recognizes a user's emotional state and adjusts the way information is presented.

[0324] "Text data" refers to data that consists of textual information, and includes technical journals and specialized papers.

[0325] A "summary generation model" is a machine learning model that extracts key points from long text data and summarizes them as short sentences.

[0326] A "text classification model" is an artificial intelligence model trained to classify text data into specific classes or categories.

[0327] "Verification points" are points of view that the user inputs as evaluation items for text data, and include, for example, "reproducibility of the problem" and "appropriateness of the countermeasures."

[0328] "Adjusting the method of providing information" refers to optimizing the content and format of the information provided based on the user's emotional state.

[0329] This invention relates to a system that utilizes a generative AI model to efficiently summarize technical journals and specialized papers related to manufacturing processes in factories, and further recognizes user emotions to optimize information provision. Specific embodiments of this system are described below.

[0330] The server initializes the generated AI model

[0331] The server initializes the generative AI model. This generative AI model includes a summary generation model for summarizing text data and a text classification model for classifying text data. These models are initialized using the HuggingFace transformers library. An emotion engine that recognizes the user's emotions is also initialized at the same time, and OpenCV is used for this emotion engine.

[0332] Factory workers upload daily reports or technical papers to the robot

[0333] A factory worker inputs the file path of the daily report or technical paper through the terminal interface and sends it to the server.

[0334] The server reads the text data.

[0335] The server opens the specified file and reads it as text data, which may include the full text of technical journals or technical papers.

[0336] Server generates summary

[0337] The server passes the text data it has read to a summary generation model, which then generates a summary by extracting the main points of the text.

[0338] The server provides the summary

[0339] The server provides the generated summary to the factory staff, who can view the summary through the terminal interface and quickly understand the contents.

[0340] Factory staff register verification points

[0341] Factory staff input verification points for evaluating technical daily reports and specialist papers and send them to the server. These verification points include "reproducibility of the problem," "appropriateness of countermeasures," and "scope of impact."

[0342] The server analyzes the text data

[0343] The server uses the verification points and text data received from the factory staff to perform analysis using a text classification model. The server classifies the corresponding part for each verification point and obtains the classification result.

[0344] The server provides the classification results

[0345] The server provides the classification results for each verification point to factory workers, who can view the evaluation results for each verification point through the terminal interface.

[0346] Recognize user emotions

[0347] The emotion engine recognizes the emotions of factory workers based on their input, operation history, and facial expressions, allowing the system to understand their current emotional state.

[0348] Emotion-based adjustment of information provision

[0349] The server adjusts the way it presents information based on the emotional state of the factory worker. For example, if the factory worker is feeling stressed, it will provide a more concise summary and classification results in an easy-to-understand format. If the factory worker is interested, it will provide detailed information or additional reference materials.

[0350] As a concrete example, consider a scenario in which a factory worker uploads a daily technical report called " / reports / machine_failure_report.txt" and the server generates the following summary: "In this daily report, machine X experienced motor overheating under certain operating conditions. As a countermeasure, the cooling system was adjusted." The factory worker also registers verification points such as "reproducibility of the problem," "appropriateness of the countermeasure," and "scope of impact," and the server provides the following classification results: "reproducibility of the problem: high," "appropriateness of the countermeasure: good," and "scope of impact: limited." The emotion engine can detect the stress level of factory workers and provide more concise information.

[0351] An example of a prompt is as follows:

[0352] "Please summarize the following technical daily report: / path / to / machine_failure_report.txt. In addition, please provide a classification and analysis based on the following verification points: 'Reproducibility of the problem', 'Appropriateness of the countermeasures', 'Scope of impact'. Please consider the feelings of the factory staff and simplify the information if stress is felt."

[0353] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0354] Step 1:

[0355] The server initializes the generative AI models, which include two models: a summary generation model and a text classification model. The server initializes these models using HuggingFace's transformers library. As input, it takes the URL or path of a pre-trained model. As output, it gets the initialized model object.

[0356] Step 2:

[0357] The user inputs the file path of the daily report or technical paper through the terminal interface and sends it to the server. The input is the file path (e.g., " / reports / machine_failure_report.txt"), and the output is that the sent file path reaches the server.

[0358] Step 3:

[0359] Based on the file path received by the server, the specified file is opened and read as text data. The input is the file path, and the output is text data containing the file contents.

[0360] Step 4:

[0361] The server passes the text data it has read to a summary generation model to generate a summary. Here, the summary generation model extracts the main points of the text data and summarizes them as short sentences. The input is the text data, and the output is the generated summary text.

[0362] Step 5:

[0363] The server provides the generated summary to the user by sending data to display the summary text on the user's terminal. The input is the summary text, and the output is the summary content displayed on the user interface.

[0364] Step 6:

[0365] The user inputs verification points for evaluating technical journals and technical papers into the terminal interface and sends them to the server. The inputs are verification points (e.g., "reproducibility of the problem," "appropriateness of the countermeasures," "scope of impact"), and the output is the verification points that reach the server.

[0366] Step 7:

[0367] The server uses the verification points and text data received from the user to perform analysis using a text classification model. The server classifies the corresponding text portion for each verification point and obtains the classification result. The input is the text data and verification points, and the output is the classification result for each verification point.

[0368] Step 8:

[0369] The server provides the user with the classification results for each verification point it has acquired. Here, it sends data to display the classification results on the user's device. The input is the classification results, and the output is the evaluation content displayed on the user interface.

[0370] Step 9:

[0371] The server uses an emotion engine to recognize the user's emotions based on the user's input, operation history, and facial expressions captured by the device's camera. The input is the user's operation history and camera footage, and the output is the recognized emotional state of the user.

[0372] Step 10:

[0373] The server adjusts the way it presents information based on the user's recognized emotional state. For example, if the user is feeling stressed, it will provide more concise information in an easy-to-understand format. The input is the user's emotional state, and the output is the adjusted information presentation data.

[0374] 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.

[0375] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) 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.

[0376] 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.

[0377] [Second embodiment]

[0378] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0379] 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.

[0380] 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).

[0381] 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.

[0382] 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.

[0383] 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).

[0384] 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.

[0385] 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.

[0386] 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.

[0387] 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.

[0388] In the smart glasses 214, the 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.

[0389] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0390] This invention is a system that uses a generative AI model to accurately grasp the content of medical papers and support the process of determining their validity. Specific embodiments of this system are described below.

[0391] The server initializes the generated AI model

[0392] First, the server initializes the generative AI model. This generative AI model includes a summary generation model for performing text summarization and a text classification model for classifying text. By initializing these models, the server prepares to process medical papers.

[0393] User uploads medical paper

[0394] Next, the user uploads a medical paper. The user inputs the file path of the medical paper and sends it to the server. This file path is used when the server reads the text of the medical paper.

[0395] The server loads medical papers

[0396] The server reads the medical paper uploaded by the user. The server opens the paper file and reads its contents as text data. This text data is then passed to the summary generation model.

[0397] Server generates summary

[0398] The server passes the text data of medical papers read into a summary generation model, which then generates a summary. This summary generation model extracts important information from long sentences and outputs a short summary. The generated summary is then provided to the user.

[0399] User registers verification point

[0400] Users register validation criteria for evaluating medical papers, including the validity of the research, statistical accuracy, and reproducibility of results, among other aspects that the user wishes to evaluate.

[0401] The server analyzes the paper

[0402] Based on the verification points received from the user, the server analyzes each part of the medical paper using a text classification model. The server uses each verification point and the text data of the paper to obtain a classification result. This classification model is trained to classify the input text into a specific category.

[0403] The server provides the classification results

[0404] The server provides the user with the classification results for each verification point, allowing the user to check the evaluation results for each aspect of the medical paper.

[0405] Specific examples

[0406] A specific example is given below.

[0407] 1. User uploads paper file

[0408] Example: A user uploads a medical article called "path / to / medical_article.txt" to the server.

[0409] 2. The server generates the summary

[0410] Example: The server reads the content of a paper and uses a summary generation model to generate the following summary:

[0411] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, showed the expected effects."

[0412] 3. User registers verification points

[0413] Example: A user registers verification points such as "research validity," "statistical accuracy," and "reproducibility of results."

[0414] 4. The server analyzes the paper

[0415] Example: The server runs a text classification model on each validation point and obtains the following classification results:

[0416] Study validity: "Positive"

[0417] Statistical accuracy: "Neutral"

[0418] Reproducibility of results: "Mixed"

[0419] In this way, by using this system, medical and pharmaceutical experts can efficiently and accurately evaluate and verify a huge amount of medical research papers.

[0420] The processing flow will be explained below.

[0421] Step 1:

[0422] The server initializes the generative AI model, specifically, loads the summary generation model and text classification model, and deploys them to storage devices to efficiently perform natural language processing tasks.

[0423] Step 2:

[0424] The user inputs the file path of the medical paper and sends it to the server. The file path is input via the interface of the terminal used by the user.

[0425] Step 3:

[0426] The server receives the file path and opens the specified file. The server reads the contents of the paper file as text data. This text data includes the full text of the medical paper.

[0427] Step 4:

[0428] The server passes the read text data to a summary generation model. The model extracts the main points of the text and generates a summary. The generated summary is short and concise, describing the main points of the paper.

[0429] Step 5:

[0430] The server provides the generated abstract to the user, who can then view the abstract through a terminal interface to quickly understand the content of the paper.

[0431] Step 6:

[0432] Users input the validation points used to evaluate papers and submit them to the server. These validation points include the validity of the research, statistical accuracy, and reproducibility of results.

[0433] Step 7:

[0434] The server uses the verification points received from the user and the text data of the medical paper to analyze the paper using a text classification model. For each verification point, the corresponding part is classified and a classification result is obtained.

[0435] Step 8:

[0436] The server provides the classification results to the user, who can then view the evaluation results for each verification point via the terminal interface, thereby confirming the analysis results for each aspect of the medical paper.

[0437] Step 9:

[0438] Users evaluate the validity and reliability of medical papers based on the provided abstracts and classification results. The server records this evaluation process and stores it in a database for future reference and analysis.

[0439] Example 1

[0440] Next, a description will be given of 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."

[0441] Efficiently evaluating and verifying large volumes of pharmaceutical research papers is a major challenge for medical and pharmaceutical experts. In particular, there is a need to quickly and accurately evaluate the validity, statistical accuracy, and reproducibility of results of research papers. However, conventional methods require a huge amount of time and effort to analyze each paper in detail, making them inefficient. Furthermore, results can vary depending on the evaluator's subjectivity, making it difficult to perform consistent evaluations.

[0442] 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.

[0443] In this invention, the server includes means for initializing the generative AI model, means for a user to upload a medical paper, means for reading the medical paper, means for passing the text data of the read medical paper to a summary generation model to generate a summary, means for providing the generated summary to the user, means for registering multiple verification points received from the user, means for classifying each part of the read medical paper using the multiple verification points received from the user with a text classification model to obtain the classification results, and means for providing the obtained classification results to the user, thereby enabling efficient summary generation of medical papers and automatic and consistent evaluation from multiple perspectives.

[0444] A "generative AI model" is a trained model that uses artificial intelligence to summarize and classify text.

[0445] The "server" is the computer system that serves as the core of the system, initializing the generative AI model, processing medical papers, and interfacing with users.

[0446] A "user" is a person or organization that uses the system to upload medical papers and receive their summaries and evaluation results.

[0447] A "medical paper" is a piece of writing that summarizes research results and findings related to medicine and pharmacy.

[0448] A "summary generation model" is a type of generative AI model used to extract important information from medical papers and generate short summaries.

[0449] A "text classification model" is a type of generative AI model used to classify parts of medical papers into specific categories.

[0450] "Verification points" are items that serve as viewpoints or criteria when a user evaluates a medical paper.

[0451] "Classification results" are evaluation results for each part of a medical paper generated by a text classification model.

[0452] "Upload" is the act of a user sending data from their own terminal to a server.

[0453] A "web interface" is a screen that uses a web browser to allow a user to access and operate the system.

[0454] This invention is a system that uses a generative AI model to efficiently summarize the contents of medical papers and assists in the process of determining their validity based on each verification point. Specific embodiments are described below.

[0455] 1. Hardware and Software Configuration

[0456] The server plays a central role in this system. The server hardware includes a powerful CPU, sufficient memory, and large storage capacity. The main software used includes Python, the Transformers library, and a suitable database management system.

[0457] 2. Initializing the Generative AI Model

[0458] When the system is initialized, the server loads generative AI models from disk. Specifically, it loads a "summary generation model" that generates summaries of medical papers and a "text classification model" that classifies text. These models have been trained in advance using large amounts of text data, and the server initializes their parameters and loads them into memory.

[0459] 3. User upload of medical papers

[0460] Users upload medical papers through the system's web interface. They use a file selection dialog to select the medical paper file they want to evaluate and click the upload button. The terminal sends the selected file path to the server, which stores it in a temporary directory.

[0461] 4. Reading medical papers

[0462] The server uses the file path sent by the user to open the uploaded medical paper, reads the file contents as text data, and stores it in memory. This text data is used in the subsequent summary generation process.

[0463] 5. Summary Generation

[0464] The server passes the text data of the loaded medical paper to a summary generation model to generate a summary. This summary generation model extracts important information from the medical paper and outputs a short summary. The generated summary is provided to users so that they can easily understand the outline of the paper.

[0465] 6. Registering Verification Points

[0466] Users register multiple validation points to evaluate the validity of a paper. These include aspects such as "research validity," "statistical accuracy," and "reproducibility of results." These validation points are entered on a web interface and sent to the server, which stores them in a database.

[0467] 7. Analysis of the paper

[0468] The server applies a text classification model using the verification points received from the user and the text data loaded, generating a classification result for each verification point and storing it in an internal data structure.

[0469] 8. Providing classification results

[0470] The server takes the generated classification results, converts them into a user-friendly format, and provides the results to the user through a web interface.

[0471] Specific examples

[0472] Initializing a generative AI model

[0473] Example: The server starts and loads a "summary generation model" and a "text classification model" using common natural language processing models.

[0474] User-uploaded medical papers

[0475] Example: User uploads "path / to / medical_article.txt" through the web interface.

[0476] Reading medical papers

[0477] Example: The server opens "medical_article.txt" from the specified path and reads its contents as text data.

[0478] Summary Generation

[0479] Example: The server generates the following summary using the summary generation model:

[0480] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, generally showed the expected effects."

[0481] Registering verification points

[0482] Example: A user enters verification points such as "study validity," "statistical accuracy," and "result reproducibility" into a web form.

[0483] Analysis of the paper

[0484] Example: The server applies a text classification model to each validation point and produces the following classification results:

[0485] Study validity: "Positive"

[0486] Statistical accuracy: "Neutral"

[0487] Reproducibility of results: "Mixed"

[0488] Providing classification results

[0489] Example: Displaying the analysis results in a table format in a web interface.

[0490] This system allows medical and pharmaceutical experts to efficiently evaluate and verify vast amounts of medical research.

