Information processing system and information processing method

Abstract

The present invention provides a novel information processing system that has enhanced convenience, usefulness, and reliability. The information processing system includes three components. The first component has the function of accepting a question and transmitting the question to the third component and the function of accepting and providing an answer. The second component accepts a first directive and generates an answer using a large-scale language model. The third component has the function of creating the first directive and transmitting the first directive to the second component and includes a database and a search engine. The first directive includes the question, background information, and an instruction to generate the answer to the question with reference to the background information. The database includes original data described in a programming language or a hardware description language and a summary described in a natural language. The search engine collects records associated with the question from the database to create the background information.

Description

Information processing system and information processing method

[0001] One embodiment of the present invention relates to an information processing system, an information processing method, or a semiconductor device.

[0002] Note that one embodiment of the present invention is not limited to the above technical field. The technical field of one embodiment of the invention disclosed in this specification relates to an object, a method, or a manufacturing method. Alternatively, one embodiment of the present invention relates to a process, a machine, a manufacture, or a composition of matter. Therefore, more specifically, examples of the technical field of one embodiment of the present invention disclosed in this specification include a data processing device, a semiconductor device, a memory device, a driving method thereof, or a manufacturing method thereof.

[0003] In recent years, the development of language models using neural networks has been actively carried out, and large-scale language models (LLMs) in particular have attracted attention. A large-scale language model is a natural language processing model trained using a large amount of data. A large-scale language model can realize a dialogue model that responds to user instructions, for example. Non-Patent Document 1 discloses GPT-4 (Generative Pre-trained Transformer 4) (registered trademark) as a large-scale language model, and also discloses ChatGPT as a dialogue model.

[0004] The use of large-scale language models has significantly increased the capabilities of natural language processing models. However, as language models become larger, it is difficult to incorporate and operate language models in-house due to the equipment and cost involved. Therefore, one way to use language models is to use external services that provide language models.

[0005] Summary of ChatGPT / GPT-4 Research and Perspective Towards the Future of Large Language Models, Yiheng Liu et al. (Submitted on 4 Apr 2023, [online], Internet <URL: https: / / arxiv.org / abs / 2304.01852>

[0006] An object of one embodiment of the present invention is to provide a novel information processing system with excellent convenience, usefulness, or reliability, or to provide a novel information processing method with excellent convenience, usefulness, or reliability, or to provide a novel information processing system, a novel information processing method, or a novel semiconductor device.

[0007] Note that the description of these problems does not preclude the existence of other problems. Note that one embodiment of the present invention does not necessarily solve all of these problems. Note that problems other than these will become apparent from the description of the specification, drawings, claims, etc., and it is possible to extract other problems from the description of the specification, drawings, claims, etc.

[0008] (1) One aspect of the present invention is an information processing system having a first component, a second component, and a third component.

[0009] The first component has a function of accepting and sending a question to the third component, and a function of accepting and providing an answer.

[0010] The second component has a function of receiving the first instruction sentence and sending a response to the third component, and a function of performing processing using a large-scale language model, where the large-scale language model has a function of generating a response in accordance with the first instruction sentence.

[0011] The third component has a function of receiving a question and sharing it within the third component, a function of creating a first instruction sentence and sending it to the second component, and a function of receiving an answer and sending it to the first component. The third component also has a first subcomponent and a second subcomponent.

[0012] The first subcomponent includes a database and a search engine. The database includes a first record, the first record including first raw data and a first summary. The first raw data is described in a programming language or a hardware description language. The first summary includes a summary of the first raw data, and the first summary is described in a natural language. The search engine also includes a function for collecting records related to the query from the database and a function for creating background information. When the first summary is related to the query, the background information includes the first summary.

[0013] The second subcomponent is responsible for creating a first instruction statement, the first instruction statement including a question, background information, and a first instruction, the first instruction including an instruction for generating an answer to the question by referencing the background information.

