Apparatus for information generation based on patent analysis model and method thereof

Abstract

The present invention relates to an information generation device comprising: a memory storing computer-executable instructions; and a processor that accesses the memory and executes the instructions, wherein the processor extracts patent information from a target document and, based on determining a target prompt indicating an analysis purpose from a user, generates grouping information through a combination of the patent information and the target prompt, applies the grouping information to a patent analysis model through batch processing to obtain a target response, stores the target response received from the patent analysis model by matching it to correspond with the grouping information, and extracts at least one patent analysis result from the target response by parsing according to a predetermined output rule.

Description

Technology Field

[0001] The present invention relates to an information generation device and a method thereof, and more specifically, to a technology for outputting information for patent search and analysis using generative artificial intelligence. Background Technology

[0002] Patent search and analysis tasks were performed manually by professional analysts or users. This included noise filtering, technical classification, validity analysis, citation verification, and legal status review for a large number of collected patent documents, requiring significant time and effort to identify and classify meaningful data from the entire patent information.

[0003] In an attempt to automate this process, there are instances where artificial neural network-based learning AI technology has been applied. Learning AI operates by training on labeled patent data to infer the relevant technical field or classification, or to extract key technical information, from new input patent documents. However, unlike general text, patent documents possess high-level semantic structures requiring technical and legal interpretation, and include long texts and complex narrative formats. Consequently, learning AI is sensitive to labeling accuracy and data volume, and faces limitations in deriving accurate interpretation results due to the difficulty of flexibly adapting to various modes of expression.

[0004] Consequently, existing manual methods are inefficient in terms of time and cost, and learning-based AI approaches lack precision in interpreting the contextual meaning of patent documents, thus presenting practical problems in the automation of patent search and analysis. The problem to be solved

[0005] Embodiments of the present invention aim to provide an information generation device and a method for generating information to effectively interpret the technical content of a patent document based on generative artificial intelligence, and for providing results in various formats according to the purpose of analysis.

[0006] Furthermore, embodiments of the present invention aim to provide an information generation device and a method for reducing user work time and increasing the consistency and quality of analysis by extracting patent information from a target document and generating a target response by having a generative artificial intelligence analyze grouped information based on a prompt corresponding to an analysis purpose input by a user.

[0007] The technical problems of the present invention are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0008] An information generation device according to an embodiment of the present invention includes a memory storing computer-executable instructions; and a processor that accesses the memory and executes the instructions. The processor can extract patent information from a target document and, based on determining a target prompt indicating an analysis purpose from a user, generate grouping information through a combination of the patent information and the target prompt, apply the grouping information to a patent analysis model through batch processing to obtain a target response, match the target response received from the patent analysis model to correspond with the grouping information and store it, and extract at least one patent analysis result from the target response by parsing according to a predetermined output rule.

[0009] In one embodiment, when the processor receives a search query from the user, it applies the search query to a database to extract the target document, and when the search query is not received, it applies output information obtained by applying the user's identification information to a keyword recommendation model to the database to extract the target document.

[0010] In one embodiment, the processor can extract from the target document bibliographic information including an application number, filing date, classification code, and inventor, and patent information including a description of invention, claims, legal status, and litigation information.

[0011] In one embodiment, the processor identifies a system prompt regarding a predetermined form of an analysis method or output format for processing the patent information and a user prompt designated by the user, and can determine the target prompt based on the system prompt and the user prompt.

[0012] In one embodiment, the processor applies the grouping information to the patent analysis model, which includes a large language model (LLM) operating in an external server or local environment, to generate the target response, which includes analysis information regarding the grouping information, and can extract the patent analysis result by converting the target response into at least one form among text, matrix, image, or graph.

[0013] In one embodiment, the processor may output the output information in at least one form among a web page, a file, electronic mail, or a social network service, based on determining the output information by combining the patent information and the patent analysis result.

[0014] In one embodiment, the processor stores at least one of the patent analysis result, the target prompt, and the grouping information in the memory, and can identify and update updatable items including a registration date, a request for examination status, or a right holder from the patent analysis result stored in the memory.

[0015] In one embodiment, the processor receives an input including a question or instruction from the user and, based on the patent analysis result and the patent information stored in the memory, can provide an output corresponding to the input to the user through an interactive user interface.

[0016] An information generation method according to an embodiment of the present invention may include: an operation of extracting patent information from a target document; an operation of generating grouping information through a combination of the patent information and the target prompt based on determining a target prompt indicating an analysis purpose from a user; an operation of obtaining a target response by applying the grouping information to a patent analysis model through batch processing; an operation of storing the target response received from the patent analysis model by matching it to correspond with the grouping information; and an operation of extracting at least one patent analysis result by parsing the target response according to a predetermined output rule.

[0017] In one embodiment, the operation of extracting the patent information may include: an operation of extracting the target document by applying the search query to a database when the search query is received from the user; and an operation of extracting the target document by applying output information obtained by applying the user's identification information to a keyword recommendation model to the database when the search query is not received.

[0018] In one embodiment, the operation of extracting the patent information may include, from the target document, bibliographic information including an application number, filing date, classification code, and inventor, and the patent information including a description of the invention, claims, legal status, and litigation information.