[0491] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0492] Step 1: Initializing the generative AI model

[0493] Specific behavior:

[0494] When the system starts up, the server loads the generative AI model from disk, loads a "summary generation model" and a "text classification model" for analyzing medical papers from around the world, and initializes them.

[0495] Input and Output: The input is the generative AI model file on disk, and the output is the initialized generative AI model (in memory).

[0496] Step 2: User uploads medical papers

[0497] Specific behavior:

[0498] The user opens the system's web interface, selects a medical paper file using the file selection dialog, and clicks the upload button. The terminal sends the selected file path to the server, which stores it in a temporary directory.

[0499] Input and Output: The input is the file path of the medical paper selected by the user, and the output is the paper file stored in the server's storage.

[0500] Step 3: Reading medical papers

[0501] Specific behavior:

[0502] The server uses the file path of the uploaded medical paper to open the paper file, read the contents of the file as text data, and save it in memory. It then records the first few lines of the read text data in a log to confirm that it was read successfully.

[0503] Input and output: The input is a paper file stored in the server's storage, and the output is text data stored in memory.

[0504] Step 4: Summary generation

[0505] Specific behavior:

[0506] The server passes the text data of medical papers stored in memory to a summary generation model, which processes the text data to generate a short summary, which is then sent to a web interface for presentation to the user.

[0507] Input and Output: The input is the text data stored in memory, and the output is the generated summary sentence.

[0508] Step 5: Register your verification points

[0509] Specific behavior:

[0510] The user enters the verification points in a form on the web interface and sends them to the server, which stores them in a database.

[0511] Input and Output: The input is the verification point entered by the user and the output is the verification point stored in the database.

[0512] Step 6: Analyzing the paper

[0513] Specific behavior:

[0514] The server applies a text classification model to the medical text data stored in memory and the validation points received from the user, generating a classification result for each validation point and storing it in an internal data structure.

[0515] Input and Output: The input is the text data stored in memory and the validation points stored in the database, and the output is the classification results stored in the internal data structure.

[0516] Step 7: Providing classification results

[0517] Specific behavior:

[0518] The server retrieves the classification results stored in an internal data structure, converts them into a user-friendly format, and sends the converted analysis results to a web interface for display to the user.

[0519] Input and Output: The input is the classification results stored in an internal data structure, and the output is the analysis results displayed in a web interface.

[0520] (Application example 1)

[0521] Next, a description will be given of Application 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."

[0522] Traditional logistics centers are required to manage a huge number of shipping lists and inventory reports, but they lack the means to process them efficiently.In addition, verifying and accurately evaluating the contents of the reports requires a lot of time and effort, and there is a risk of human error.

[0523] 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.

[0524] In this invention, the server includes means for initializing a generative AI model, means for reading a document, means for passing the text data of the read document to a summary generation model and generating a summary, means for providing the generated summary to a user, means for classifying each part of the read document using a text classification model using multiple verification points received from the user and obtaining the classification results, and means for providing the obtained classification results to the user. This enables efficient and accurate processing and evaluation of documents such as shipping lists and inventory reports, even in logistics centers.

[0525] A "generative AI model" is an algorithmic model designed to use artificial intelligence to perform tasks such as text generation, summarization, and classification.

[0526] A "document" is a collection of information, including text data, recorded on paper or in digital form.

[0527] "Means of reading" refers to the process of inputting a specified document file into the system and capturing its contents as digital data.

[0528] A "summary generation model" is an algorithmic model that extracts important information from long pieces of text and summarizes it concisely.

[0529] The "means for generating a summary" refers to a process for generating a summary from the input text data using a summary generation model.

[0530] "Means for providing" refers to the process for displaying or communicating the generated information or classification results to the user.

[0531] "Verification points" refer to specific perspectives or criteria that users evaluate.

[0532] A "text classification model" is an algorithmic model for classifying input text data into specific categories.

[0533] The "means for obtaining classification results" refers to the process of using a text classification model to determine the category of text data for each particular validation point and obtain a result.

[0534] A "smart device" refers to an electronic device that has internet connectivity, computing power, and can run applications.

[0535] This invention relates to a system that uses generative AI models to efficiently manage shipping lists and inventory reports within a logistics center and accurately evaluate their contents. This system allows managers and staff to acquire, analyze, and evaluate information using smart devices, thereby improving work efficiency and accuracy.

[0536] The server first initializes the generative AI models, which include a summary generation model and a text classification model. The summary generation model is used to extract important information from long texts and provide them as short summaries, while the text classification model is used to classify texts into specific categories.

[0537] Users upload logistics documents such as shipping lists and inventory reports, which are then loaded onto the server, where they are converted into text format and prepared for analysis.

[0538] The server then passes the read text data to a summary generation model to generate a summary. This summary is automatically generated in a short sentence format, allowing users to quickly grasp the content. The generated summary is then provided to the user via their smart device.

[0539] Users can set multiple evaluation points for shipping and inventory management. These verification points include evaluation criteria such as "shipping accuracy," "inventory accuracy," and "report validity."

[0540] The server analyzes the loaded text data using a text classification model based on the verification points received from the user. As a result of the analysis, a classification result is obtained for each verification point. This classification result is provided to the user via their smart device, and the user can check the evaluation results for each verification point.

[0541] The hardware used includes servers, smartphones, smart glasses, or logistics robots, while the software used is primarily Python programs and the Hugging Face transformers library.

[0542] For example, let's take a concrete example of how the system is used within a logistics center. When a user uploads a shipping report called "path / to / shipping_report.txt" to the system, the server reads the contents of the report and generates a summary. The summary is provided in the form of, "This shipping report describes the shipping status within a specific period, and shows that most tasks are being carried out according to schedule." Next, when the user registers verification points such as "Shipping accuracy," "Inventory accuracy," and "Report validity," the server classifies each verification point and produces classification results such as "Shipping accuracy: Positive," "Inventory accuracy: Neutral," and "Report validity: Positive."

[0543] Example prompt sentence:

[0544] "Design a system to assist distribution centers in the process of accurately capturing and validating the contents of shipping lists and inventory reports. The system will utilize a generative AI model to generate summaries and provide classification results based on validation points. Specifically, we will use a Python program and the Hugging Face transformers library."

[0545] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0546] Step 1:

[0547] The server initializes the generative AI model. Specifically, it initializes a summary generation model for summarizing text data and a text classification model for classifying text. As input, the configuration file and initialization data for the generative AI model are passed. As output, the launched summary generation model and text classification model are obtained.

[0548] Step 2:

[0549] The user uploads a shipping list or inventory report through a terminal. Specifically, the user enters the file path specified by the user and sends it to the server. The input includes the file path information selected by the user. The output is the file path information sent to the server.

[0550] Step 3:

[0551] The server reads the uploaded document file. Specifically, it opens the specified file path and reads its contents as text data. The input includes the file path information sent by the user. The output is the read text data.

[0552] Step 4:

[0553] The server passes text data to a summary generation model to generate a summary. Specifically, the text data is input to the summary generation model, and the model generates a summary. The input includes the read text data. The output is the generated summary text.

[0554] Step 5:

[0555] The server provides the generated summary to the user. Specifically, it sends the summary text to the terminal so that the user can view it on their smart device. The input includes the generated summary text. The output is display data that the user can view on their terminal.

[0556] Step 6:

[0557] The user sets the verification points for evaluation. Specifically, the user sends the aspects they want to evaluate (e.g., "shipping accuracy," "inventory accuracy," "report validity") to the server via their terminal. The input includes the verification point information selected by the user. The output is the verification point information sent to the server.

[0558] Step 7:

[0559] The server analyzes the text data loaded by the server using a text classification model based on the verification points. Specifically, the text data is classified for each verification point using the verification point information. The input includes the loaded text data and verification point information. The output is a classification result for each verification point.

[0560] Step 8:

[0561] The classification results acquired by the server are provided to the user. Specifically, the classification results are sent to the terminal, allowing the user to check the evaluation results for each perspective on their smart device. The input includes the classification results. The output is display data that the user can check on their terminal.

[0562] 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.

[0563] This invention provides more flexible and personalized information by combining a system that uses a generative AI model to accurately grasp the content of medical papers and assist in the process of determining their validity with an emotion engine that recognizes the user's emotions. Specific embodiments of this system are described below.

[0564] The server initializes the generated AI model

[0565] First, the server initializes the generative AI model. This generative AI model includes a summary generation model for performing text summarization and a text classification model for classifying text. By initializing these models, the server prepares for the automatic processing of medical papers. At the same time, an emotion engine for recognizing user emotions is also initialized.

[0566] User uploads medical paper

[0567] Next, the user uploads the medical paper. The user inputs the file path of the medical paper and sends it to the server. The file path is input via the interface of the terminal used by the user.

[0568] The server loads medical papers

[0569] The server receives the file path and opens the specified file. The server reads the contents of the paper file as text data. This text data includes the full text of the medical paper.

[0570] Server generates summary

[0571] The server passes the text data it has read to a summary generation model, which then generates a summary. The summary generation model extracts the main points of the text and generates a summary. The generated summary is short and concise, describing the main points of the paper.

[0572] The server provides the summary

[0573] The server provides the generated abstract to the user, who can then view the abstract through a terminal interface to quickly understand the content of the paper.

[0574] User registers verification point

[0575] Users input verification criteria for evaluating medical papers and submit them to the server. These verification criteria include the validity of the research, statistical accuracy, and reproducibility of the results.

[0576] The server analyzes the paper

[0577] The server uses the verification points received from the user and the text data of the medical paper to analyze the paper using a text classification model, and classifies the corresponding part for each verification point to obtain the classification result.

[0578] The server provides the classification results

[0579] The server provides the user with the classification results for each verification point. The user can view the evaluation results for each verification point through the terminal interface, allowing them to check the analysis results for each aspect of the medical paper.

[0580] Recognize user emotions

[0581] The emotion engine recognizes the user's emotions based on their input and operation history. For example, the emotion engine can detect when the user is feeling stressed or interested. This allows the system to understand the user's current emotional state.

[0582] Emotion-based adjustment of information provision

[0583] The server can then adapt the way it presents information based on the user's perceived emotional state. For example, if the user is feeling stressed, it can provide a more concise summary or classification in an easy-to-understand format. It can also provide more detailed information or additional references if the user is interested.

[0584] Specific examples

[0585] 1. User uploads paper file

[0586] Example: A user uploads a medical article called "path / to / medical_article.txt" to the server.

[0587] 2. The server generates the summary

[0588] Example: The server reads the content of a paper and uses a summary generation model to generate the following summary:

[0589] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, showed the expected effects."

[0590] 3. User registers verification points

[0591] Example: A user registers verification points such as "research validity," "statistical accuracy," and "reproducibility of results."

[0592] 4. The server analyzes the paper

[0593] Example: The server runs a text classification model on each validation point and obtains the following classification results:

[0594] Study validity: "Positive"

[0595] Statistical accuracy: "Neutral"

[0596] Reproducibility of results: "Mixed"

[0597] 5. Emotion recognition and information provision coordination

[0598] Example: If the emotion engine detects a stressed state from the user's operation history, the server will present the information more concisely, making it easier to understand.

[0599] This system allows users to efficiently evaluate the validity and reliability of medical papers and receive personalized information based on the user's emotional state.

[0600] The processing flow will be explained below.

[0601] Step 1:

[0602] The server initializes the generative AI model and emotion engine. Specifically, it loads the summary generation model, text classification model, and emotion engine that recognizes the user's emotions into memory and makes them ready for use.

[0603] Step 2:

[0604] The user inputs the file path of the medical paper using the terminal and sends it to the server. The file path indicates the location information of the medical paper specified by the user.

[0605] Step 3:

[0606] The server opens the specified medical paper file based on the file path received from the user. The server reads the contents of the medical paper as text data and stores it in memory.

[0607] Step 4:

[0608] The server inputs the text data into a summary generation model to generate a summary, which is a short sentence that succinctly summarizes the main points of the paper.

[0609] Step 5:

[0610] The server sends the generated abstract to the user's terminal, and the user can view the abstract through the terminal interface. The user can refer to the abstract and quickly understand the content of the paper.

[0611] Step 6:

[0612] Users input the validation points used to evaluate medical papers through their terminals and send them to the server. These validation points include aspects that users want to evaluate, such as the validity of the research, statistical accuracy, and reproducibility of results.

[0613] Step 7:

[0614] The server uses the verification points received from the user and the text data of the medical paper it has read to analyze the paper using a text classification model. For each verification point, the model classifies the corresponding part and obtains the classification result.

[0615] Step 8:

[0616] The server sends the classification results for each verification point to the user's device, and the user can view the classification results through the device's interface, allowing the user to check the evaluation results of the paper for each perspective.

[0617] Step 9:

[0618] The emotion engine recognizes the user's emotional state from their input, operation history, reactions, etc. For example, if a user browses a lot of literature in a short period of time, the emotion engine can determine that the user is in a stressful state.

[0619] Step 10:

[0620] The server adjusts the way it presents information based on the user's emotional state. For example, if the user is feeling stressed, it will adjust the way it presents information by making summaries and classification results more concise and easy to understand.

[0621] Step 11:

[0622] The server then sends the adjusted information to the user's device, where the user can view the adjusted summary and classification results through the device interface. The adjusted information is optimized to fit the user's emotional state.

[0623] Step 12:

[0624] Users can efficiently evaluate the validity and reliability of medical papers based on the provided abstracts and classification results, and the server records this evaluation process and stores it in a database for future reference and analysis.

[0625] In this way, a system that combines an emotion engine can accurately grasp the content of medical papers and provide information that corresponds to the user's emotional state, thereby providing personalized support that meets the user's needs.

[0626] Example 2

[0627] Next, a description will be given of 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."

[0628] Modern medical papers contain a vast amount of information, making it extremely difficult to quickly and accurately grasp and evaluate its validity. Furthermore, because information is not provided in response to the user's emotions or state, there are concerns that this will increase the burden on researchers. There is a need to solve these problems and provide an environment in which users can more efficiently and comfortably evaluate and understand medical papers.

[0629] 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.

[0630] In this invention, the server includes means for initializing the generation AI model, means for reading electronic documents, and means for passing the text data of the read electronic documents to the summary generation model and generating summaries, thereby enabling users to quickly grasp the vast amount of information in electronic documents and providing personalized information according to the user's emotional state.

[0631] A "generative AI model" refers to an artificial intelligence model for summarizing and classifying text data.

[0632] A "summary generation model" refers to an artificial intelligence model that extracts the main points of input text data and generates a concise summary.

[0633] "Text classification model" refers to an artificial intelligence model for classifying input text data based on specific criteria.

[0634] "Electronic literature" refers to text data such as papers and articles stored in digital format.

[0635] "User" refers to a person who uses this system and operates it for the purpose of understanding and evaluating the content of electronic documents.

[0636] An "emotion engine" refers to software or a system that recognizes emotions and states based on user input and operation history.

[0637] "Verification points" refer to specific viewpoints or criteria that users set for evaluating electronic documents.

[0638] "Means for adjusting the way information is presented" refers to processes or software functions for changing the format or content of the information provided depending on the perceived user's emotions or state.