[0014] As a result, even if the first original data is written in an artificial language other than a natural language, such as a programming language or a hardware description language, the main points can be described in a first summary using natural language. Furthermore, a search can be performed using natural language to find the first summary. Furthermore, records containing summaries related to a question can be collected from a database. Furthermore, summaries related to a question can be collected from a database to serve as background information. Furthermore, an answer to a question can be generated by referring to the background information using a large-scale language model. Furthermore, original data related to a question can be provided. As a result, a novel information processing system with excellent convenience, usefulness, and reliability can be provided.

[0015] (2) Another aspect of the present invention is the information processing system described above, wherein the first component has a function of accepting a registration request and transmitting the request to the third component.

[0016] The registration request includes second original data, and the second original data is written in a programming language or a hardware description language.

[0017] The second component has a function of receiving the second instruction and sending the second summary to the third component. The large-scale language model has a function of generating the second summary in accordance with the second instruction. The second summary includes a summary of the second original data.

[0018] The third component also has a function of accepting a registration request and sharing the registration request within the third component, a function of creating a second instruction sentence and sending it to the second component, and a function of accepting a second summary and sharing it within the third component.

[0019] The third component also includes a third subcomponent that has a function of extracting second original data from the registration request and sharing the second original data within the third component, and a function of creating a second instruction statement. The second instruction statement includes the second original data and a second command, and the second command includes a command for generating a second summary written in a natural language from the second original data.

[0020] The first subcomponent also has a function of creating a second record and registering the second record in the database, the second record including the second original data and the second summary.

[0021] As a result, even if the second original data is written in an artificial language other than a natural language, such as a programming language or a hardware description language, the large-scale language model can be made to describe the main points of the data in a second summary using natural language. Furthermore, the second original data written in an artificial language other than a natural language can be associated with the second summary written in a natural language and registered in a database. As a result, a novel information processing system with excellent convenience, usefulness, and reliability can be provided.

[0022] (3) Another aspect of the present invention is the above-mentioned information processing system, wherein the third subcomponent has a function of identifying the type of the second original data and a function of selecting one command from a plurality of commands and adopting it as the second command according to the identified type.

[0023] This allows useful content to be included in the second summary depending on the type of the second original data, thereby providing a novel information processing system that is highly convenient, useful, and reliable.

[0024] (4) In accordance with another aspect of the present invention, the second component has a function of receiving a third instruction statement and sending a third summary to the third component, and a function of receiving a fourth instruction statement and sending the fourth summary to the third component. The large-scale language model has a function of generating the third summary in accordance with the third instruction statement and generating the fourth summary in accordance with the fourth instruction statement.

[0025] The third subcomponent also includes a preprocessing function. The preprocessing function includes a function of extracting a first chunk from the beginning of the second original data, a function of extracting a second chunk subsequent to the first chunk from the second original data, and a function of creating a third directive and a fourth directive. The third directive includes the first chunk and a second command, and the fourth directive includes the second chunk and the second command.

[0026] The third component also has a function of sending the third instruction sentence and the fourth instruction sentence to the second component in a predetermined order, and a function of accepting the third summary and the fourth summary in a predetermined order and sharing them within the third component.

[0027] The third component also has a function of creating a second instruction statement, which includes a third summary, a fourth summary, and the second instruction instead of the second original data and the second instruction.

[0028] This allows the second original data to be divided and processed even if it is very long. Furthermore, even if the second original data is very long, a second summary can be generated. As a result, a novel information processing system that is highly convenient, useful, and reliable can be provided.

[0029] (5) One aspect of the present invention is an information processing method including first to eighth steps.

[0030] In a first step, a first component accepts a question and sends it to a second component, where the question is written in natural language.

[0031] In the second step, the second component receives the question and shares it within the second component, where the second component includes a first subcomponent and a second subcomponent.

[0032] In a third step, the search engine of the first subcomponent collects records related to the query from a database to create background information. The database includes records, each of which includes original data and a summary. The original data is written in a programming language or a hardware description language, and the summary includes the gist of the original data and is written in a natural language. When the summary is related to the query, the background information also includes the summary.

[0033] In a fourth step, the second subcomponent creates and sends to the third component an instruction statement, the instruction statement including a question, background information, and instructions, including instructions for generating an answer to the question by referencing the background information.

[0034] In the fifth step, the third component accepts the instruction sentence and generates an answer using a large-scale language model.