[0019] In one embodiment, the operation of generating the grouping information may include: an operation of identifying a system prompt and a user prompt specified by the user regarding a predetermined form of an analysis method or output format for processing the patent information; and an operation of determining the target prompt based on the system prompt and the user prompt.

[0020] In one embodiment, the operation of extracting the patent analysis result may include: applying the grouping information to the patent analysis model including a large language model (LLM) operating in an external server or local environment to generate the target response including analysis information regarding the grouping information; and extracting the patent analysis result by converting the target response into at least one form among text, matrix, image, or graph.

[0021] In one embodiment, the operation of extracting the patent analysis result may include the operation of outputting the output information in at least one form among a web page, a file, electronic mail, or a social network service, based on determining the output information by combining the patent information and the patent analysis result.

[0022] In one embodiment, the operation of extracting the patent analysis result may include: the operation of storing at least one of the patent analysis result, the target prompt, and the grouping information in memory; and the operation of identifying and updating updatable items including a registration date, a request for examination status, or a right holder from the patent analysis result stored in memory.

[0023] In one embodiment, the operation of extracting the patent analysis result may include: receiving an input including a question or instruction from the user; and providing an output corresponding to the input to the user through an interactive user interface based on the patent analysis result and the patent information stored in the memory. Effects of the invention

[0024] The effects of the information generating device and method according to the present invention are described as follows.

[0025] According to at least one of the embodiments of the present invention, information for effectively interpreting the technical content of a patent document based on generative artificial intelligence can be generated, and results in various formats can be provided according to the purpose of analysis.

[0026] In addition, according to at least one embodiment of the present invention, patent information is extracted from a target document, and a generative artificial intelligence analyzes grouped information based on a prompt corresponding to an analysis purpose entered by a user to generate a target response, thereby having the effect of reducing the user's work time and increasing the consistency and quality of the analysis.

[0027] In addition, various effects that can be identified directly or indirectly through this document may be provided. Brief explanation of the drawing

[0028] FIG. 1 is a drawing illustrating an information generating device according to one embodiment of the present invention. FIG. 2 is a flowchart illustrating a method for extracting patent analysis results from a target document in an information generation device according to an embodiment of the present invention. FIG. 3 is a flowchart illustrating a method for extracting patent analysis results based on a target document in an information generation device according to an embodiment of the present invention. FIG. 4 is a drawing illustrating an example of patent information or patent analysis results extracted from a target document in an information generation device according to one embodiment of the present invention. In relation to the description of the drawings, the same or similar reference numerals may be used for identical or similar components. Specific details for implementing the invention

[0029] Some embodiments of the present invention will be described in detail below with reference to exemplary drawings. It should be noted that in assigning reference numerals to the components of each drawing, the same components are given the same reference numeral whenever possible, even if they are shown in different drawings. Furthermore, in describing embodiments of the present invention, if it is determined that a detailed description of related known components or functions would hinder understanding of the embodiments of the present invention, such detailed description is omitted. In particular, various embodiments of this document are described with reference to the accompanying drawings. However, this is not intended to limit the technology described in this document to specific embodiments, and it should be understood that it includes various modifications, equivalents, and / or alternatives to the embodiments of this document. In relation to the description of the drawings, similar reference numerals may be used for similar components.

[0030] In describing the components of the embodiments of the present invention, terms such as first, second, A, B, (a), (b), etc., may be used. These terms are used merely to distinguish the components from other components, and the essence, order, or sequence of the components is not limited by such terms. Furthermore, unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the present invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application. For example, expressions such as "first," "second," "first," or "second" used in this document may modify various components regardless of order and / or importance, and are used merely to distinguish one component from another, without limiting the components. For example, the first user device and the second user device may represent different user devices regardless of order or importance. For example, without departing from the scope of rights set forth in this document, the first component may be named the second component, and similarly, the second component may be renamed the first component.

[0031] In this document, expressions such as "have," "can have," "include," or "can include" refer to the existence of the relevant feature (e.g., numerical values, functions, actions, or components, etc.) and do not exclude the existence of additional features.

[0032] Where it is stated that a certain component (e.g., a first component) is "(operatively or communicatively) coupled with" or "connected to" another component (e.g., a second component), it should be understood that the said certain component may be directly connected to the said other component or connected through another component (e.g., a third component). On the other hand, where it is stated that a certain component (e.g., a first component) is "directly connected" or "directly connected" to another component (e.g., a second component), it may be understood that no other component (e.g., a third component) exists between the said certain component and the said other component.

[0033] The expression "configured to" as used in this document may be replaced with, depending on the context, for example, "suitable for," "having the capacity to," "designed to," "adapted to," "made to," or "capable of."