[0639] This invention provides more flexible and personalized information by combining a system that uses a generative AI model to accurately grasp the content of electronic documents and assist in the process of evaluating their validity with an emotion engine that recognizes the user's emotions. Specific embodiments of this system are described below.

[0640] First, the server initializes the generative AI model. This generative AI model includes a "summary generation model" for summarizing text and a "text classification model" for classifying text. By initializing these models, the server prepares for the automatic processing of electronic documents. At the same time, the server also initializes an "emotion engine" for recognizing user emotions. This initialization process includes loading model parameters and configuration files.

[0641] Next, the user uploads the electronic document. The user uses the terminal interface to input the file path of the electronic document and sends it to the server. Specifically, the user can specify the file path as "path / to / medical_article.txt". This input information is received by the server.

[0642] The server uses the received file path to open the specified file. Using the standard Python library, the file is opened and the contents are read as text data. This allows the full text of the electronic document to be obtained as data.

[0643] The acquired text data is passed to a summary generation model, which extracts the key points of the text and generates a summary. For example, the model can generate a summary like this: "A study evaluated the efficacy and safety of drug A under specific conditions. The results, although condition dependent, showed the expected effects." This summary is then sent from the server to the user's device.

[0644] After viewing an abstract, users input specific verification criteria for evaluating the electronic literature, such as "study validity," "statistical accuracy," and "reproducibility of results." These verification criteria are then sent to the server.

[0645] The server uses the received validation points and the text data of the electronic document to analyze the document using a text classification model. It classifies the corresponding part for each validation point and obtains a classification result. For example, it can generate classification results such as "Research validity: Positive," "Statistical accuracy: Neutral," and "Reproducibility of results: Mixed." The classification results are then sent to the user's device, where the user can view the evaluation results for each validation point.

[0646] Furthermore, the emotion engine recognizes emotions based on the user's input and operation history. For example, the emotion engine detects whether the user is feeling stressed or interested. Based on the recognized emotional state, the server adjusts the way information is provided. For example, if the user is feeling stressed, the server can provide summaries and classification results in a more concise and easy-to-understand format. If the user is interested, the server can provide detailed information and additional reference materials.

[0647] As an example of a prompt sentence, a user might upload a medical paper called "path / to / medical_article.txt" to a server, and a summary generation model might generate the following summary: "The study evaluated the efficacy and safety of drug A under specific conditions. The results, although dependent on the conditions, showed the expected effects."

[0648] This allows users to efficiently evaluate the validity and reliability of medical papers and receive personalized information according to their emotional state.

[0649] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0650] Step 1:

[0651] The server initializes the generative AI model. Specifically, the server loads and initializes the summary generation model and text classification model. It also initializes the emotion engine to recognize the user's emotions. This includes loading the model parameters and configuration files. The model file and configuration file are required as input, and an initialized model object is generated as output.

[0652] Step 2:

[0653] The user uploads an electronic document. The user specifies the file path on the terminal interface and sends it to the server. Specifically, the user enters the file path "path / to / medical_article.txt". The file path is provided as input, and the file path information is sent to the server.

[0654] Step 3:

[0655] The server reads the electronic document. Based on the received file path, the server opens the specified file and reads the contents as text data. Depending on the programming language used, for example, Python's "open()" function is used. The file path is received as input, and the full text data of the electronic document is generated as output.

[0656] Step 4:

[0657] The server generates summaries of electronic literature. The text data is passed to a summary generation model, which extracts the main points of the text and generates a summary. Specifically, a generative AI model such as BERT is used to summarize the text data. The text data of the electronic literature is required as input, and summary text is generated as output. For example, a summary might be generated that reads, "A study evaluated the efficacy and safety of drug A under specific conditions. The results, while dependent on the conditions, showed the expected effects."

[0658] Step 5:

[0659] The server provides the generated summary to the user. The summary text is sent from the server to the user's terminal, and the user views the summary on the terminal interface. The generated summary text is required as input, and the summary text is displayed on the terminal screen as output.

[0660] Step 6:

[0661] The user registers validation points. The user enters multiple validation points into the terminal interface and sends them to the server. Specifically, the user enters items such as "research validity," "statistical accuracy," and "result reproducibility." A list of validation points is provided as input, and the information is sent to the server.

[0662] Step 7:

[0663] The server analyzes the electronic document. It uses a text classification model to classify the specified part of the electronic document and obtains classification results for each verification point. The input requires the text data of the electronic document and a list of verification points, and the output generates classification results for each verification point. For example, classification results such as "Research validity: Positive," "Statistical accuracy: Neutral," and "Reproducibility of results: Mixed" can be obtained.

[0664] Step 8:

[0665] The server provides the classification results to the user. The acquired classification results are sent from the server to the user's device, and the user can view the evaluation results of each verification point on the device interface. The classification results are required as input, and are displayed on the screen as output.

[0666] Step 9:

[0667] The emotion engine recognizes the user's emotions. The emotion engine analyzes emotions from the user's input and operation history, and detects emotional states such as stress and interest. Input data includes operation history such as input speed and click frequency. The user's emotional state is recognized as the output.

[0668] Step 10:

[0669] The server adjusts the way information is presented. The server adjusts the format of information presentation based on the recognized emotional state. For example, if the user is feeling stressed, the server will provide information in a concise and easy-to-understand format. The emotional state is required as input, and tailored information is provided as output.

[0670] (Application example 2)

[0671] Next, a description will be given of Application 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."

[0672] Problems often occur in the manufacturing process of modern factories, and dealing with them can take a lot of time and effort. It is also difficult to quickly and accurately grasp the large amount of information available in technical journals and specialized papers, and to implement appropriate countermeasures. Furthermore, dealing with problems often causes stress for on-site workers. In these situations, there is a growing need for a system that can efficiently summarize information and provide it while taking emotions into consideration.

[0673] 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.

[0674] In this invention, the server includes means for initializing a generative AI model, means for reading text data, means for passing the read text data to a summary generation model and generating a summary, means for providing the generated summary to a user, means for classifying each portion of the read text data using a text classification model and obtaining a classification result using multiple verification points received from the user, means for providing the obtained classification result to the user, means for initializing an emotion engine that recognizes the user's emotions, and means for adjusting a method of providing information based on the user's emotional state recognized by the emotion engine. This speeds up troubleshooting in factories and enables support that takes user emotions into consideration.

[0675] A "generative AI model" is an artificial intelligence model trained to perform tasks such as summarizing, classifying, and generating text data.

[0676] An "emotion engine" is software or hardware that recognizes a user's emotional state and adjusts the way information is presented.

[0677] "Text data" refers to data that consists of textual information, and includes technical journals and specialized papers.

[0678] A "summary generation model" is a machine learning model that extracts key points from long text data and summarizes them as short sentences.

[0679] A "text classification model" is an artificial intelligence model trained to classify text data into specific classes or categories.

[0680] "Verification points" are points of view that the user inputs as evaluation items for text data, and include, for example, "reproducibility of the problem" and "appropriateness of the countermeasures."

[0681] "Adjusting the method of providing information" refers to optimizing the content and format of the information provided based on the user's emotional state.

[0682] This invention relates to a system that utilizes a generative AI model to efficiently summarize technical journals and specialized papers related to manufacturing processes in factories, and further recognizes user emotions to optimize information provision. Specific embodiments of this system are described below.

[0683] The server initializes the generated AI model

[0684] The server initializes the generative AI model. This generative AI model includes a summary generation model for summarizing text data and a text classification model for classifying text data. These models are initialized using the HuggingFace transformers library. An emotion engine that recognizes the user's emotions is also initialized at the same time, and OpenCV is used for this emotion engine.

[0685] Factory workers upload daily reports or technical papers to the robot

[0686] A factory worker inputs the file path of the daily report or technical paper through the terminal interface and sends it to the server.

[0687] The server reads the text data.

[0688] The server opens the specified file and reads it as text data, which may include the full text of technical journals or technical papers.

[0689] Server generates summary

[0690] The server passes the text data it has read to a summary generation model, which then generates a summary by extracting the main points of the text.

[0691] The server provides the summary

[0692] The server provides the generated summary to the factory staff, who can view the summary through the terminal interface and quickly understand the contents.

[0693] Factory staff register verification points

[0694] Factory staff input verification points for evaluating technical daily reports and specialist papers and send them to the server. These verification points include "reproducibility of the problem," "appropriateness of countermeasures," and "scope of impact."

[0695] The server analyzes the text data

[0696] The server uses the verification points and text data received from the factory staff to perform analysis using a text classification model. The server classifies the corresponding part for each verification point and obtains the classification result.

[0697] The server provides the classification results

[0698] The server provides the classification results for each verification point to factory workers, who can view the evaluation results for each verification point through the terminal interface.

[0699] Recognize user emotions

[0700] The emotion engine recognizes the emotions of factory workers based on their input, operation history, and facial expressions, allowing the system to understand their current emotional state.

[0701] Emotion-based adjustment of information provision

[0702] The server adjusts the way it presents information based on the emotional state of the factory worker. For example, if the factory worker is feeling stressed, it will provide a more concise summary and classification results in an easy-to-understand format. If the factory worker is interested, it will provide detailed information or additional reference materials.

[0703] As a concrete example, consider a scenario in which a factory worker uploads a daily technical report called " / reports / machine_failure_report.txt" and the server generates the following summary: "In this daily report, machine X experienced motor overheating under certain operating conditions. As a countermeasure, the cooling system was adjusted." The factory worker also registers verification points such as "reproducibility of the problem," "appropriateness of the countermeasure," and "scope of impact," and the server provides the following classification results: "reproducibility of the problem: high," "appropriateness of the countermeasure: good," and "scope of impact: limited." The emotion engine can detect the stress level of factory workers and provide more concise information.

[0704] An example of a prompt is as follows:

[0705] "Please summarize the following technical daily report: / path / to / machine_failure_report.txt. In addition, please provide a classification and analysis based on the following verification points: 'Reproducibility of the problem', 'Appropriateness of the countermeasures', 'Scope of impact'. Please consider the feelings of the factory staff and simplify the information if stress is felt."

[0706] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0707] Step 1:

[0708] The server initializes the generative AI models, which include two models: a summary generation model and a text classification model. The server initializes these models using HuggingFace's transformers library. As input, it takes the URL or path of a pre-trained model. As output, it gets the initialized model object.

[0709] Step 2:

[0710] The user inputs the file path of the daily report or technical paper through the terminal interface and sends it to the server. The input is the file path (e.g., " / reports / machine_failure_report.txt"), and the output is that the sent file path reaches the server.

[0711] Step 3:

[0712] Based on the file path received by the server, the specified file is opened and read as text data. The input is the file path, and the output is text data containing the file contents.

[0713] Step 4:

[0714] The server passes the text data it has read to a summary generation model to generate a summary. Here, the summary generation model extracts the main points of the text data and summarizes them as short sentences. The input is the text data, and the output is the generated summary text.

[0715] Step 5:

[0716] The server provides the generated summary to the user by sending data to display the summary text on the user's terminal. The input is the summary text, and the output is the summary content displayed on the user interface.

[0717] Step 6:

[0718] The user inputs verification points for evaluating technical journals and technical papers into the terminal interface and sends them to the server. The inputs are verification points (e.g., "reproducibility of the problem," "appropriateness of the countermeasures," "scope of impact"), and the output is the verification points that reach the server.

[0719] Step 7:

[0720] The server uses the verification points and text data received from the user to perform analysis using a text classification model. The server classifies the corresponding text portion for each verification point and obtains the classification result. The input is the text data and verification points, and the output is the classification result for each verification point.

[0721] Step 8:

[0722] The server provides the user with the classification results for each verification point it has acquired. Here, it sends data to display the classification results on the user's device. The input is the classification results, and the output is the evaluation content displayed on the user interface.

[0723] Step 9:

[0724] The server uses an emotion engine to recognize the user's emotions based on the user's input, operation history, and facial expressions captured by the device's camera. The input is the user's operation history and camera footage, and the output is the recognized emotional state of the user.

[0725] Step 10:

[0726] The server adjusts the way it presents information based on the user's recognized emotional state. For example, if the user is feeling stressed, it will provide more concise information in an easy-to-understand format. The input is the user's emotional state, and the output is the adjusted information presentation data.

[0727] 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.

[0728] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) 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.

[0729] 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.

[0730] [Third embodiment]

[0731] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0732] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0733] 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).

[0734] 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.

[0735] 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.

[0736] 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).

[0737] 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.

[0738] 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.

[0739] 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.

[0740] 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.

[0741] 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.

[0742] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0743] This invention is a system that uses a generative AI model to accurately grasp the content of medical papers and support the process of determining their validity. Specific embodiments of this system are described below.

[0744] The server initializes the generated AI model

[0745] First, the server initializes the generative AI model. This generative AI model includes a summary generation model for performing text summarization and a text classification model for classifying text. By initializing these models, the server prepares to process medical papers.

[0746] User uploads medical paper

[0747] Next, the user uploads a medical paper. The user inputs the file path of the medical paper and sends it to the server. This file path is used when the server reads the text of the medical paper.

[0748] The server loads medical papers

[0749] The server reads the medical paper uploaded by the user. The server opens the paper file and reads its contents as text data. This text data is then passed to the summary generation model.

[0750] Server generates summary

[0751] The server passes the text data of medical papers read into a summary generation model, which then generates a summary. This summary generation model extracts important information from long sentences and outputs a short summary. The generated summary is then provided to the user.

[0752] User registers verification point

[0753] Users register validation criteria for evaluating medical papers, including the validity of the research, statistical accuracy, and reproducibility of results, among other aspects that the user wishes to evaluate.

[0754] The server analyzes the paper

[0755] Based on the verification points received from the user, the server analyzes each part of the medical paper using a text classification model. The server uses each verification point and the text data of the paper to obtain a classification result. This classification model is trained to classify the input text into a specific category.

[0756] The server provides the classification results

[0757] The server provides the user with the classification results for each verification point, allowing the user to check the evaluation results for each aspect of the medical paper.

[0758] Specific examples

[0759] A specific example is given below.

[0760] 1. User uploads paper file

[0761] Example: A user uploads a medical article called "path / to / medical_article.txt" to the server.

[0762] 2. The server generates the summary

[0763] Example: The server reads the content of a paper and uses a summary generation model to generate the following summary:

[0764] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, showed the expected effects."

[0765] 3. User registers verification points

[0766] Example: A user registers verification points such as "research validity," "statistical accuracy," and "reproducibility of results."

[0767] 4. The server analyzes the paper

[0768] Example: The server runs a text classification model on each validation point and obtains the following classification results:

[0769] Study validity: "Positive"

[0770] Statistical accuracy: "Neutral"

[0771] Reproducibility of results: "Mixed"

[0772] In this way, by using this system, medical and pharmaceutical experts can efficiently and accurately evaluate and verify a huge amount of medical research papers.

[0773] The processing flow will be explained below.

[0774] Step 1:

[0775] The server initializes the generative AI model, specifically, loads the summary generation model and text classification model, and deploys them to storage devices to efficiently perform natural language processing tasks.