[0035] In a sixth step, the third component sends a response to the second component.

[0036] In the seventh step, the second component accepts the response and sends it to the first component.

[0037] In the eighth step, the first component accepts and provides the answer.

[0038] As a result, even if the first original data is written in an artificial language other than a natural language, such as a programming language or a hardware description language, the main points can be described in a first summary using natural language. Furthermore, a search can be performed using natural language to find the first summary. Furthermore, records containing summaries related to a question can be collected from a database. Furthermore, summaries related to a question can be collected from a database to serve as background information. Furthermore, an answer to a question can be generated by referring to the background information using a large-scale language model. Furthermore, original data related to a question can be provided. As a result, a novel information processing method that is highly convenient, useful, and reliable can be provided.

[0039] (6) One aspect of the present invention is an information processing method having a first phase. Note that the first phase includes first to eighth steps.

[0040] In a first step of a first phase, a first component has a function of receiving a registration request and transmitting the registration request to a second component, the registration request including source data, the source data being written in a programming language or a hardware description language.

[0041] In the second step of the first phase, the second component accepts the registration request and shares it within the second component, where the second component includes a first subcomponent and a second subcomponent.

[0042] In the third step of the first phase, the second subcomponent extracts the original data from the registration request and shares it within the second component.

[0043] In the fourth step of the first phase, the second subcomponent creates and sends to the third component an instruction statement, the instruction statement including the original data and instructions, the instructions including instructions for generating a summary written in natural language from the original data.

[0044] In the fifth step of the first phase, the third component accepts the directive and generates a summary using a large-scale language model.

[0045] In the sixth step of the first phase, the third component sends the summary to the second component.

[0046] In the seventh step of the first phase, the second component accepts the summary and shares it within the second component.

[0047] In the eighth step of the first phase, the first subcomponent creates a first record and registers the first record in the database, where the first record includes the original data and the summary.

[0048] This allows the large-scale language model to describe the main points of original data in a natural language summary, even if the original data is written in an artificial language other than a natural language, such as a programming language or a hardware description language. Furthermore, the original data written in an artificial language other than a natural language can be associated with the summary written in a natural language and registered in a database. As a result, a novel information processing method that is highly convenient, useful, and reliable can be provided.

[0049] (7) Another aspect of the present invention is an information processing method having a first phase and a second phase, where the second phase follows the first phase and includes first to sixth steps.

[0050] In the first step of the second phase, the second component sends the first record to the first component.

[0051] In the second step of the second phase, the first component provides the first record and waits for input.

[0052] In the third step of the second phase, the first component terminates the process when approval is entered, and advances the process to the fourth step of the second phase when a revised document is entered.

[0053] In the fourth step of the second phase, the first component replaces the summary of the first record with the modified document to create a second record and sends it to the second component.

[0054] In the fifth step of the second phase, the second component accepts the second record and shares it within the second component.

[0055] In the sixth step of the second phase, the first subcomponent replaces the first record with the second record.

[0056] This allows, for example, a user of the information processing system to check whether the summary generated by the large-scale language model is valid. Furthermore, for example, a user of the information processing system can correct the summary generated by the large-scale language model. As a result, a novel information processing method that is highly convenient, useful, and reliable can be provided.

[0057] One embodiment of the present invention can provide a novel information processing system with excellent convenience, usefulness, or reliability, or a novel information processing method with excellent convenience, usefulness, or reliability, or a novel information processing system, a novel information processing method, or a novel semiconductor device.

[0058] Note that the description of these effects does not preclude the existence of other effects. Note that one embodiment of the present invention does not necessarily have all of these effects. Note that effects other than these will become apparent from the description in the specification, drawings, claims, etc., and it is possible to extract other effects from the description in the specification, drawings, claims, etc.