[0034] The term “configured (or set to)” may not necessarily mean only that which is “specifically designed to” in hardware. Instead, in some cases, the expression “device configured to” may mean that the device is “capable of” doing so in conjunction with other devices or components. For example, the phrase “processor configured (or set to) perform A, B, and C” may mean a dedicated processor for performing the said operations (e.g., an embedded processor), or a generic-purpose processor (e.g., a CPU or an application processor) capable of performing said operations by executing one or more software programs stored in a memory device. The terms used herein are used merely to describe specific embodiments and are not intended to limit the scope of other embodiments. A singular expression may include a plural expression unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as generally understood by a person of ordinary skill in the art described herein. Terms used in this document that are defined in general dictionaries may be interpreted as having the same or similar meaning as they do in the context of the relevant technology, and are not to be interpreted in an ideal or overly formal sense unless explicitly defined in this document. In some cases, even terms defined in this document shall not be interpreted to exclude the embodiments of this document.

[0035] In this document, expressions such as "A or B," "at least one of A or / and B," or "one or more of A or / and B" may include all possible combinations of the items listed together. For example, "A or B," "at least one of A and B," or "at least one of A or B" may refer to cases including (1) at least one A, (2) at least one B, or (3) both at least one A and at least one B. Additionally, in describing the components of an embodiment of the present invention, phrases such as "A or B," "at least one of A and B," "at least one of A or B," "A, B or C," "at least one of A, B and C," "at least one of A, B, or C," and "at least one of A, B, C, or any combination thereof" may each include any one of the items listed together with the corresponding phrase, or all possible combinations thereof. In particular, phrases such as "at least one of A, B, C, or any combination thereof" may include A, B, or C, or combinations thereof such as AB or ABC, etc.

[0036] Hereinafter, embodiments of the present invention will be described in detail with reference to FIGS. 1 to 4.

[0038] FIG. 1 is a drawing illustrating an information generating device according to one embodiment of the present invention.

[0039] An information generating device (100) according to one embodiment may include a processor (110), a memory (120) including instructions (122), and a communication unit (130).

[0040] The processor (110) can execute software and control at least one other component (e.g., a hardware or software component) connected to the processor (110). The processor (110) can also perform various data processing or operations.

[0041] For reference, the processor (110) can perform all operations performed by the information generating device (100). Therefore, for convenience of explanation, the operations performed by the information generating device (100) are mainly described as operations performed by the processor (110).

[0042] Additionally, for convenience of explanation in this specification, the processor (110) is primarily described as a single processor, but is not limited thereto. For example, the information generating device (100) may include processors.

[0043] The memory (120) can temporarily and / or permanently store various data and / or information required to perform the operation of generating patent analysis results by receiving one or more patent documents, extracting patent information therefrom, generating grouping information by combining it with a prompt corresponding to the purpose of analysis entered by the user, and applying it to a generative artificial intelligence-based patent analysis model.

[0044] Generative artificial intelligence refers to AI technology that performs functions such as generating new information based on input data or prompts, or processing, summarizing, and interpreting existing information. For example, generative AI can represent a Large Language Model (LLM) specialized for natural language processing and is characterized by high-level language generation capabilities that can understand the meaning of sentences and generate text tailored to the user's purpose.

[0045] In the present invention, a generative artificial intelligence receives technical information extracted from a patent document and a prompt together, and is utilized to generate various patent analysis results, such as technical classification, keyword extraction, technical summary, term interpretation, and context analysis, according to the purpose of analysis. Through this, the processor (110) enables context-based interpretation and summarization of complex technical content, which were difficult for existing learning-type AI to provide, thereby significantly improving the efficiency and precision of patent search and analysis tasks.

[0046] The communication unit (130) can support communication between the information generating device (100) and the external server (140). For example, the communication unit (130) may include one or more components that enable communication between the information generating device (100) and the external server (140). For example, the communication unit (130) may include a short-range wireless communication unit, a microphone, etc. At this time, short-range communication technologies may include wireless LAN (Wi-Fi), Bluetooth, Zigbee, WFD (Wi-Fi Direct), UWB (ultra-wideband), infrared communication (IrDA, infrared Data Association), BLE (Bluetooth Low Energy), NFC (Near Field Communication), etc., but are not limited thereto.

[0048] FIG. 2 is a flowchart illustrating a method for extracting patent analysis results from a target document in an information generation device according to an embodiment of the present invention.

[0049] A processor according to one embodiment (e.g., processor (110) of FIG. 1) can extract patent information from a target document in operation S210.

[0050] For example, the target document refers to one or more patent-related documents subject to patent analysis and may include application publications, registration publications, international publications, examination history documents, cited literature, or publicly disclosed patent documents of a similar format. The target document may be uploaded directly by a user, collected from a patent database by a search query, or identified through an artificial intelligence-based recommendation system.

[0051] For example, the target document may be a single document or a set of multiple documents, and may be selectively configured depending on the purpose of analysis. The target document may represent input data processed by the processor of the patent information generation device.

[0052] For example, patent information may refer to data elements containing technical, legal, and administrative content unique to the patent, extracted from a target document. Patent information may be extracted in whole or in part depending on the purpose of analysis, and the extraction method may be performed using user-specific, AI-based extraction, or recommendation-based methods. Specifically, the processor may extract patent information from a target document according to the purpose of extraction. The processor may identify items that can be extracted from the target document and extract patent information from the target document based on the user's selection.