[0776] Step 2:

[0777] The user inputs the file path of the medical paper and sends it to the server. The file path is input via the interface of the terminal used by the user.

[0778] Step 3:

[0779] The server receives the file path and opens the specified file. The server reads the contents of the paper file as text data. This text data includes the full text of the medical paper.

[0780] Step 4:

[0781] The server passes the read text data to a summary generation model. The model extracts the main points of the text and generates a summary. The generated summary is short and concise, describing the main points of the paper.

[0782] Step 5:

[0783] The server provides the generated abstract to the user, who can then view the abstract through a terminal interface to quickly understand the content of the paper.

[0784] Step 6:

[0785] Users input the validation points used to evaluate papers and submit them to the server. These validation points include the validity of the research, statistical accuracy, and reproducibility of results.

[0786] Step 7:

[0787] The server uses the verification points received from the user and the text data of the medical paper to analyze the paper using a text classification model. For each verification point, the corresponding part is classified and a classification result is obtained.

[0788] Step 8:

[0789] The server provides the classification results to the user, who can then view the evaluation results for each verification point via the terminal interface, thereby confirming the analysis results for each aspect of the medical paper.

[0790] Step 9:

[0791] Users evaluate the validity and reliability of medical papers based on the provided abstracts and classification results. The server records this evaluation process and stores it in a database for future reference and analysis.

[0792] Example 1

[0793] Next, a description will be given of 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."

[0794] Efficiently evaluating and verifying large volumes of pharmaceutical research papers is a major challenge for medical and pharmaceutical experts. In particular, there is a need to quickly and accurately evaluate the validity, statistical accuracy, and reproducibility of results of research papers. However, conventional methods require a huge amount of time and effort to analyze each paper in detail, making them inefficient. Furthermore, results can vary depending on the evaluator's subjectivity, making it difficult to perform consistent evaluations.

[0795] 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.

[0796] In this invention, the server includes means for initializing the generative AI model, means for a user to upload a medical paper, means for reading the medical paper, means for passing the text data of the read medical paper to a summary generation model to generate a summary, means for providing the generated summary to the user, means for registering multiple verification points received from the user, means for classifying each part of the read medical paper using the multiple verification points received from the user with a text classification model to obtain the classification results, and means for providing the obtained classification results to the user, thereby enabling efficient summary generation of medical papers and automatic and consistent evaluation from multiple perspectives.

[0797] A "generative AI model" is a trained model that uses artificial intelligence to summarize and classify text.

[0798] The "server" is the computer system that serves as the core of the system, initializing the generative AI model, processing medical papers, and interfacing with users.

[0799] A "user" is a person or organization that uses the system to upload medical papers and receive their summaries and evaluation results.

[0800] A "medical paper" is a piece of writing that summarizes research results and findings related to medicine and pharmacy.

[0801] A "summary generation model" is a type of generative AI model used to extract important information from medical papers and generate short summaries.

[0802] A "text classification model" is a type of generative AI model used to classify parts of medical papers into specific categories.

[0803] "Verification points" are items that serve as viewpoints or criteria when a user evaluates a medical paper.

[0804] "Classification results" are evaluation results for each part of a medical paper generated by a text classification model.

[0805] "Upload" is the act of a user sending data from their own terminal to a server.

[0806] A "web interface" is a screen that uses a web browser to allow a user to access and operate the system.

[0807] This invention is a system that uses a generative AI model to efficiently summarize the contents of medical papers and assists in the process of determining their validity based on each verification point. Specific embodiments are described below.

[0808] 1. Hardware and Software Configuration

[0809] The server plays a central role in this system. The server hardware includes a powerful CPU, sufficient memory, and large storage capacity. The main software used includes Python, the Transformers library, and a suitable database management system.

[0810] 2. Initializing the Generative AI Model

[0811] When the system is initialized, the server loads generative AI models from disk. Specifically, it loads a "summary generation model" that generates summaries of medical papers and a "text classification model" that classifies text. These models have been trained in advance using large amounts of text data, and the server initializes their parameters and loads them into memory.

[0812] 3. User upload of medical papers

[0813] Users upload medical papers through the system's web interface. They use a file selection dialog to select the medical paper file they want to evaluate and click the upload button. The terminal sends the selected file path to the server, which stores it in a temporary directory.

[0814] 4. Reading medical papers

[0815] The server uses the file path sent by the user to open the uploaded medical paper, reads the file contents as text data, and stores it in memory. This text data is used in the subsequent summary generation process.

[0816] 5. Summary Generation

[0817] The server passes the text data of the loaded medical paper to a summary generation model to generate a summary. This summary generation model extracts important information from the medical paper and outputs a short summary. The generated summary is provided to users so that they can easily understand the outline of the paper.

[0818] 6. Registering Verification Points

[0819] Users register multiple validation points to evaluate the validity of a paper. These include aspects such as "research validity," "statistical accuracy," and "reproducibility of results." These validation points are entered on a web interface and sent to the server, which stores them in a database.

[0820] 7. Analysis of the paper

[0821] The server applies a text classification model using the verification points received from the user and the text data loaded, generating a classification result for each verification point and storing it in an internal data structure.

[0822] 8. Providing classification results

[0823] The server takes the generated classification results, converts them into a user-friendly format, and provides the results to the user through a web interface.

[0824] Specific examples

[0825] Initializing a generative AI model

[0826] Example: The server starts and loads a "summary generation model" and a "text classification model" using common natural language processing models.

[0827] User-uploaded medical papers

[0828] Example: User uploads "path / to / medical_article.txt" through the web interface.

[0829] Reading medical papers

[0830] Example: The server opens "medical_article.txt" from the specified path and reads its contents as text data.

[0831] Summary Generation

[0832] Example: The server generates the following summary using the summary generation model:

[0833] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, generally showed the expected effects."

[0834] Registering verification points

[0835] Example: A user enters verification points such as "study validity," "statistical accuracy," and "result reproducibility" into a web form.

[0836] Analysis of the paper

[0837] Example: The server applies a text classification model to each validation point and produces the following classification results:

[0838] Study validity: "Positive"

[0839] Statistical accuracy: "Neutral"

[0840] Reproducibility of results: "Mixed"

[0841] Providing classification results

[0842] Example: Displaying the analysis results in a table format in a web interface.

[0843] This system allows medical and pharmaceutical experts to efficiently evaluate and verify vast amounts of medical research.

[0844] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0845] Step 1: Initializing the generative AI model

[0846] Specific behavior:

[0847] When the system starts up, the server loads the generative AI model from disk, loads a "summary generation model" and a "text classification model" for analyzing medical papers from around the world, and initializes them.

[0848] Input and Output: The input is the generative AI model file on disk, and the output is the initialized generative AI model (in memory).

[0849] Step 2: User uploads medical papers

[0850] Specific behavior:

[0851] The user opens the system's web interface, selects a medical paper file using the file selection dialog, and clicks the upload button. The terminal sends the selected file path to the server, which stores it in a temporary directory.

[0852] Input and Output: The input is the file path of the medical paper selected by the user, and the output is the paper file stored in the server's storage.

[0853] Step 3: Reading medical papers

[0854] Specific behavior:

[0855] The server uses the file path of the uploaded medical paper to open the paper file, read the contents of the file as text data, and save it in memory. It then records the first few lines of the read text data in a log to confirm that it was read successfully.

[0856] Input and output: The input is a paper file stored in the server's storage, and the output is text data stored in memory.

[0857] Step 4: Summary generation

[0858] Specific behavior:

[0859] The server passes the text data of medical papers stored in memory to a summary generation model, which processes the text data to generate a short summary, which is then sent to a web interface for presentation to the user.

[0860] Input and Output: The input is the text data stored in memory, and the output is the generated summary sentence.

[0861] Step 5: Register your verification points

[0862] Specific behavior:

[0863] The user enters the verification points in a form on the web interface and sends them to the server, which stores them in a database.

[0864] Input and Output: The input is the verification point entered by the user and the output is the verification point stored in the database.

[0865] Step 6: Analyzing the paper

[0866] Specific behavior:

[0867] The server applies a text classification model to the medical text data stored in memory and the validation points received from the user, generating a classification result for each validation point and storing it in an internal data structure.

[0868] Input and Output: The input is the text data stored in memory and the validation points stored in the database, and the output is the classification results stored in the internal data structure.

[0869] Step 7: Providing classification results

[0870] Specific behavior:

[0871] The server retrieves the classification results stored in an internal data structure, converts them into a user-friendly format, and sends the converted analysis results to a web interface for display to the user.

[0872] Input and Output: The input is the classification results stored in an internal data structure, and the output is the analysis results displayed in a web interface.

[0873] (Application example 1)

[0874] Next, a description will be given of Application 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."

[0875] Traditional logistics centers are required to manage a huge number of shipping lists and inventory reports, but they lack the means to process them efficiently.In addition, verifying and accurately evaluating the contents of the reports requires a lot of time and effort, and there is a risk of human error.

[0876] 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.

[0877] In this invention, the server includes means for initializing a generative AI model, means for reading a document, means for passing the text data of the read document to a summary generation model and generating a summary, means for providing the generated summary to a user, means for classifying each part of the read document using a text classification model using multiple verification points received from the user and obtaining the classification results, and means for providing the obtained classification results to the user. This enables efficient and accurate processing and evaluation of documents such as shipping lists and inventory reports, even in logistics centers.

[0878] A "generative AI model" is an algorithmic model designed to use artificial intelligence to perform tasks such as text generation, summarization, and classification.

[0879] A "document" is a collection of information, including text data, recorded on paper or in digital form.

[0880] "Means of reading" refers to the process of inputting a specified document file into the system and capturing its contents as digital data.

[0881] A "summary generation model" is an algorithmic model that extracts important information from long pieces of text and summarizes it concisely.

[0882] The "means for generating a summary" refers to a process for generating a summary from the input text data using a summary generation model.

[0883] "Means for providing" refers to the process for displaying or communicating the generated information or classification results to the user.

[0884] "Verification points" refer to specific perspectives or criteria that users evaluate.

[0885] A "text classification model" is an algorithmic model for classifying input text data into specific categories.

[0886] The "means for obtaining classification results" refers to the process of using a text classification model to determine the category of text data for each particular validation point and obtain a result.

[0887] A "smart device" refers to an electronic device that has internet connectivity, computing power, and can run applications.

[0888] This invention relates to a system that uses generative AI models to efficiently manage shipping lists and inventory reports within a logistics center and accurately evaluate their contents. This system allows managers and staff to acquire, analyze, and evaluate information using smart devices, thereby improving work efficiency and accuracy.

[0889] The server first initializes the generative AI models, which include a summary generation model and a text classification model. The summary generation model is used to extract important information from long texts and provide them as short summaries, while the text classification model is used to classify texts into specific categories.

[0890] Users upload logistics documents such as shipping lists and inventory reports, which are then loaded onto the server, where they are converted into text format and prepared for analysis.

[0891] The server then passes the read text data to a summary generation model to generate a summary. This summary is automatically generated in a short sentence format, allowing users to quickly grasp the content. The generated summary is then provided to the user via their smart device.

[0892] Users can set multiple evaluation points for shipping and inventory management. These verification points include evaluation criteria such as "shipping accuracy," "inventory accuracy," and "report validity."

[0893] The server analyzes the loaded text data using a text classification model based on the verification points received from the user. As a result of the analysis, a classification result is obtained for each verification point. This classification result is provided to the user via their smart device, and the user can check the evaluation results for each verification point.

[0894] The hardware used includes servers, smartphones, smart glasses, or logistics robots, while the software used is primarily Python programs and the Hugging Face transformers library.

[0895] For example, let's take a concrete example of how the system is used within a logistics center. When a user uploads a shipping report called "path / to / shipping_report.txt" to the system, the server reads the contents of the report and generates a summary. The summary is provided in the form of, "This shipping report describes the shipping status within a specific period, and shows that most tasks are being carried out according to schedule." Next, when the user registers verification points such as "Shipping accuracy," "Inventory accuracy," and "Report validity," the server classifies each verification point and produces classification results such as "Shipping accuracy: Positive," "Inventory accuracy: Neutral," and "Report validity: Positive."

[0896] Example prompt sentence:

[0897] "Design a system to assist distribution centers in the process of accurately capturing and validating the contents of shipping lists and inventory reports. The system will utilize a generative AI model to generate summaries and provide classification results based on validation points. Specifically, we will use a Python program and the Hugging Face transformers library."

[0898] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0899] Step 1:

[0900] The server initializes the generative AI model. Specifically, it initializes a summary generation model for summarizing text data and a text classification model for classifying text. As input, the configuration file and initialization data for the generative AI model are passed. As output, the launched summary generation model and text classification model are obtained.

[0901] Step 2:

[0902] The user uploads a shipping list or inventory report through a terminal. Specifically, the user enters the file path specified by the user and sends it to the server. The input includes the file path information selected by the user. The output is the file path information sent to the server.

[0903] Step 3:

[0904] The server reads the uploaded document file. Specifically, it opens the specified file path and reads its contents as text data. The input includes the file path information sent by the user. The output is the read text data.

[0905] Step 4:

[0906] The server passes text data to a summary generation model to generate a summary. Specifically, the text data is input to the summary generation model, and the model generates a summary. The input includes the read text data. The output is the generated summary text.

[0907] Step 5:

[0908] The server provides the generated summary to the user. Specifically, it sends the summary text to the terminal so that the user can view it on their smart device. The input includes the generated summary text. The output is display data that the user can view on their terminal.

[0909] Step 6:

[0910] The user sets the verification points for evaluation. Specifically, the user sends the aspects they want to evaluate (e.g., "shipping accuracy," "inventory accuracy," "report validity") to the server via their terminal. The input includes the verification point information selected by the user. The output is the verification point information sent to the server.

[0911] Step 7:

[0912] The server analyzes the text data loaded by the server using a text classification model based on the verification points. Specifically, the text data is classified for each verification point using the verification point information. The input includes the loaded text data and verification point information. The output is a classification result for each verification point.

[0913] Step 8:

[0914] The classification results acquired by the server are provided to the user. Specifically, the classification results are sent to the terminal, allowing the user to check the evaluation results for each perspective on their smart device. The input includes the classification results. The output is display data that the user can check on their terminal.

[0915] 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.

[0916] This invention provides more flexible and personalized information by combining a system that uses a generative AI model to accurately grasp the content of medical papers and assist in the process of determining their validity with an emotion engine that recognizes the user's emotions. Specific embodiments of this system are described below.

[0917] The server initializes the generated AI model

[0918] First, the server initializes the generative AI model. This generative AI model includes a summary generation model for performing text summarization and a text classification model for classifying text. By initializing these models, the server prepares for the automatic processing of medical papers. At the same time, an emotion engine for recognizing user emotions is also initialized.

[0919] User uploads medical paper

[0920] Next, the user uploads the medical paper. The user inputs the file path of the medical paper and sends it to the server. The file path is input via the interface of the terminal used by the user.

[0921] The server loads medical papers

[0922] The server receives the file path and opens the specified file. The server reads the contents of the paper file as text data. This text data includes the full text of the medical paper.