[0059] FIG. 1 is a diagram illustrating the configuration of an information processing system according to an embodiment. FIG. 2 is a diagram illustrating the configuration of components used in the information processing system according to an embodiment. FIGS. 3A and 3B are diagrams illustrating the configuration of a database and the configuration of background information according to an embodiment. FIG. 4 is a diagram illustrating the configuration of a directive statement according to an embodiment. FIG. 5 is a diagram illustrating the configuration of an information processing system according to an embodiment. FIG. 6A is a diagram illustrating the configuration of a registration request according to an embodiment, and FIG. 6B is a diagram illustrating the configuration of original data. FIG. 7 is a diagram illustrating the configuration of components used in the information processing system according to an embodiment. FIGS. 8A to 8C are diagrams illustrating the configuration of directive statements and record structures according to an embodiment. FIGS. 9A to 9C are diagrams illustrating the configuration of directive statements according to an embodiment. FIG. 10 is a diagram illustrating the configuration of an information processing device used in the information processing system according to an embodiment. FIG. 11 is a diagram illustrating an information processing method according to an embodiment. FIG. 12 is a diagram illustrating the information processing method according to an embodiment.

[0060] An information processing system according to one aspect of the present invention includes a first component, a second component, and a third component. The first component has a function of accepting a question and sending it to the third component, and a function of accepting and providing an answer. The second component has a function of accepting a first instruction sentence and sending the answer to the third component, and a function of performing processing using a large-scale language model, the large-scale language model having a function of generating an answer according to the first instruction sentence. The third component has a function of accepting a question and sharing it within the third component, a function of creating a first instruction sentence and sending it to the second component, and a function of accepting an answer and sending it to the first component. The third component also has a first subcomponent and a second subcomponent, the first subcomponent including a database and a search engine. The database includes a first record, the first record including first original data and a first summary. The first raw data is described in a programming language or a hardware description language, the first summary includes a gist of the first raw data, and the first summary is written in a natural language. The search engine has a function of collecting records related to the question from a database and a function of creating background information, and when the first summary is related to the question, the background information includes the first summary. The second subcomponent has a function of creating a first instruction statement, the first instruction statement including the question, the background information, and a first instruction, and the first instruction includes an instruction for generating an answer to the question by referring to the background information.

[0061] As a result, even if the first original data is written in an artificial language other than a natural language, such as a programming language or a hardware description language, the main points can be described in a first summary using natural language. Furthermore, a search can be performed using natural language to find the first summary. Furthermore, records containing summaries related to a question can be collected from a database. Furthermore, summaries related to a question can be collected from a database to serve as background information. Furthermore, an answer to a question can be generated by referring to the background information using a large-scale language model. Furthermore, original data related to a question can be provided. As a result, a novel information processing system with excellent convenience, usefulness, and reliability can be provided.

[0062] The embodiments will be described in detail with reference to the drawings. However, the present invention is not limited to the following description, and it will be readily understood by those skilled in the art that various changes in form and details can be made without departing from the spirit and scope of the present invention. Therefore, the present invention should not be interpreted as being limited to the description of the embodiments shown below. In the configuration of the invention described below, the same parts or parts having similar functions will be denoted by the same reference numerals in different drawings, and repeated explanations will be omitted.

[0063] In this specification, ordinal numbers such as "first" and "second" are used to avoid confusion between components and do not limit the number of components or the order of the components (for example, the order of processes or the order of stacking). Furthermore, even if a term does not have an ordinal number in this specification, an ordinal number may be added in the claims to avoid confusion between the components. Even if a term has an ordinal number in this specification, a different ordinal number may be added in the claims. Even if a term has an ordinal number in this specification, the ordinal number may be omitted in the claims.

[0064] In the drawings accompanying this specification, components are classified by function and shown as block diagrams that are independent of each other, but in reality, it is difficult to completely separate components by function, and one component may be involved in multiple functions.

[0065] Embodiment 1 In this embodiment, an information processing system according to one embodiment of the present invention will be described with reference to FIGS.

[0066] FIG. 1 is a diagram illustrating a configuration of an information processing system according to one embodiment of the present invention.

[0067] FIG. 2 is a diagram illustrating the configuration of components used in the information processing system according to the embodiment.

[0068] FIG. 3A is a diagram illustrating a configuration of a database used in an information processing system of one embodiment of the present invention, and FIG. 3B is a diagram illustrating a configuration of background information created by the information processing system of one embodiment of the present invention.