[0053] Specifically, patent information may include not only bibliographic information such as application number, filing date, publication date, registration date, International Patent Classification (IPC) code and CPC classification code, inventor, applicant, and right holder, but also technical information such as description of invention, claims, drawings, and abstract, and legal or procedural information such as the date of request for examination, notice of grounds for rejection, registration status, and litigation history.

[0054] The processor can generate grouping information through a combination of patent information and a target prompt based on determining a target prompt indicating an analysis purpose from the user in operation S220.

[0055] For example, the purpose of analysis refers to the technical or legal objectives of the analysis that the user intends to achieve by utilizing patent information. The purpose of analysis can be defined in various ways depending on the user's inquiries or requirements, and may, for example, consist of reviewing prior art, comparing similar patents, identifying application trends, interpreting the scope of rights, identifying technological trends, or analyzing differences between claims.

[0056] For example, the purpose of analysis serves as a criterion for determining the output format of the analysis results, the precision of the analysis, and the range of comparison targets, and can act as a key element of the query conditions or instructions required for generative artificial intelligence to perform analysis tasks.

[0057] For example, a target prompt refers to an input instruction that directs the direction and method of the analysis task to be performed by the generative AI, reflecting the analysis purpose defined by the user or the system. The target prompt is expressed in natural language and can specify or elicit the purpose of the analysis, the scope of comparison, the output format, interpretation criteria, etc.

[0058] For example, the target prompt is determined based on the system prompt and the user prompt, and may include conditions regarding how to perform interpretation or analysis on a specific patent document or patent information.

[0059] For example, grouping information refers to information that organizes the correspondence relationship between a target prompt determined according to the purpose of analysis and a target document or patent information extracted therefrom into a single unit of analysis. In grouping information, one or more pieces of patent information may be linked to a single target prompt, and conversely, multiple prompts may be linked to a single piece of patent information.

[0060] In addition, grouping information may be composed of a combination between a representative document and a comparison target document, or a combination between already stored analysis results and new instructions.

[0061] In operation S230, the processor can obtain a target response by applying grouping information to a patent analysis model through batch processing.

[0062] For example, the processor can apply grouping information to a patent analysis model to extract a patent analysis result corresponding to the grouping information.

[0063] Specifically, the processor can transmit (or input) multiple grouping information in batches to a patent analysis model. For example, the processor may organize a combination of first patent information and a target prompt, a combination of second patent information and the same target prompt, and a combination of Nth patent information and a target prompt into a single analysis unit, and transmit and / or input them in a batch manner to a patent analysis model including a Large Language Model (LLM). The processor may implement a method of transmitting input to the patent analysis model using an Application Programming Interface (API) for communication.

[0064] In operation S240, the processor can store the target response received from the patent analysis model by matching it with the grouping information.

[0065] Additionally, the processor can receive multiple responses corresponding to each grouping information from the patent analysis model, and can store the received responses by matching them with the first to Nth patent information corresponding to the input order. That is, the processor can maintain a correspondence relationship between the analysis target and the analysis result by distinguishing and linking the response corresponding to the first patent information, the response corresponding to the second patent information, and the response corresponding to the Nth patent information.

[0066] The processor can extract at least one patent analysis result by parsing the target response according to a predetermined output rule in operation S250.

[0067] For example, the processor can extract patent analysis results by parsing the response according to predefined rules. The aforementioned rules may be defined by output format instructions included in the target prompt, and, for example, if the generative AI is configured to include key analysis information between specific symbols (e.g., “@@”, “@#”, etc.), it can identify and extract the text between the symbols.

[0068] In addition, if the patent analysis model is configured to generate a response in a structured file format such as JSON (JavaScript Object Notation), the processor can determine the final output information by interpreting the format and extracting the value corresponding to a specific key.

[0069] In this way, the processor can derive patent analysis results suitable for the purpose of analysis by processing multiple grouping information in batches, sorting the received responses according to correspondence relationships, and interpreting the responses based on format rules defined by the prompt.

[0070] For example, a patent analysis model may refer to a model that includes generative artificial intelligence or a large language model (LLM) for interpreting the meaning of the patent information and generating a target response suitable for the purpose of analysis, based on the patent information included in the grouping information and the target prompt.

[0071] For example, the patent analysis model may be a commercial or open-source large-scale language model running in an external server environment, or a standalone analysis engine running in a local environment. The patent analysis model can understand technical content and legal expressions within patent documents expressed in natural language, grasp contextual meaning based on analysis instructions presented at target prompts, and generate target responses in various forms such as text, tables, images, or graphs.

[0072] Specifically, the processor can train a patent analysis model. For example, the patent analysis model may include a neural network. The neural network may include multiple layers, and each layer may include multiple nodes. A node may have a node value determined based on an activation function. A node in any layer may be connected to a node in another layer (e.g., another node) through a link (e.g., a connection edge) having a connection weight. A node's node value may be propagated to other nodes through the link. In the inference operation of the neural network, node values ​​may be forward propagated from the previous layer toward the next layer.

[0073] For example, in a patent analysis model, the forward propagation operation can represent an operation that propagates node values ​​based on input data in the direction from the input layer to the output layer of the patent analysis model. That is, the node value of a node can be propagated (e.g., forward propagation) to a node in the next layer (e.g., the next node) connected to the node via a connection line. For example, a node can receive a value weighted by connection weights from a previous node (e.g., multiple nodes) connected via a connection line.