[0923] Server generates summary

[0924] The server passes the text data it has read to a summary generation model, which then generates a summary. The summary generation model extracts the main points of the text and generates a summary. The generated summary is short and concise, describing the main points of the paper.

[0925] The server provides the summary

[0926] The server provides the generated abstract to the user, who can then view the abstract through a terminal interface to quickly understand the content of the paper.

[0927] User registers verification point

[0928] Users input verification criteria for evaluating medical papers and submit them to the server. These verification criteria include the validity of the research, statistical accuracy, and reproducibility of the results.

[0929] The server analyzes the paper

[0930] The server uses the verification points received from the user and the text data of the medical paper to analyze the paper using a text classification model, and classifies the corresponding part for each verification point to obtain the classification result.

[0931] The server provides the classification results

[0932] The server provides the user with the classification results for each verification point. The user can view the evaluation results for each verification point through the terminal interface, allowing them to check the analysis results for each aspect of the medical paper.

[0933] Recognize user emotions

[0934] The emotion engine recognizes the user's emotions based on their input and operation history. For example, the emotion engine can detect when the user is feeling stressed or interested. This allows the system to understand the user's current emotional state.

[0935] Emotion-based adjustment of information provision

[0936] The server can then adapt the way it presents information based on the user's perceived emotional state. For example, if the user is feeling stressed, it can provide a more concise summary or classification in an easy-to-understand format. It can also provide more detailed information or additional references if the user is interested.

[0937] Specific examples

[0938] 1. User uploads paper file

[0939] Example: A user uploads a medical article called "path / to / medical_article.txt" to the server.

[0940] 2. The server generates the summary

[0941] Example: The server reads the content of a paper and uses a summary generation model to generate the following summary:

[0942] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, showed the expected effects."

[0943] 3. User registers verification points

[0944] Example: A user registers verification points such as "research validity," "statistical accuracy," and "reproducibility of results."

[0945] 4. The server analyzes the paper

[0946] Example: The server runs a text classification model on each validation point and obtains the following classification results:

[0947] Study validity: "Positive"

[0948] Statistical accuracy: "Neutral"

[0949] Reproducibility of results: "Mixed"

[0950] 5. Emotion recognition and information provision coordination

[0951] Example: If the emotion engine detects a stressed state from the user's operation history, the server will present the information more concisely, making it easier to understand.

[0952] This system allows users to efficiently evaluate the validity and reliability of medical papers and receive personalized information based on the user's emotional state.

[0953] The processing flow will be explained below.

[0954] Step 1:

[0955] The server initializes the generative AI model and emotion engine. Specifically, it loads the summary generation model, text classification model, and emotion engine that recognizes the user's emotions into memory and makes them ready for use.

[0956] Step 2:

[0957] The user inputs the file path of the medical paper using the terminal and sends it to the server. The file path indicates the location information of the medical paper specified by the user.

[0958] Step 3:

[0959] The server opens the specified medical paper file based on the file path received from the user. The server reads the contents of the medical paper as text data and stores it in memory.

[0960] Step 4:

[0961] The server inputs the text data into a summary generation model to generate a summary, which is a short sentence that succinctly summarizes the main points of the paper.

[0962] Step 5:

[0963] The server sends the generated abstract to the user's terminal, and the user can view the abstract through the terminal interface. The user can refer to the abstract and quickly understand the content of the paper.

[0964] Step 6:

[0965] Users input the validation points used to evaluate medical papers through their terminals and send them to the server. These validation points include aspects that users want to evaluate, such as the validity of the research, statistical accuracy, and reproducibility of results.

[0966] Step 7:

[0967] The server uses the verification points received from the user and the text data of the medical paper it has read to analyze the paper using a text classification model. For each verification point, the model classifies the corresponding part and obtains the classification result.

[0968] Step 8:

[0969] The server sends the classification results for each verification point to the user's device, and the user can view the classification results through the device's interface, allowing the user to check the evaluation results of the paper for each perspective.

[0970] Step 9:

[0971] The emotion engine recognizes the user's emotional state from their input, operation history, reactions, etc. For example, if a user browses a lot of literature in a short period of time, the emotion engine can determine that the user is in a stressful state.

[0972] Step 10:

[0973] The server adjusts the way it presents information based on the user's emotional state. For example, if the user is feeling stressed, it will adjust the way it presents information by making summaries and classification results more concise and easy to understand.

[0974] Step 11:

[0975] The server then sends the adjusted information to the user's device, where the user can view the adjusted summary and classification results through the device interface. The adjusted information is optimized to fit the user's emotional state.

[0976] Step 12:

[0977] Users can efficiently evaluate the validity and reliability of medical papers based on the provided abstracts and classification results, and the server records this evaluation process and stores it in a database for future reference and analysis.

[0978] In this way, a system that combines an emotion engine can accurately grasp the content of medical papers and provide information that corresponds to the user's emotional state, thereby providing personalized support that meets the user's needs.

[0979] Example 2

[0980] Next, a description will be given of 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."

[0981] Modern medical papers contain a vast amount of information, making it extremely difficult to quickly and accurately grasp and evaluate its validity. Furthermore, because information is not provided in response to the user's emotions or state, there are concerns that this will increase the burden on researchers. There is a need to solve these problems and provide an environment in which users can more efficiently and comfortably evaluate and understand medical papers.

[0982] 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.

[0983] In this invention, the server includes means for initializing the generation AI model, means for reading electronic documents, and means for passing the text data of the read electronic documents to the summary generation model and generating summaries, thereby enabling users to quickly grasp the vast amount of information in electronic documents and providing personalized information according to the user's emotional state.

[0984] A "generative AI model" refers to an artificial intelligence model for summarizing and classifying text data.

[0985] A "summary generation model" refers to an artificial intelligence model that extracts the main points of input text data and generates a concise summary.

[0986] "Text classification model" refers to an artificial intelligence model for classifying input text data based on specific criteria.

[0987] "Electronic literature" refers to text data such as papers and articles stored in digital format.

[0988] "User" refers to a person who uses this system and operates it for the purpose of understanding and evaluating the content of electronic documents.

[0989] An "emotion engine" refers to software or a system that recognizes emotions and states based on user input and operation history.

[0990] "Verification points" refer to specific viewpoints or criteria that users set for evaluating electronic documents.

[0991] "Means for adjusting the way information is presented" refers to processes or software functions for changing the format or content of the information provided depending on the perceived user's emotions or state.

[0992] This invention provides more flexible and personalized information by combining a system that uses a generative AI model to accurately grasp the content of electronic documents and assist in the process of evaluating their validity with an emotion engine that recognizes the user's emotions. Specific embodiments of this system are described below.

[0993] First, the server initializes the generative AI model. This generative AI model includes a "summary generation model" for summarizing text and a "text classification model" for classifying text. By initializing these models, the server prepares for the automatic processing of electronic documents. At the same time, the server also initializes an "emotion engine" for recognizing user emotions. This initialization process includes loading model parameters and configuration files.

[0994] Next, the user uploads the electronic document. The user uses the terminal interface to input the file path of the electronic document and sends it to the server. Specifically, the user can specify the file path as "path / to / medical_article.txt". This input information is received by the server.

[0995] The server uses the received file path to open the specified file. Using the standard Python library, the file is opened and the contents are read as text data. This allows the full text of the electronic document to be obtained as data.

[0996] The acquired text data is passed to a summary generation model, which extracts the key points of the text and generates a summary. For example, the model can generate a summary like this: "A study evaluated the efficacy and safety of drug A under specific conditions. The results, although condition dependent, showed the expected effects." This summary is then sent from the server to the user's device.

[0997] After viewing an abstract, users input specific verification criteria for evaluating the electronic literature, such as "study validity," "statistical accuracy," and "reproducibility of results." These verification criteria are then sent to the server.

[0998] The server uses the received validation points and the text data of the electronic document to analyze the document using a text classification model. It classifies the corresponding part for each validation point and obtains a classification result. For example, it can generate classification results such as "Research validity: Positive," "Statistical accuracy: Neutral," and "Reproducibility of results: Mixed." The classification results are then sent to the user's device, where the user can view the evaluation results for each validation point.

[0999] Furthermore, the emotion engine recognizes emotions based on the user's input and operation history. For example, the emotion engine detects whether the user is feeling stressed or interested. Based on the recognized emotional state, the server adjusts the way information is provided. For example, if the user is feeling stressed, the server can provide summaries and classification results in a more concise and easy-to-understand format. If the user is interested, the server can provide detailed information and additional reference materials.

[1000] As an example of a prompt sentence, a user might upload a medical paper called "path / to / medical_article.txt" to a server, and a summary generation model might generate the following summary: "The study evaluated the efficacy and safety of drug A under specific conditions. The results, although dependent on the conditions, showed the expected effects."

[1001] This allows users to efficiently evaluate the validity and reliability of medical papers and receive personalized information according to their emotional state.

[1002] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1003] Step 1:

[1004] The server initializes the generative AI model. Specifically, the server loads and initializes the summary generation model and text classification model. It also initializes the emotion engine to recognize the user's emotions. This includes loading the model parameters and configuration files. The model file and configuration file are required as input, and an initialized model object is generated as output.

[1005] Step 2:

[1006] The user uploads an electronic document. The user specifies the file path on the terminal interface and sends it to the server. Specifically, the user enters the file path "path / to / medical_article.txt". The file path is provided as input, and the file path information is sent to the server.

[1007] Step 3:

[1008] The server reads the electronic document. Based on the received file path, the server opens the specified file and reads the contents as text data. Depending on the programming language used, for example, Python's "open()" function is used. The file path is received as input, and the full text data of the electronic document is generated as output.

[1009] Step 4:

[1010] The server generates summaries of electronic literature. The text data is passed to a summary generation model, which extracts the main points of the text and generates a summary. Specifically, a generative AI model such as BERT is used to summarize the text data. The text data of the electronic literature is required as input, and summary text is generated as output. For example, a summary might be generated that reads, "A study evaluated the efficacy and safety of drug A under specific conditions. The results, while dependent on the conditions, showed the expected effects."

[1011] Step 5:

[1012] The server provides the generated summary to the user. The summary text is sent from the server to the user's terminal, and the user views the summary on the terminal interface. The generated summary text is required as input, and the summary text is displayed on the terminal screen as output.

[1013] Step 6:

[1014] The user registers validation points. The user enters multiple validation points into the terminal interface and sends them to the server. Specifically, the user enters items such as "research validity," "statistical accuracy," and "result reproducibility." A list of validation points is provided as input, and the information is sent to the server.

[1015] Step 7:

[1016] The server analyzes the electronic document. It uses a text classification model to classify the specified part of the electronic document and obtains classification results for each verification point. The input requires the text data of the electronic document and a list of verification points, and the output generates classification results for each verification point. For example, classification results such as "Research validity: Positive," "Statistical accuracy: Neutral," and "Reproducibility of results: Mixed" can be obtained.

[1017] Step 8:

[1018] The server provides the classification results to the user. The acquired classification results are sent from the server to the user's device, and the user can view the evaluation results of each verification point on the device interface. The classification results are required as input, and are displayed on the screen as output.

[1019] Step 9:

[1020] The emotion engine recognizes the user's emotions. It analyzes emotions from the user's input and operation history, and detects emotional states such as stress and interest. Input data includes operation history such as input speed and click frequency. The user's emotional state is recognized as the output.

[1021] Step 10:

[1022] The server adjusts the way information is presented. The server adjusts the format of information presentation based on the recognized emotional state. For example, if the user is feeling stressed, the server will provide information in a concise and easy-to-understand format. The emotional state is required as input, and tailored information is provided as output.

[1023] (Application example 2)

[1024] Next, a description will be given of Application 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."

[1025] Problems often occur in the manufacturing process of modern factories, and dealing with them can take a lot of time and effort. It is also difficult to quickly and accurately grasp the large amount of information available in technical journals and specialized papers, and to implement appropriate countermeasures. Furthermore, dealing with problems often causes stress for on-site workers. In these situations, there is a growing need for a system that can efficiently summarize information and provide it while taking emotions into consideration.

[1026] 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.

[1027] In this invention, the server includes means for initializing a generative AI model, means for reading text data, means for passing the read text data to a summary generation model and generating a summary, means for providing the generated summary to a user, means for classifying each portion of the read text data using a text classification model and obtaining a classification result using multiple verification points received from the user, means for providing the obtained classification result to the user, means for initializing an emotion engine that recognizes the user's emotions, and means for adjusting a method of providing information based on the user's emotional state recognized by the emotion engine. This speeds up troubleshooting in factories and enables support that takes user emotions into consideration.

[1028] A "generative AI model" is an artificial intelligence model trained to perform tasks such as summarizing, classifying, and generating text data.

[1029] An "emotion engine" is software or hardware that recognizes a user's emotional state and adjusts the way information is presented.

[1030] "Text data" refers to data that consists of textual information, and includes technical journals and specialized papers.

[1031] A "summary generation model" is a machine learning model that extracts key points from long text data and summarizes them as short sentences.

[1032] A "text classification model" is an artificial intelligence model trained to classify text data into specific classes or categories.

[1033] "Verification points" are points of view that the user inputs as evaluation items for text data, and include, for example, "reproducibility of the problem" and "appropriateness of the countermeasures."

[1034] "Adjusting the method of providing information" refers to optimizing the content and format of the information provided based on the user's emotional state.

[1035] This invention relates to a system that utilizes a generative AI model to efficiently summarize technical journals and specialized papers related to manufacturing processes in factories, and further recognizes user emotions to optimize information provision. Specific embodiments of this system are described below.

[1036] The server initializes the generated AI model

[1037] The server initializes the generative AI model. This generative AI model includes a summary generation model for summarizing text data and a text classification model for classifying text data. These models are initialized using the HuggingFace transformers library. An emotion engine that recognizes the user's emotions is also initialized at the same time, and OpenCV is used for this emotion engine.

[1038] Factory workers upload daily reports or technical papers to the robot

[1039] A factory worker inputs the file path of the daily report or technical paper through the terminal interface and sends it to the server.

[1040] The server reads the text data.

[1041] The server opens the specified file and reads it as text data, which may include the full text of technical journals or technical papers.

[1042] Server generates summary

[1043] The server passes the text data it has read to a summary generation model, which then generates a summary by extracting the main points of the text.

[1044] The server provides the summary

[1045] The server provides the generated summary to the factory staff, who can view the summary through the terminal interface and quickly understand the contents.

[1046] Factory staff register verification points

[1047] Factory staff input verification points for evaluating technical daily reports and specialist papers and send them to the server. These verification points include "reproducibility of the problem," "appropriateness of countermeasures," and "scope of impact."

[1048] The server analyzes the text data

[1049] The server uses the verification points and text data received from the factory staff to perform analysis using a text classification model. The server classifies the corresponding part for each verification point and obtains the classification result.

[1050] The server provides the classification results

[1051] The server provides the classification results for each verification point to factory workers, who can view the evaluation results for each verification point through the terminal interface.

[1052] Recognize user emotions

[1053] The emotion engine recognizes the emotions of factory workers based on their input, operation history, and facial expressions, allowing the system to understand their current emotional state.

[1054] Emotion-based adjustment of information provision

[1055] The server adjusts the way it presents information based on the emotional state of the factory worker. For example, if the factory worker is feeling stressed, it will provide a more concise summary and classification results in an easy-to-understand format. If the factory worker is interested, it will provide detailed information or additional reference materials.

[1056] As a concrete example, consider a scenario in which a factory worker uploads a daily technical report called " / reports / machine_failure_report.txt" and the server generates the following summary: "In this daily report, machine X experienced motor overheating under certain operating conditions. As a countermeasure, the cooling system was adjusted." The factory worker also registers verification points such as "reproducibility of the problem," "appropriateness of the countermeasure," and "scope of impact," and the server provides the following classification results: "reproducibility of the problem: high," "appropriateness of the countermeasure: good," and "scope of impact: limited." The emotion engine can detect the stress level of factory workers and provide more concise information.

[1057] An example of a prompt is as follows:

[1058] "Please summarize the following technical daily report: / path / to / machine_failure_report.txt. In addition, please provide a classification and analysis based on the following verification points: 'Reproducibility of the problem', 'Appropriateness of the countermeasures', 'Scope of impact'. Please consider the feelings of the factory staff and simplify the information if stress is felt."

[1059] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1060] Step 1:

[1061] The server initializes the generative AI models, which include two models: a summary generation model and a text classification model. The server initializes these models using HuggingFace's transformers library. As input, it takes the URL or path of a pre-trained model. As output, it gets the initialized model object.

[1062] Step 2:

[1063] The user inputs the file path of the daily report or technical paper through the terminal interface and sends it to the server. The input is the file path (e.g., " / reports / machine_failure_report.txt"), and the output is that the sent file path reaches the server.

[1064] Step 3:

[1065] Based on the file path received by the server, the specified file is opened and read as text data. The input is the file path, and the output is text data containing the file contents.

[1066] Step 4:

[1067] The server passes the text data it has read to a summary generation model to generate a summary. Here, the summary generation model extracts the main points of the text data and summarizes them as short sentences. The input is the text data, and the output is the generated summary text.

[1068] Step 5:

[1069] The server provides the generated summary to the user by sending data to display the summary text on the user's terminal. The input is the summary text, and the output is the summary content displayed on the user interface.

[1070] Step 6:

[1071] The user inputs verification points for evaluating technical journals and technical papers into the terminal interface and sends them to the server. The inputs are verification points (e.g., "reproducibility of the problem," "appropriateness of the countermeasures," "scope of impact"), and the output is the verification points that reach the server.

[1072] Step 7:

[1073] The server uses the verification points and text data received from the user to perform analysis using a text classification model. The server classifies the corresponding text portion for each verification point and obtains the classification result. The input is the text data and verification points, and the output is the classification result for each verification point.

[1074] Step 8:

[1075] The server provides the user with the classification results for each verification point it has acquired. Here, it sends data to display the classification results on the user's device. The input is the classification results, and the output is the evaluation content displayed on the user interface.

[1076] Step 9:

[1077] The server uses an emotion engine to recognize the user's emotions based on the user's input, operation history, and facial expressions captured by the device's camera. The input is the user's operation history and camera footage, and the output is the recognized emotional state of the user.

[1078] Step 10:

[1079] The server adjusts the way it presents information based on the user's recognized emotional state. For example, if the user is feeling stressed, it will provide more concise information in an easy-to-understand format. The input is the user's emotional state, and the output is the adjusted information presentation data.

[1080] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 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.

[1081] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) 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.

[1082] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1083] [Fourth embodiment]

[1084] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1085] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1086] 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).

[1087] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. 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 control target 443 are also connected to the bus 52.

[1088] 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.

[1089] 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).

[1090] 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.

[1091] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1092] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, 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.

[1093] 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.

[1094] 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.

[1095] In the robot 414, 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.

[1096] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1097] This invention is a system that uses a generative AI model to accurately grasp the content of medical papers and support the process of determining their validity. Specific embodiments of this system are described below.

[1098] The server initializes the generated AI model

[1099] First, the server initializes the generative AI model. This generative AI model includes a summary generation model for performing text summarization and a text classification model for classifying text. By initializing these models, the server prepares to process medical papers.

[1100] User uploads medical paper

[1101] Next, the user uploads a medical paper. The user inputs the file path of the medical paper and sends it to the server. This file path is used when the server reads the text of the medical paper.

[1102] The server loads medical papers

[1103] The server reads the medical paper uploaded by the user. The server opens the paper file and reads its contents as text data. This text data is then passed to the summary generation model.

[1104] Server generates summary

[1105] The server passes the text data of medical papers read into a summary generation model, which then generates a summary. This summary generation model extracts important information from long sentences and outputs a short summary. The generated summary is then provided to the user.

[1106] User registers verification point

[1107] Users register validation criteria for evaluating medical papers, including the validity of the research, statistical accuracy, and reproducibility of results, among other aspects that the user wishes to evaluate.

[1108] The server analyzes the paper

[1109] Based on the verification points received from the user, the server analyzes each part of the medical paper using a text classification model. The server uses each verification point and the text data of the paper to obtain a classification result. This classification model is trained to classify the input text into a specific category.

[1110] The server provides the classification results

[1111] The server provides the user with the classification results for each verification point, allowing the user to check the evaluation results for each aspect of the medical paper.

[1112] Specific examples

[1113] A specific example is given below.

[1114] 1. User uploads paper file

[1115] Example: A user uploads a medical article called "path / to / medical_article.txt" to the server.

[1116] 2. The server generates the summary

[1117] Example: The server reads the content of a paper and uses a summary generation model to generate the following summary:

[1118] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, showed the expected effects."

[1119] 3. User registers verification points

[1120] Example: A user registers verification points such as "research validity," "statistical accuracy," and "reproducibility of results."

[1121] 4. The server analyzes the paper

[1122] Example: The server runs a text classification model on each validation point and obtains the following classification results:

[1123] Study validity: "Positive"

[1124] Statistical accuracy: "Neutral"

[1125] Reproducibility of results: "Mixed"

[1126] In this way, by using this system, medical and pharmaceutical experts can efficiently and accurately evaluate and verify a huge amount of medical research papers.

[1127] The processing flow will be explained below.

[1128] Step 1:

[1129] The server initializes the generative AI model, specifically, loads the summary generation model and text classification model, and deploys them to storage devices to efficiently perform natural language processing tasks.

[1130] Step 2:

[1131] The user inputs the file path of the medical paper and sends it to the server. The file path is input via the interface of the terminal used by the user.

[1132] Step 3:

[1133] The server receives the file path and opens the specified file. The server reads the contents of the paper file as text data. This text data includes the full text of the medical paper.

[1134] Step 4:

[1135] The server passes the read text data to a summary generation model. The model extracts the main points of the text and generates a summary. The generated summary is short and concise, describing the main points of the paper.

[1136] Step 5:

[1137] The server provides the generated abstract to the user, who can then view the abstract through a terminal interface to quickly understand the content of the paper.

[1138] Step 6:

[1139] Users input the validation points used to evaluate papers and submit them to the server. These validation points include the validity of the research, statistical accuracy, and reproducibility of results.

[1140] Step 7:

[1141] The server uses the verification points received from the user and the text data of the medical paper to analyze the paper using a text classification model. For each verification point, the corresponding part is classified and a classification result is obtained.

[1142] Step 8:

[1143] The server provides the classification results to the user, who can then view the evaluation results for each verification point via the terminal interface, thereby confirming the analysis results for each aspect of the medical paper.

[1144] Step 9:

[1145] Users evaluate the validity and reliability of medical papers based on the provided abstracts and classification results. The server records this evaluation process and stores it in a database for future reference and analysis.

[1146] Example 1

[1147] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1148] Efficiently evaluating and verifying large volumes of pharmaceutical research papers is a major challenge for medical and pharmaceutical experts. In particular, there is a need to quickly and accurately evaluate the validity, statistical accuracy, and reproducibility of results of research papers. However, conventional methods require a huge amount of time and effort to analyze each paper in detail, making them inefficient. Furthermore, results can vary depending on the evaluator's subjectivity, making it difficult to perform consistent evaluations.

[1149] 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.

[1150] In this invention, the server includes means for initializing the generative AI model, means for a user to upload a medical paper, means for reading the medical paper, means for passing the text data of the read medical paper to a summary generation model to generate a summary, means for providing the generated summary to the user, means for registering multiple verification points received from the user, means for classifying each part of the read medical paper using the multiple verification points received from the user with a text classification model to obtain the classification results, and means for providing the obtained classification results to the user, thereby enabling efficient summary generation of medical papers and automatic and consistent evaluation from multiple perspectives.

[1151] A "generative AI model" is a trained model that uses artificial intelligence to summarize and classify text.

[1152] The "server" is the computer system that serves as the core of the system, initializing the generative AI model, processing medical papers, and interfacing with users.

[1153] A "user" is a person or organization that uses the system to upload medical papers and receive their summaries and evaluation results.

[1154] A "medical paper" is a piece of writing that summarizes research results and findings related to medicine and pharmacy.

[1155] A "summary generation model" is a type of generative AI model used to extract important information from medical papers and generate short summaries.

[1156] A "text classification model" is a type of generative AI model used to classify parts of medical papers into specific categories.

[1157] "Verification points" are items that serve as viewpoints or criteria when a user evaluates a medical paper.

[1158] "Classification results" are evaluation results for each part of a medical paper generated by a text classification model.

[1159] "Upload" is the act of a user sending data from their own terminal to a server.

[1160] A "web interface" is a screen that uses a web browser to allow a user to access and operate the system.

[1161] This invention is a system that uses a generative AI model to efficiently summarize the contents of medical papers and assists in the process of determining their validity based on each verification point. Specific embodiments are described below.

[1162] 1. Hardware and Software Configuration

[1163] The server plays a central role in this system. The server hardware includes a powerful CPU, sufficient memory, and large storage capacity. The main software used includes Python, the Transformers library, and a suitable database management system.

[1164] 2. Initializing the Generative AI Model

[1165] When the system is initialized, the server loads generative AI models from disk. Specifically, it loads a "summary generation model" that generates summaries of medical papers and a "text classification model" that classifies text. These models have been trained in advance using large amounts of text data, and the server initializes their parameters and loads them into memory.

[1166] 3. User upload of medical papers

[1167] Users upload medical papers through the system's web interface. They use a file selection dialog to select the medical paper file they want to evaluate and click the upload button. The terminal sends the selected file path to the server, which stores it in a temporary directory.

[1168] 4. Reading medical papers

[1169] The server uses the file path sent by the user to open the uploaded medical paper, reads the file contents as text data, and stores it in memory. This text data is used in the subsequent summary generation process.

[1170] 5. Summary Generation

[1171] The server passes the text data of the loaded medical paper to a summary generation model to generate a summary. This summary generation model extracts important information from the medical paper and outputs a short summary. The generated summary is provided to users so that they can easily understand the outline of the paper.

[1172] 6. Registering Verification Points

[1173] Users register multiple validation points to evaluate the validity of a paper. These include aspects such as "research validity," "statistical accuracy," and "reproducibility of results." These validation points are entered on a web interface and sent to the server, which stores them in a database.

[1174] 7. Analysis of the paper

[1175] The server applies a text classification model using the verification points received from the user and the text data loaded, generating a classification result for each verification point and storing it in an internal data structure.

[1176] 8. Providing classification results

[1177] The server takes the generated classification results, converts them into a user-friendly format, and provides the results to the user through a web interface.

[1178] Specific examples

[1179] Initializing a generative AI model

[1180] Example: The server starts and loads a "summary generation model" and a "text classification model" using common natural language processing models.

[1181] User-uploaded medical papers

[1182] Example: User uploads "path / to / medical_article.txt" through the web interface.

[1183] Reading medical papers

[1184] Example: The server opens "medical_article.txt" from the specified path and reads its contents as text data.

[1185] Summary Generation

[1186] Example: The server generates the following summary using the summary generation model:

[1187] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, generally showed the expected effects."

[1188] Registering verification points

[1189] Example: A user enters verification points such as "study validity," "statistical accuracy," and "result reproducibility" into a web form.

[1190] Analysis of the paper

[1191] Example: The server applies a text classification model to each validation point and produces the following classification results:

[1192] Study validity: "Positive"

[1193] Statistical accuracy: "Neutral"

[1194] Reproducibility of results: "Mixed"

[1195] Providing classification results

[1196] Example: Displaying the analysis results in a table format in a web interface.

[1197] This system allows medical and pharmaceutical experts to efficiently evaluate and verify vast amounts of medical research.

[1198] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1199] Step 1: Initializing the generative AI model

[1200] Specific behavior:

[1201] When the system starts up, the server loads the generative AI model from disk, loads a "summary generation model" and a "text classification model" for analyzing medical papers from around the world, and initializes them.

[1202] Input and Output: The input is the generative AI model file on disk, and the output is the initialized generative AI model (in memory).

[1203] Step 2: User uploads medical papers

[1204] Specific behavior:

[1205] The user opens the system's web interface, selects a medical paper file using the file selection dialog, and clicks the upload button. The terminal sends the selected file path to the server, which stores it in a temporary directory.

[1206] Input and Output: The input is the file path of the medical paper selected by the user, and the output is the paper file stored in the server's storage.

[1207] Step 3: Reading medical papers

[1208] Specific behavior:

[1209] The server uses the file path of the uploaded medical paper to open the paper file, read the contents of the file as text data, and save it in memory. It then records the first few lines of the read text data in a log to confirm that it was read successfully.

[1210] Input and output: The input is a paper file stored in the server's storage, and the output is text data stored in memory.

[1211] Step 4: Summary generation

[1212] Specific behavior:

[1213] The server passes the text data of medical papers stored in memory to a summary generation model, which processes the text data to generate a short summary, which is then sent to a web interface for presentation to the user.

[1214] Input and Output: The input is the text data stored in memory, and the output is the generated summary sentence.

[1215] Step 5: Register your verification points

[1216] Specific behavior:

[1217] The user enters the verification points in a form on the web interface and sends them to the server, which stores them in a database.

[1218] Input and Output: The input is the verification point entered by the user and the output is the verification point stored in the database.

[1219] Step 6: Analyzing the paper

[1220] Specific behavior:

[1221] The server applies a text classification model to the medical text data stored in memory and the validation points received from the user, generating a classification result for each validation point and storing it in an internal data structure.

[1222] Input and Output: The input is the text data stored in memory and the validation points stored in the database, and the output is the classification results stored in the internal data structure.

[1223] Step 7: Providing classification results

[1224] Specific behavior:

[1225] The server retrieves the classification results stored in an internal data structure, converts them into a user-friendly format, and sends the converted analysis results to a web interface for display to the user.

[1226] Input and Output: The input is the classification results stored in an internal data structure, and the output is the analysis results displayed in a web interface.