[0074] For example, the node value of a node can be determined based on applying an activation function to the sum of weighted values ​​received from previous nodes (e.g., weighted sum). The parameters of the neural network may include the aforementioned connection weights as an example. The parameters of the neural network can be updated so that the objective function value described below changes in a targeted direction (e.g., the direction in which loss is minimized).

[0075] For example, a learned patent analysis model may represent a model learned through machine learning and may be a learned machine learning model that outputs a training output from a training input. The machine learning model (e.g., the learned patent analysis model) may be generated through machine learning. The learning algorithm may include, for example, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but is not limited to the examples described above.

[0076] For example, a machine learning model may include multiple artificial neural network layers. The artificial neural network may be at least one combination of a deep neural network (DNN), a convolutional neural network (CNN), a U-Net (U-Net for Image Segmentation), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), deep Q-networks, or a combination thereof, but is not limited to the examples described above.

[0077] For example, in the case of supervised learning, the machine learning model described above may be trained based on training data comprising pairs of training inputs and training outputs mapped to those training inputs. For example, the machine learning model may be trained to output a training output from a training input. During training, the machine learning model may generate a temporary output in response to the training input and may be trained to minimize the loss between the temporary output and the training output (e.g., the target of training). During the training process, the parameters of the machine learning model (e.g., connection weights between nodes / layers in a neural network) may be updated according to the loss. This training may be performed, for example, on the information generating device itself where the machine learning model is executed, or through a separate server. The machine learning model that has completed training (e.g., the trained patent analysis model) may be stored in memory (e.g., memory (120) of FIG. 1).

[0078] For example, patent analysis results refer to information selected or reconstructed according to user requirements from among target responses generated by inputting grouping information into a patent analysis model, and may include a format and structure suitable for the purpose of analysis. Patent analysis results may include information representations suitable for various analysis purposes, such as, for example, comparison tables between technical features, the presence or absence of similar claims, interpretation of correlations between technical components, summaries of application trends, and summaries of the scope of rights.

[0079] For example, patent analysis results can be configured in the form of text, numerical matrix, visual image, graphical graph, etc., and can be modified in various ways depending on user output settings or system settings.

[0080] In addition, the patent analysis results may refer to a result that is utilized in subsequent components within the information generation device (e.g., output unit, storage unit, interactive analysis unit, report generation unit, etc.) to realize a knowledge-based output that meets the purpose of the analysis.

[0082] FIG. 3 is a flowchart illustrating a method for extracting patent analysis results based on a target document in an information generation device according to an embodiment of the present invention.

[0083] A processor according to one embodiment (e.g., processor (110) of FIG. 1) can identify user input and database in operation S310.

[0084] For example, the processor may or may not be connected to a patent database. If the processor is connected to a patent database, it may collect target documents based on a search query or collect target documents based on an artificial intelligence recommendation method.

[0085] For example, the processor can upload a pre-determined target document without using a patent database. If not connected to a patent database, the processor can upload pre-determined patent information.

[0086] For example, user input refers to information or instruction commands provided by a user using an information generating device. User input may be received through various interfaces such as a keyboard, touchscreen, voice recognition, or API calls over a network, and may include instructions for analysis purposes, input of search queries, selection of prompts, uploading of patent documents, or requests for specific patents.

[0087] For example, a patent database refers to a data repository or searchable server system in which one or more patent documents are stored, and includes patent information repositories provided by public or private institutions. A patent database includes application publications, registration publications, internationally published documents, citation information, legal status, classification codes, and related metadata, and may be configured to enable searching or linking through an external network.

[0088] For example, a patent database can be utilized as a collection path for target documents based on search queries or AI recommendation methods, and may or may not be connected to the system as needed. If not connected, the system sets documents uploaded directly by the user as processing targets.

[0089] For example, a search expression refers to a text-based query command written by a user to query a patent database and extract relevant patent documents. A search expression is constructed by combining conditions such as keywords, Boolean operators, filing dates, or classification codes, and users can refine or filter the search targets to suit the purpose of analysis. For instance, a search expression targeting patents containing specific technical keywords, patents containing specific applicants, or documents filed during a specific period may be used. The search expression is entered into the patent database and can be used to extract a set of related patent documents.

[0090] For example, an AI-based recommendation method is a method in which an AI model suggests highly relevant patent documents based on user input information, past search history, technical fields, analysis purposes, etc. If the user has not explicitly entered a search query (i.e., user input), the processor can extract target documents by utilizing a keyword recommendation model or a vector-based similarity model based on user account information or previous input data.

[0091] For example, the AI ​​recommendation method can improve user convenience and the speed of analysis initiation, and enable the system's autonomous document selection function. The AI ​​recommendation method can operate in conjunction with a patent database, and the generated recommendation results can be confirmed as target documents after user review.

[0092] The processor can extract the target document from user input and the database in operation S320.

[0093] For example, when a processor receives a search query from a user, it can apply the search query to a database to extract target documents.