[1227] (Application example 1)

[1228] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1229] Traditional logistics centers are required to manage a huge number of shipping lists and inventory reports, but they lack the means to process them efficiently.In addition, verifying and accurately evaluating the contents of the reports requires a lot of time and effort, and there is a risk of human error.

[1230] 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.

[1231] In this invention, the server includes means for initializing a generative AI model, means for reading a document, means for passing the text data of the read document to a summary generation model and generating a summary, means for providing the generated summary to a user, means for classifying each part of the read document using a text classification model using multiple verification points received from the user and obtaining the classification results, and means for providing the obtained classification results to the user. This enables efficient and accurate processing and evaluation of documents such as shipping lists and inventory reports, even in logistics centers.

[1232] A "generative AI model" is an algorithmic model designed to use artificial intelligence to perform tasks such as text generation, summarization, and classification.

[1233] A "document" is a collection of information, including text data, recorded on paper or in digital form.

[1234] "Means of reading" refers to the process of inputting a specified document file into the system and capturing its contents as digital data.

[1235] A "summary generation model" is an algorithmic model that extracts important information from long pieces of text and summarizes it concisely.

[1236] The "means for generating a summary" refers to a process for generating a summary from the input text data using a summary generation model.

[1237] "Means for providing" refers to the process for displaying or communicating the generated information or classification results to the user.

[1238] "Verification points" refer to specific perspectives or criteria that users evaluate.

[1239] A "text classification model" is an algorithmic model for classifying input text data into specific categories.

[1240] The "means for obtaining classification results" refers to the process of using a text classification model to determine the category of text data for each particular validation point and obtain a result.

[1241] A "smart device" refers to an electronic device that has internet connectivity, computing power, and can run applications.

[1242] This invention relates to a system that uses generative AI models to efficiently manage shipping lists and inventory reports within a logistics center and accurately evaluate their contents. This system allows managers and staff to acquire, analyze, and evaluate information using smart devices, thereby improving work efficiency and accuracy.

[1243] The server first initializes the generative AI models, which include a summary generation model and a text classification model. The summary generation model is used to extract important information from long texts and provide them as short summaries, while the text classification model is used to classify texts into specific categories.

[1244] Users upload logistics documents such as shipping lists and inventory reports, which are then loaded onto the server, where they are converted into text format and prepared for analysis.

[1245] The server then passes the read text data to a summary generation model to generate a summary. This summary is automatically generated in a short sentence format, allowing users to quickly grasp the content. The generated summary is then provided to the user via their smart device.

[1246] Users can set multiple evaluation points for shipping and inventory management. These verification points include evaluation criteria such as "shipping accuracy," "inventory accuracy," and "report validity."

[1247] The server analyzes the loaded text data using a text classification model based on the verification points received from the user. As a result of the analysis, a classification result is obtained for each verification point. This classification result is provided to the user via their smart device, and the user can check the evaluation results for each verification point.

[1248] The hardware used includes servers, smartphones, smart glasses, or logistics robots, while the software used is primarily Python programs and the Hugging Face transformers library.

[1249] For example, let's take a concrete example of how the system is used within a logistics center. When a user uploads a shipping report called "path / to / shipping_report.txt" to the system, the server reads the contents of the report and generates a summary. The summary is provided in the form of, "This shipping report describes the shipping status within a specific period, and shows that most tasks are being carried out according to schedule." Next, when the user registers verification points such as "Shipping accuracy," "Inventory accuracy," and "Report validity," the server classifies each verification point and produces classification results such as "Shipping accuracy: Positive," "Inventory accuracy: Neutral," and "Report validity: Positive."

[1250] Example prompt sentence:

[1251] "Design a system to assist distribution centers in the process of accurately capturing and validating the contents of shipping lists and inventory reports. The system will utilize a generative AI model to generate summaries and provide classification results based on validation points. Specifically, we will use a Python program and the Hugging Face transformers library."

[1252] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1253] Step 1:

[1254] The server initializes the generative AI model. Specifically, it initializes a summary generation model for summarizing text data and a text classification model for classifying text. As input, the configuration file and initialization data for the generative AI model are passed. As output, the launched summary generation model and text classification model are obtained.

[1255] Step 2:

[1256] The user uploads a shipping list or inventory report through a terminal. Specifically, the user enters the file path specified by the user and sends it to the server. The input includes the file path information selected by the user. The output is the file path information sent to the server.

[1257] Step 3:

[1258] The server reads the uploaded document file. Specifically, it opens the specified file path and reads its contents as text data. The input includes the file path information sent by the user. The output is the read text data.

[1259] Step 4:

[1260] The server passes text data to a summary generation model to generate a summary. Specifically, the text data is input to the summary generation model, and the model generates a summary. The input includes the read text data. The output is the generated summary text.

[1261] Step 5:

[1262] The server provides the generated summary to the user. Specifically, it sends the summary text to the terminal so that the user can view it on their smart device. The input includes the generated summary text. The output is display data that the user can view on their terminal.

[1263] Step 6:

[1264] The user sets the verification points for evaluation. Specifically, the user sends the aspects they want to evaluate (e.g., "shipping accuracy," "inventory accuracy," "report validity") to the server via their terminal. The input includes the verification point information selected by the user. The output is the verification point information sent to the server.

[1265] Step 7:

[1266] The server analyzes the text data loaded by the server using a text classification model based on the verification points. Specifically, the text data is classified for each verification point using the verification point information. The input includes the loaded text data and verification point information. The output is a classification result for each verification point.

[1267] Step 8:

[1268] The classification results acquired by the server are provided to the user. Specifically, the classification results are sent to the terminal, allowing the user to check the evaluation results for each perspective on their smart device. The input includes the classification results. The output is display data that the user can check on their terminal.

[1269] 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.

[1270] This invention provides more flexible and personalized information by combining a system that uses a generative AI model to accurately grasp the content of medical papers and assist in the process of determining their validity with an emotion engine that recognizes the user's emotions. Specific embodiments of this system are described below.

[1271] The server initializes the generated AI model

[1272] First, the server initializes the generative AI model. This generative AI model includes a summary generation model for performing text summarization and a text classification model for classifying text. By initializing these models, the server prepares for the automatic processing of medical papers. At the same time, an emotion engine for recognizing user emotions is also initialized.

[1273] User uploads medical paper

[1274] Next, the user uploads the medical paper. The user inputs the file path of the medical paper and sends it to the server. The file path is input via the interface of the terminal used by the user.

[1275] The server loads medical papers

[1276] The server receives the file path and opens the specified file. The server reads the contents of the paper file as text data. This text data includes the full text of the medical paper.

[1277] Server generates summary

[1278] The server passes the text data it has read to a summary generation model, which then generates a summary. The summary generation model extracts the main points of the text and generates a summary. The generated summary is short and concise, describing the main points of the paper.

[1279] The server provides the summary

[1280] The server provides the generated abstract to the user, who can then view the abstract through a terminal interface to quickly understand the content of the paper.

[1281] User registers verification point

[1282] Users input verification criteria for evaluating medical papers and submit them to the server. These verification criteria include the validity of the research, statistical accuracy, and reproducibility of the results.

[1283] The server analyzes the paper

[1284] The server uses the verification points received from the user and the text data of the medical paper to analyze the paper using a text classification model, and classifies the corresponding part for each verification point to obtain the classification result.

[1285] The server provides the classification results

[1286] The server provides the user with the classification results for each verification point. The user can view the evaluation results for each verification point through the terminal interface, allowing them to check the analysis results for each aspect of the medical paper.

[1287] Recognize user emotions

[1288] The emotion engine recognizes the user's emotions based on their input and operation history. For example, the emotion engine can detect when the user is feeling stressed or interested. This allows the system to understand the user's current emotional state.

[1289] Emotion-based adjustment of information provision

[1290] The server can then adapt the way it presents information based on the user's perceived emotional state. For example, if the user is feeling stressed, it can provide a more concise summary or classification in an easy-to-understand format. It can also provide more detailed information or additional references if the user is interested.

[1291] Specific examples

[1292] 1. User uploads paper file

[1293] Example: A user uploads a medical article called "path / to / medical_article.txt" to the server.

[1294] 2. The server generates the summary

[1295] Example: The server reads the content of a paper and uses a summary generation model to generate the following summary:

[1296] "The study evaluated the efficacy and safety of Drug A under specific conditions. The results, although condition dependent, showed the expected effects."

[1297] 3. User registers verification points

[1298] Example: A user registers verification points such as "research validity," "statistical accuracy," and "reproducibility of results."

[1299] 4. The server analyzes the paper

[1300] Example: The server runs a text classification model on each validation point and obtains the following classification results:

[1301] Study validity: "Positive"

[1302] Statistical accuracy: "Neutral"

[1303] Reproducibility of results: "Mixed"

[1304] 5. Emotion recognition and information provision coordination

[1305] Example: If the emotion engine detects a stressed state from the user's operation history, the server will present the information more concisely, making it easier to understand.

[1306] This system allows users to efficiently evaluate the validity and reliability of medical papers and receive personalized information based on the user's emotional state.

[1307] The processing flow will be explained below.

[1308] Step 1:

[1309] The server initializes the generative AI model and emotion engine. Specifically, it loads the summary generation model, text classification model, and emotion engine that recognizes the user's emotions into memory and makes them ready for use.

[1310] Step 2:

[1311] The user inputs the file path of the medical paper using the terminal and sends it to the server. The file path indicates the location information of the medical paper specified by the user.

[1312] Step 3:

[1313] The server opens the specified medical paper file based on the file path received from the user. The server reads the contents of the medical paper as text data and stores it in memory.

[1314] Step 4:

[1315] The server inputs the text data into a summary generation model to generate a summary, which is a short sentence that succinctly summarizes the main points of the paper.

[1316] Step 5:

[1317] The server sends the generated abstract to the user's terminal, and the user can view the abstract through the terminal interface. The user can refer to the abstract and quickly understand the content of the paper.

[1318] Step 6:

[1319] Users input the validation points used to evaluate medical papers through their terminals and send them to the server. These validation points include aspects that users want to evaluate, such as the validity of the research, statistical accuracy, and reproducibility of results.

[1320] Step 7:

[1321] The server uses the verification points received from the user and the text data of the medical paper it has read to analyze the paper using a text classification model. For each verification point, the model classifies the corresponding part and obtains the classification result.

[1322] Step 8:

[1323] The server sends the classification results for each verification point to the user's device, and the user can view the classification results through the device's interface, allowing the user to check the evaluation results of the paper for each perspective.

[1324] Step 9:

[1325] The emotion engine recognizes the user's emotional state from their input, operation history, reactions, etc. For example, if a user browses a lot of literature in a short period of time, the emotion engine can determine that the user is in a stressful state.

[1326] Step 10:

[1327] The server adjusts the way it presents information based on the user's emotional state. For example, if the user is feeling stressed, it will adjust the way it presents information by making summaries and classification results more concise and easy to understand.

[1328] Step 11:

[1329] The server then sends the adjusted information to the user's device, where the user can view the adjusted summary and classification results through the device interface. The adjusted information is optimized to fit the user's emotional state.

[1330] Step 12:

[1331] Users can efficiently evaluate the validity and reliability of medical papers based on the provided abstracts and classification results, and the server records this evaluation process and stores it in a database for future reference and analysis.

[1332] In this way, a system that combines an emotion engine can accurately grasp the content of medical papers and provide information that corresponds to the user's emotional state, thereby providing personalized support that meets the user's needs.

[1333] Example 2

[1334] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1335] Modern medical papers contain a vast amount of information, making it extremely difficult to quickly and accurately grasp and evaluate its validity. Furthermore, because information is not provided in response to the user's emotions or state, there are concerns that this will increase the burden on researchers. There is a need to solve these problems and provide an environment in which users can more efficiently and comfortably evaluate and understand medical papers.

[1336] 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.

[1337] In this invention, the server includes means for initializing the generation AI model, means for reading electronic documents, and means for passing the text data of the read electronic documents to the summary generation model and generating summaries, thereby enabling users to quickly grasp the vast amount of information in electronic documents and providing personalized information according to the user's emotional state.

[1338] A "generative AI model" refers to an artificial intelligence model for summarizing and classifying text data.

[1339] A "summary generation model" refers to an artificial intelligence model that extracts the main points of input text data and generates a concise summary.

[1340] "Text classification model" refers to an artificial intelligence model for classifying input text data based on specific criteria.

[1341] "Electronic literature" refers to text data such as papers and articles stored in digital format.

[1342] "User" refers to a person who uses this system and operates it for the purpose of understanding and evaluating the content of electronic documents.

[1343] An "emotion engine" refers to software or a system that recognizes emotions and states based on user input and operation history.

[1344] "Verification points" refer to specific viewpoints or criteria that users set for evaluating electronic documents.

[1345] "Means for adjusting the way information is presented" refers to processes or software functions for changing the format or content of the information provided depending on the perceived user's emotions or state.

[1346] This invention provides more flexible and personalized information by combining a system that uses a generative AI model to accurately grasp the content of electronic documents and assist in the process of evaluating their validity with an emotion engine that recognizes the user's emotions. Specific embodiments of this system are described below.

[1347] First, the server initializes the generative AI model. This generative AI model includes a "summary generation model" for summarizing text and a "text classification model" for classifying text. By initializing these models, the server prepares for the automatic processing of electronic documents. At the same time, the server also initializes an "emotion engine" for recognizing user emotions. This initialization process includes loading model parameters and configuration files.

[1348] Next, the user uploads the electronic document. The user uses the terminal interface to input the file path of the electronic document and sends it to the server. Specifically, the user can specify the file path as "path / to / medical_article.txt". This input information is received by the server.

[1349] The server uses the received file path to open the specified file. Using the standard Python library, the file is opened and the contents are read as text data. This allows the full text of the electronic document to be obtained as data.

[1350] The acquired text data is passed to a summary generation model, which extracts the key points of the text and generates a summary. For example, the model can generate a summary like this: "A study evaluated the efficacy and safety of drug A under specific conditions. The results, although condition dependent, showed the expected effects." This summary is then sent from the server to the user's device.

[1351] After viewing an abstract, users input specific verification criteria for evaluating the electronic literature, such as "study validity," "statistical accuracy," and "reproducibility of results." These verification criteria are then sent to the server.

[1352] The server uses the received validation points and the text data of the electronic document to analyze the document using a text classification model. It classifies the corresponding part for each validation point and obtains a classification result. For example, it can generate classification results such as "Research validity: Positive," "Statistical accuracy: Neutral," and "Reproducibility of results: Mixed." The classification results are then sent to the user's device, where the user can view the evaluation results for each validation point.

[1353] Furthermore, the emotion engine recognizes emotions based on the user's input and operation history. For example, the emotion engine detects whether the user is feeling stressed or interested. Based on the recognized emotional state, the server adjusts the way information is provided. For example, if the user is feeling stressed, the server can provide summaries and classification results in a more concise and easy-to-understand format. If the user is interested, the server can provide detailed information and additional reference materials.

[1354] As an example of a prompt sentence, a user might upload a medical paper called "path / to / medical_article.txt" to a server, and a summary generation model might generate the following summary: "The study evaluated the efficacy and safety of drug A under specific conditions. The results, although dependent on the conditions, showed the expected effects."