[0094] For example, if the processor does not receive a search expression, it can apply the output information obtained by applying the user's identification information to a keyword recommendation model to the database to extract the target document.

[0095] For example, a processor may receive a search query from a user that includes the keyword "AI-based voice recognition technology." By applying the search query to a patent database, the processor may identify one or more patent documents in which the keyword is included in the abstract, claims, or description of the invention. Each identified document is set as a target document, and in a subsequent step, patent information is extracted and utilized as a subject for analysis. In this case, the search query may be written in a form such as, for example, "(AI AND voice recognition) AND IPC=G10L," and the processor may filter a set of documents within the database according to the search logic.

[0096] For example, if the user has not entered an explicit search query, the processor can apply identifying information associated with the user account (e.g., areas of technical interest, past search history, saved analysis settings, etc.) to the keyword recommendation model. For instance, if the user has a history of frequently analyzing technologies related to "image sensors" or "TOF sensors," the processor can derive associated keywords such as "depth camera," "optical sensing," and "3D imaging."

[0097] For example, the derived keywords are converted into search conditions and applied to the patent database, and one or more patent documents containing the keywords are extracted as target documents.

[0098] For example, one or more target documents generated by a search query may include patent information containing keywords based on a search query written in the same manner as existing methods. One or more patent documents generated by artificial intelligence or a recommendation method may extract one or more patent documents by using artificial intelligence based on information entered by a user, or extract one or more patent documents by sequentially proceeding with a method of suggestion based on information entered by a user. The predetermined patent documents may include a patent list that identifies one or more patent documents using application numbers, etc., and can be entered into the system.

[0099] The processor can extract patent information from the target document in operation S330.

[0100] For example, the processor can extract from the target document bibliographic information including an application number, filing date, classification code, and inventor, and patent information including a description of invention, claims, legal status, and litigation information.

[0101] For example, the processor extracts one or more patent information from patent documents to be applied to a patent analysis model. The information to be extracted can be selected directly by the user, extracted by artificial intelligence according to the purpose, or recommended to the user.

[0102] The processor can determine the target prompt based on the system prompt and the user prompt in operation S340.

[0103] For example, the processor can identify a system prompt regarding a predetermined form of an analysis method or output format for processing patent information and a user prompt specified by the user. Subsequently, the processor can determine a target prompt based on the system prompt and the user prompt.

[0104] For example, the processor is configured to generate prompts, be input by artificial intelligence, or provide recommendations to the user for the purpose of analyzing patent information.

[0105] In particular, the system prompt may include basic prompts for defined analysis methods, such as how the patent analysis model processes patent information or the output data format. Additionally, the user prompt may include prompts for additional analysis in addition to the system prompt.

[0106] The processor can generate grouping information through a combination of patent information and target prompts in operation S350.

[0107] For example, the processor may apply grouping information to a patent analysis model including a large language model (LLM) operating on an external server or in a local environment to generate a target response including analysis information regarding the grouping information.

[0108] Specifically, the patent information may be patent information extracted from a single target document, information extracted from one or more target documents, or patent information that has already been extracted. The extracted patent information may be information generated by the information generating device according to the present invention or information generated through another source.

[0109] For example, the processor generates grouping information by grouping patent information with target prompts. In this case, one grouping information may be generated, or one or more grouping information may be generated.

[0110] For example, the processor can generate multiple grouping information by grouping patent information extracted from multiple target documents and each target prompt.

[0111] For example, the processor can generate multiple grouping information by grouping the patent information of a representative document among multiple target documents and the patent information of the remaining other target documents with a target prompt, respectively.

[0112] For example, the processor can generate multiple grouping information by grouping patent information of one or more of the multiple target documents with a target prompt, respectively.

[0113] For example, the processor can generate one grouping information by grouping patent information of multiple target documents with one target prompt.

[0114] For example, the processor can generate a single grouping information by grouping already extracted patent information with the target prompt, in addition to the information extracted from the target document.

[0115] For example, the processor can generate grouping information by grouping patent information and one or more target prompts.

[0116] For example, the processor can generate grouping information by grouping patent information of a single target document and target prompts having different purposes, respectively.

[0117] The processor can apply grouping information to a patent analysis model in operation S360 to extract patent analysis results.

[0118] For example, the processor can extract patent analysis results by converting the target response into at least one of text, matrix, image, or graph.

[0119] For example, the processor may generate analyzed information (i.e., patent analysis results) for each of the grouping information, including patent information and target prompts, based on a patent analysis model operating on an external server or in memory. The analyzed information may be in the form of text or a matrix, and may include images, graphs, etc.

[0120] The processor can combine patent information and patent analysis results in operation S370 and provide output information to the user.

[0121] For example, the processor may match grouping information with patent analysis results generated in correspondence and finally determine output information and an output format. The output information may include patent information included in the grouping information and analysis results derived from the target response of the patent analysis model, and in some cases, may include only the analysis results.

[0122] For example, the output information may be configured in a form that combines the application number and all or part of the analysis results, or it may be configured to include only some or all of the analysis results.

[0123] For example, output information may be output in at least one of the following forms: visual display via a web page, creation in the form of a file, or delivery to a user via electronic mail or social network service. Here, the output form may be determined according to predefined settings or may be selectively determined by artificial intelligence processing logic within the system.

[0124] That is, the processor can output the output information in at least one of a web page, a file, electronic mail, or a social network service, based on determining the output information by combining patent information and patent analysis results.

[0125] Additionally, the processor stores at least one of the patent analysis results, target prompts, and grouping information in memory, and can perform updates by identifying updateable items, including a registration date, request for examination status, or right holder, from the patent analysis results stored in memory. For example, the processor can identify updated patent information among the stored patent information and add a new registration date or change the request for examination status, right holder, etc.

[0126] For example, the processor may receive input including a user's question or instruction and, based on patent analysis results and patent information stored in memory, provide an output corresponding to the input to the user through an interactive user interface.

[0127] That is, the processor can generate and provide a target response corresponding to user input through an interactive user interface based on patent analysis results and patent information stored in memory. Here, user input is not limited to simple questions and may include instructions for adjusting the target response format, output method, analysis scope, etc.

[0128] For example, the processor may generate one or more analysis reports using patent analysis results and related information stored in memory. The analysis reports may include various reports related to patent information, such as a prior art search report, an application trend report, a technology flow report, and an examination history report.

[0129] For example, the processor may generate a target response by referring to a predetermined prompt corresponding to the type of analysis report and inputting the prompt and information stored in the patent analysis storage unit into generative artificial intelligence.

[0130] For example, the generated target response can be inserted into the corresponding analysis report according to a predefined method or a user-specified custom configuration method. Here, the configuration format of the report may be predetermined based on system settings or may vary according to a template customized by the user.

[0131] In addition, the processor can construct an analysis report by extracting the entire content of the patent document corresponding to the analysis target and inputting it into a generative artificial intelligence along with a prompt to obtain the analysis target response.

[0133] FIG. 4 is a drawing illustrating an example of patent information or patent analysis results extracted from a target document in an information generation device according to one embodiment of the present invention.

[0134] A processor according to one embodiment (e.g., processor (110) of FIG. 1) can generate and / or extract patent information (410) extracted from a target document, a target prompt (420), grouping information (430) generated based on a combination thereof, a target response (440) to the grouping information (430), and a final patent analysis result (450).

[0135] For example, patent information (410) is structured information extracted from one or more target documents and may include key technical information such as the application number, filing date, claims, description of the invention, and rights holder of each patent. Patent information (410) is obtained from a patent document selected by a user or system and is used as basic data for subsequent analysis.

[0136] For example, the target prompt (420) includes instructions defined according to the user's analysis purpose. For example, the target prompt (420) includes a set of commands to be provided to the generative artificial intelligence according to the analysis purpose, such as matching based on technology classification criteria, extraction of core keywords, technology summary, and context analysis, and may be configured in the form of a system prompt or a user prompt.

[0137] For example, grouping information (430) is combination data generated by combining patent information (410) and target prompt (420) into a single processing unit. Grouping information (430) may, for example, be composed of "Prompt A and Patent Information (410) A", "Prompt A and Patent Information (410) B", etc., and each combination is transmitted to the patent analysis model as a single analysis unit.

[0138] For example, the target response (440) represents result data generated by the patent analysis model. Depending on each grouping information (430), the patent analysis model can return a response according to the specified analysis purpose.

[0139] That is, the target response (440) may include various forms of analysis information such as keywords, classifications, summarized technical content, and key sentences.

[0140] For example, the patent analysis result (450) may refer to an output result that processes or visualizes the target response (440) into a form that is easy for humans to understand. The patent analysis result (450) may be implemented in the form of a table on a user interface, a downloadable file format, or a report item, and may include various patent analysis elements such as application number, technical summary, major IPCs, and related companies.

[0141] Accordingly, FIG. 4 specifically illustrates a processing process for efficiently extracting information suitable for the purpose of analysis from a patent document and generating and outputting a structured analysis result based on the response of a patent analysis model.

[0142] Additionally, a computing system relating to an information generating device (e.g., the information generating device (100) of FIG. 1) or an information generating method according to one embodiment may include at least one processor, memory, user interface input device, user interface output device, storage, and network interface connected via a bus.

[0143] A processor may be a semiconductor device that executes processing on instructions stored in a central processing unit (CPU) or in memory and / or storage. Memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include read-only memory (ROM) and random access memory (RAM). Accordingly, steps of a method or algorithm described in connection with the embodiments disclosed herein may be directly implemented by hardware, software modules, or a combination of both, executed by the processor. Software modules may reside in storage media such as RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disks, removable disks, and CD-ROMs.

[0144] An exemplary storage medium is coupled to a processor, and the processor can read information from the storage medium and write information to the storage medium. Alternatively, the storage medium may be integrated with the processor. The processor and the storage medium may reside within an Application-Specific Integrated Circuit (ASIC). The ASIC may reside within a user terminal. Alternatively, the processor and the storage medium may reside as separate components within the user terminal.

[0146] The above description is merely an illustrative explanation of the technical concept of the present invention, and those skilled in the art to which the present invention pertains will be able to make various modifications and variations within the scope of the essential characteristics of the present invention.