[1355] This allows users to efficiently evaluate the validity and reliability of medical papers and receive personalized information according to their emotional state.

[1356] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1357] Step 1:

[1358] The server initializes the generative AI model. Specifically, the server loads and initializes the summary generation model and text classification model. It also initializes the emotion engine to recognize the user's emotions. This includes loading the model parameters and configuration files. The model file and configuration file are required as input, and an initialized model object is generated as output.

[1359] Step 2:

[1360] The user uploads an electronic document. The user specifies the file path on the terminal interface and sends it to the server. Specifically, the user enters the file path "path / to / medical_article.txt". The file path is provided as input, and the file path information is sent to the server.

[1361] Step 3:

[1362] The server reads the electronic document. Based on the received file path, the server opens the specified file and reads the contents as text data. Depending on the programming language used, for example, Python's "open()" function is used. The file path is received as input, and the full text data of the electronic document is generated as output.

[1363] Step 4:

[1364] The server generates summaries of electronic literature. The text data is passed to a summary generation model, which extracts the main points of the text and generates a summary. Specifically, a generative AI model such as BERT is used to summarize the text data. The text data of the electronic literature is required as input, and summary text is generated as output. For example, a summary might be generated that reads, "A study evaluated the efficacy and safety of drug A under specific conditions. The results, while dependent on the conditions, showed the expected effects."

[1365] Step 5:

[1366] The server provides the generated summary to the user. The summary text is sent from the server to the user's terminal, and the user views the summary on the terminal interface. The generated summary text is required as input, and the summary text is displayed on the terminal screen as output.

[1367] Step 6:

[1368] The user registers validation points. The user enters multiple validation points into the terminal interface and sends them to the server. Specifically, the user enters items such as "research validity," "statistical accuracy," and "result reproducibility." A list of validation points is provided as input, and the information is sent to the server.

[1369] Step 7:

[1370] The server analyzes the electronic document. It uses a text classification model to classify the specified part of the electronic document and obtains classification results for each verification point. The input requires the text data of the electronic document and a list of verification points, and the output generates classification results for each verification point. For example, classification results such as "Research validity: Positive," "Statistical accuracy: Neutral," and "Reproducibility of results: Mixed" can be obtained.

[1371] Step 8:

[1372] The server provides the classification results to the user. The acquired classification results are sent from the server to the user's device, and the user can view the evaluation results of each verification point on the device interface. The classification results are required as input, and are displayed on the screen as output.

[1373] Step 9:

[1374] The emotion engine recognizes the user's emotions. It analyzes emotions from the user's input and operation history, and detects emotional states such as stress and interest. Input data includes operation history such as input speed and click frequency. The user's emotional state is recognized as the output.

[1375] Step 10:

[1376] The server adjusts the way information is presented. The server adjusts the format of information presentation based on the recognized emotional state. For example, if the user is feeling stressed, the server will provide information in a concise and easy-to-understand format. The emotional state is required as input, and tailored information is provided as output.

[1377] (Application example 2)

[1378] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1379] Problems often occur in the manufacturing process of modern factories, and dealing with them can take a lot of time and effort. It is also difficult to quickly and accurately grasp the large amount of information available in technical journals and specialized papers, and to implement appropriate countermeasures. Furthermore, dealing with problems often causes stress for on-site workers. In these situations, there is a growing need for a system that can efficiently summarize information and provide it while taking emotions into consideration.

[1380] 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.

[1381] In this invention, the server includes means for initializing a generative AI model, means for reading text data, means for passing the read text data to a summary generation model and generating a summary, means for providing the generated summary to a user, means for classifying each portion of the read text data using a text classification model and obtaining a classification result using multiple verification points received from the user, means for providing the obtained classification result to the user, means for initializing an emotion engine that recognizes the user's emotions, and means for adjusting a method of providing information based on the user's emotional state recognized by the emotion engine. This speeds up troubleshooting in factories and enables support that takes user emotions into consideration.

[1382] A "generative AI model" is an artificial intelligence model trained to perform tasks such as summarizing, classifying, and generating text data.

[1383] An "emotion engine" is software or hardware that recognizes a user's emotional state and adjusts the way information is presented.

[1384] "Text data" refers to data that consists of textual information, and includes technical journals and specialized papers.

[1385] A "summary generation model" is a machine learning model that extracts key points from long text data and summarizes them as short sentences.

[1386] A "text classification model" is an artificial intelligence model trained to classify text data into specific classes or categories.

[1387] "Verification points" are points of view that the user inputs as evaluation items for text data, and include, for example, "reproducibility of the problem" and "appropriateness of the countermeasures."

[1388] "Adjusting the method of providing information" refers to optimizing the content and format of the information provided based on the user's emotional state.

[1389] This invention relates to a system that utilizes a generative AI model to efficiently summarize technical journals and specialized papers related to manufacturing processes in factories, and further recognizes user emotions to optimize information provision. Specific embodiments of this system are described below.

[1390] The server initializes the generated AI model

[1391] The server initializes the generative AI model. This generative AI model includes a summary generation model for summarizing text data and a text classification model for classifying text data. These models are initialized using the HuggingFace transformers library. An emotion engine that recognizes the user's emotions is also initialized at the same time, and OpenCV is used for this emotion engine.

[1392] Factory workers upload daily reports or technical papers to the robot

[1393] A factory worker inputs the file path of the daily report or technical paper through the terminal interface and sends it to the server.

[1394] The server reads the text data.

[1395] The server opens the specified file and reads it as text data, which may include the full text of technical journals or technical papers.

[1396] Server generates summary

[1397] The server passes the text data it has read to a summary generation model, which then generates a summary by extracting the main points of the text.

[1398] The server provides the summary

[1399] The server provides the generated summary to the factory staff, who can view the summary through the terminal interface and quickly understand the contents.

[1400] Factory staff register verification points

[1401] Factory staff input verification points for evaluating technical daily reports and specialist papers and send them to the server. These verification points include "reproducibility of the problem," "appropriateness of countermeasures," and "scope of impact."

[1402] The server analyzes the text data

[1403] The server uses the verification points and text data received from the factory staff to perform analysis using a text classification model. The server classifies the corresponding part for each verification point and obtains the classification result.

[1404] The server provides the classification results

[1405] The server provides the classification results for each verification point to factory workers, who can view the evaluation results for each verification point through the terminal interface.

[1406] Recognize user emotions

[1407] The emotion engine recognizes the emotions of factory workers based on their input, operation history, and facial expressions, allowing the system to understand their current emotional state.

[1408] Emotion-based adjustment of information provision

[1409] The server adjusts the way it presents information based on the emotional state of the factory worker. For example, if the factory worker is feeling stressed, it will provide a more concise summary and classification results in an easy-to-understand format. If the factory worker is interested, it will provide detailed information or additional reference materials.

[1410] As a concrete example, consider a scenario in which a factory worker uploads a daily technical report called " / reports / machine_failure_report.txt" and the server generates the following summary: "In this daily report, machine X experienced motor overheating under certain operating conditions. As a countermeasure, the cooling system was adjusted." The factory worker also registers verification points such as "reproducibility of the problem," "appropriateness of the countermeasure," and "scope of impact," and the server provides the following classification results: "reproducibility of the problem: high," "appropriateness of the countermeasure: good," and "scope of impact: limited." The emotion engine can detect the stress level of factory workers and provide more concise information.

[1411] An example of a prompt is as follows:

[1412] "Please summarize the following technical daily report: / path / to / machine_failure_report.txt. In addition, please provide a classification and analysis based on the following verification points: 'Reproducibility of the problem', 'Appropriateness of the countermeasures', 'Scope of impact'. Please consider the feelings of the factory staff and simplify the information if stress is felt."

[1413] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1414] Step 1:

[1415] The server initializes the generative AI models, which include two models: a summary generation model and a text classification model. The server initializes these models using HuggingFace's transformers library. As input, it takes the URL or path of a pre-trained model. As output, it gets the initialized model object.

[1416] Step 2:

[1417] The user inputs the file path of the daily report or technical paper through the terminal interface and sends it to the server. The input is the file path (e.g., " / reports / machine_failure_report.txt"), and the output is that the sent file path reaches the server.

[1418] Step 3:

[1419] Based on the file path received by the server, the specified file is opened and read as text data. The input is the file path, and the output is text data containing the file contents.

[1420] Step 4:

[1421] The server passes the text data it has read to a summary generation model to generate a summary. Here, the summary generation model extracts the main points of the text data and summarizes them as short sentences. The input is the text data, and the output is the generated summary text.

[1422] Step 5:

[1423] The server provides the generated summary to the user by sending data to display the summary text on the user's terminal. The input is the summary text, and the output is the summary content displayed on the user interface.

[1424] Step 6:

[1425] The user inputs verification points for evaluating technical journals and technical papers into the terminal interface and sends them to the server. The inputs are verification points (e.g., "reproducibility of the problem," "appropriateness of the countermeasures," "scope of impact"), and the output is the verification points that reach the server.

[1426] Step 7:

[1427] The server uses the verification points and text data received from the user to perform analysis using a text classification model. The server classifies the corresponding text portion for each verification point and obtains the classification result. The input is the text data and verification points, and the output is the classification result for each verification point.

[1428] Step 8:

[1429] The server provides the user with the classification results for each verification point it has acquired. Here, it sends data to display the classification results on the user's device. The input is the classification results, and the output is the evaluation content displayed on the user interface.

[1430] Step 9:

[1431] The server uses an emotion engine to recognize the user's emotions based on the user's input, operation history, and facial expressions captured by the device's camera. The input is the user's operation history and camera footage, and the output is the recognized emotional state of the user.

[1432] Step 10:

[1433] The server adjusts the way it presents information based on the user's recognized emotional state. For example, if the user is feeling stressed, it will provide more concise information in an easy-to-understand format. The input is the user's emotional state, and the output is the adjusted information presentation data.

[1434] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice 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 voice data.

[1435] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) 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.

[1436] 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 robot 414.

[1437] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1438] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1439] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1440] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1441] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1442] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1443] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1444] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1445] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1446] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1447] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1448] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1449] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1450] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1451] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1452] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1453] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1454] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1455] The following is further disclosed regarding the above embodiment.

[1456] (Claim 1)

[1457] a means for initializing the generative AI model;

[1458] A means of reading medical papers,

[1459] A means for passing the read text data of the medical paper to a summary generation model and generating a summary;

[1460] means for providing the generated summary to a user;

[1461] A means for classifying each part of the read medical paper by a text classification model using a plurality of verification points received from a user, and obtaining a classification result;

[1462] means for providing the acquired classification results to a user;

[1463] A system including:

[1464] (Claim 2)

[1465] 2. The system of claim 1, further comprising means for setting a maximum length and a minimum length for the generated summary.

[1466] (Claim 3)

[1467] 10. The system of claim 1, wherein the text classification model includes means for obtaining a classification result for each verification point registered by a user.

[1468] "Example 1"

[1469] (Claim 1)

[1470] a means for initializing the generative AI model;

[1471] a means for users to upload medical papers;

[1472] A means of reading medical papers,

[1473] A means for passing the read text data of the medical paper to a summary generation model and generating a summary;

[1474] means for providing the generated summary to a user;

[1475] means for registering a plurality of verification points received from a user;

[1476] A means for classifying each part of the read medical paper by a text classification model using a plurality of verification points received from a user, and obtaining a classification result;

[1477] means for providing the acquired classification results to a user;

[1478] A system including:

[1479] (Claim 2)

[1480] 2. The system of claim 1, further comprising means for setting a maximum length and a minimum length for the generated summary.

[1481] (Claim 3)

[1482] 10. The system of claim 1, wherein the text classification model includes means for obtaining a classification result for each verification point registered by a user.

[1483] "Application Example 1"

[1484] (Claim 1)

[1485] a means for initializing the generative AI model;

[1486] a means for reading a document;

[1487] a means for passing the text data of the read document to a summary generation model and generating a summary;

[1488] means for providing the generated summary to a user;

[1489] A means for classifying each part of the read document using a text classification model by using a plurality of verification points received from a user, and obtaining a classification result;

[1490] means for providing the acquired classification results to a user;

[1491] A system including:

[1492] (Claim 2)

[1493] 2. The system of claim 1, further comprising means for setting a maximum length and a minimum length for the generated summary.

[1494] (Claim 3)

[1495] 10. The system of claim 1, wherein the text classification model includes means for obtaining a classification result for each verification point registered by a user.

[1496] (Claim 4)

[1497] The system of claim 1 , wherein the document is a document including a logistics document.

[1498] (Claim 5)

[1499] The system of claim 1, further comprising means for displaying the summary generation and classification results on a smart device.

[1500] "Example 2: Combining Emotion Engines"

[1501] (Claim 1)

[1502] a means for initializing the generative AI model;

[1503] a means for reading electronic documents;

[1504] a means for passing the text data of the electronic document read to a summary generation model and generating a summary;

[1505] means for providing the generated summary to a user;

[1506] A means for classifying each part of the read electronic document using a text classification model by using a plurality of verification points received from a user, and obtaining a classification result;

[1507] A means for recognizing emotions from user input and operation history;

[1508] means for adjusting a method of providing information based on the recognized emotion;

[1509] means for providing the acquired classification result to a user;

[1510] A system including:

[1511] (Claim 2)

[1512] 2. The system of claim 1, further comprising means for setting a maximum length and a minimum length for the generated summary.

[1513] (Claim 3)

[1514] 10. The system of claim 1, wherein the text classification model includes means for obtaining a classification result for each verification point registered by a user.

[1515] "Application example 2 when combining emotion engines"

[1516] (Claim 1)

[1517] a means for initializing the generative AI model;

[1518] A means for reading text data;

[1519] A means for passing the read text data to a summary generation model and generating a summary;

[1520] means for providing the generated summary to a user;

[1521] a means for classifying each portion of the read text data using a text classification model by utilizing a plurality of verification points received from a user, and obtaining a classification result;

[1522] means for providing the acquired classification results to a user;

[1523] means for initializing an emotion engine that recognizes the user's emotions;

[1524] means for adjusting a method of providing information based on the emotional state of the user recognized by the emotion engine;

[1525] A system including:

[1526] (Claim 2)

[1527] 2. The system of claim 1, further comprising means for setting a maximum length and a minimum length for the generated summary.

[1528] (Claim 3)

[1529] 10. The system of claim 1, wherein the text classification model includes means for obtaining a classification result for each verification point registered by a user. [Explanation of symbols]

[1530] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. a means for initializing the generative AI model; A means of reading medical papers, A means for passing the read text data of the medical paper to a summary generation model and generating a summary; means for providing the generated summary to a user; A means for classifying each part of the read medical paper by a text classification model using a plurality of verification points received from a user, and obtaining a classification result; means for providing the acquired classification results to a user; A system including:

2. 2. The system of claim 1, further comprising means for setting a maximum length and a minimum length for the generated summary.

3. The system of claim 1 , wherein the text classification model includes means for obtaining a classification result for each verification point registered by a user.

Images