[0147] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using a general-purpose computer or a special-purpose computer, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing instructions and responding to targets. The processing unit may execute an operating system (OS) and software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0148] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on computer-readable recording media.

[0149] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination, and the program instructions recorded on the medium may be those specifically designed and configured for the embodiment or those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

[0150] The hardware device described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.

[0151] Although the embodiments described above have been explained with reference to limited drawings, those skilled in the art can apply various technical modifications and variations based thereon. For example, appropriate results can be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

[0152] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below.

[0153] Accordingly, the embodiments disclosed in this invention are intended to illustrate, not limit, the technical concept of the invention, and the scope of the technical concept of the invention is not limited by these embodiments. The scope of protection of this invention shall be interpreted by the claims below, and all technical concepts within an equivalent scope shall be interpreted as being included within the scope of rights of this invention.

Claims

Claim 1 An information generating device comprising: a memory storing computer-executable instructions; and a processor that accesses the memory and executes the instructions, wherein the processor extracts patent information from a target document, generates grouping information through a combination of the patent information and the target prompt based on determining a target prompt indicating an analysis purpose from a user, applies the grouping information to a patent analysis model through batch processing to obtain a target response, stores the target response received from the patent analysis model by matching it to correspond with the grouping information, and extracts at least one patent analysis result from the target response by parsing according to a predetermined output rule. Claim 2 In claim 1, the information generating device, wherein the processor, when receiving a search query from the user, applies the search query to a database to extract the target document, and when not receiving the search query, applies output information obtained by applying the user's identification information to a keyword recommendation model to the database to extract the target document. Claim 3 In claim 1, the processor is an information generating device that extracts from the target document bibliographic information including an application number, filing date, classification code, and inventor, and patent information including a description of invention, claims, legal status, and litigation information. Claim 4 In claim 1, the processor identifies a system prompt regarding a predetermined form of an analysis method or output format for processing the patent information and a user prompt designated by the user, and determines the target prompt based on the system prompt and the user prompt. Claim 5 An information generating device according to claim 1, wherein the processor applies the grouping information to the patent analysis model including a large language model (LLM) operating in an external server or local environment to generate the target response including analysis information regarding the grouping information, and extracts the patent analysis result by converting the target response into at least one form among text, matrix, image, or graph. Claim 6 In claim 1, the information generating device, wherein the processor determines output information by combining the patent information and the patent analysis result, and outputs the output information in at least one form among a web page, a file, electronic mail, or a social network service. Claim 7 An information generating device according to claim 1, wherein the processor stores at least one of the patent analysis result, the target prompt, and the grouping information in the memory, and identifies and updates updatable items including a registration date, a request for examination status, or a right holder from the patent analysis result stored in the memory. Claim 8 In claim 7, the processor receives an input including a question or instruction from the user, and provides an output corresponding to the input to the user through an interactive user interface based on the patent analysis result and the patent information stored in the memory. Claim 9 A method for generating information comprising: an operation of extracting patent information from a target document; an operation of generating grouping information through a combination of said patent information and said target prompt based on determining a target prompt indicating an analysis purpose from a user; an operation of obtaining a target response by applying said grouping information to a patent analysis model through batch processing; an operation of storing said target response received from said patent analysis model by matching it to correspond with said grouping information; and an operation of extracting at least one patent analysis result from said target response by parsing according to a predetermined output rule. Claim 10 In claim 9, the operation of extracting the patent information comprises: an operation of applying the search query to a database to extract the target document when the search query is received from the user; and an operation of applying output information obtained by applying the user's identification information to a keyword recommendation model to the database to extract the target document when the search query is not received. Claim 11 In claim 9, the operation of extracting the patent information comprises, from the target document, an operation of extracting bibliographic information including an application number, filing date, classification code, and inventor, and the patent information including a description of invention, claims, legal status, and litigation information. Claim 12 In claim 9, the operation of generating the grouping information comprises: an operation of identifying a system prompt and a user prompt specified by the user regarding a predetermined form of an analysis method or output format for processing the patent information; and an operation of determining the target prompt based on the system prompt and the user prompt. Claim 13 In claim 9, the operation of extracting the patent analysis result comprises: applying the grouping information to the patent analysis model including a large language model (LLM) operating in an external server or local environment to generate the target response including analysis information regarding the grouping information; and the operation of extracting the patent analysis result by converting the target response into at least one form among text, matrix, image, or graph. Claim 14 In claim 9, the operation of extracting the patent analysis result comprises an operation of outputting the output information in at least one form among a web page, a file, electronic mail, or a social network service, based on determining output information by combining the patent information and the patent analysis result. Claim 15 In claim 9, the operation of extracting the patent analysis result comprises: the operation of storing at least one of the patent analysis result, the target prompt, and the grouping information in memory; and the operation of identifying and updating updatable items including a registration date, a request for examination status, or a right holder from the patent analysis result stored in memory. Claim 16 In claim 15, the operation of extracting the patent analysis result comprises: receiving an input including a question or instruction from the user; and providing an output corresponding to the input to the user through an interactive user interface based on the patent analysis result and the patent information stored in the memory.

Images