[0069] FIG. 4 is a diagram illustrating the configuration of instruction statements transmitted and received within an information processing system according to an aspect of the present invention.

[0070] FIG. 5 is a diagram illustrating a configuration of an information processing system according to one embodiment of the present invention.

[0071] FIG. 6A is a diagram illustrating the configuration of a registration request transmitted and received within an information processing system according to an aspect of the present invention, and FIG. 6B is a diagram illustrating the configuration of original data.

[0072] FIG. 7 is a diagram illustrating the configuration of components used in the information processing system according to the embodiment.

[0073] Figure 8A is a diagram explaining the structure of an instruction statement sent and received within an information processing system of one embodiment of the present invention, Figure 8B is a diagram explaining the structure of a record to be registered in a database, and Figure 8C is a diagram explaining the structure of a modified record.

[0074] 9A to 9C are diagrams illustrating the configuration of instruction statements transmitted and received within an information processing system according to one embodiment of the present invention.

[0075] FIG. 10 is a block diagram illustrating a configuration of an information processing device that can be used in an information processing system of one embodiment of the present invention.

[0076] <Configuration Example 1 of Information Processing System> The information processing system described in this embodiment includes a component 30, a component 20, and a component 21 (see FIG. 1 ). Note that the information processing device that performs the functions of component 30, the information processing device that performs the functions of component 21, and the information processing device that performs the functions of component 20 each include a calculation device and a communication device. Furthermore, they are connected using a network 51, for example.

[0077] <Configuration Example 1 of Component 30> The component 30 has a function of accepting a question Qre and transmitting it to the component 21. For example, a user of the information processing system inputs the question Qre to the component 30. Specifically, the user of the information processing system inputs the question Qre to the component 30 in natural language using an input device such as a keyboard, a mouse, an eye-gaze input device, or a microphone.

[0078] The component 30 also has a function of receiving an answer Ans from the component 21 and providing it to the user of the information processing system, for example.

[0079] <Configuration Example 1 of Component 20> The component 20 has a function of receiving a directive Pt-I and transmitting an answer Ans to the component 21, and a function of performing processing using the large-scale language model LLM.

[0080] <<Configuration Example 1 of Large-Scale Language Model LLM>> The large-scale language model LLM has a function of generating an answer Ans in accordance with a directive Pt-I. For example, a large-scale language model that has been trained on one or both of a programming language and a hardware description language, and a natural language, can be used for the large-scale language model LLM.

[0081] <Configuration example 1 of component 21> Component 21 has the function of accepting a question Qre and sharing it within component 21, the function of creating an instruction statement Pt-I and sending it to component 20, and the function of accepting an answer Ans and sending it to component 30 (see Figure 2).

[0082] The component 21 also includes a subcomponent 21 A and a subcomponent 21 B. For ease of explanation, in this specification, a configuration having a single function or multiple functions will be referred to as a component or a subcomponent.

[0083] <<Configuration Example 1 of Subcomponent 21A>> The subcomponent 21A includes a database DB and a search engine SE.

[0084] [Configuration Example of Database DB] The database DB includes one or more records, for example, n records (n is an integer greater than 1) with record ID_1 to record ID_n (see FIG. 3A).

[0085] For example, record ID_1 includes original data OD_1 and summary Sum_1. Furthermore, record ID_n includes original data OD_n. For example, data created by a user of an information processing system according to an embodiment of the present invention can be used as original data OD_1. Furthermore, data created by an organization to which a user of an information processing system according to an embodiment of the present invention belongs can be used as original data OD_1. Note that a distributed representation or embedded representation of original data OD_1 can also be included in record ID_1.

[0086] For example, data written in a programming language or a hardware description language can be used for the original data OD_1. Specifically, data written in a programming language such as Python, shell script, or Perl script can be used for the original data OD_1. Furthermore, data written in a register transfer level (RTL) hardware description language or data (e.g., a netlist) written in a gate-level hardware description language can be used for the original data OD_1. Note that data written in the native language of the user of the information processing system or another natural language can also be used for the original data OD_1.

[0087] The summary Sum_1 includes the gist of the original data OD_1 and is written in natural language.

[0088] [Configuration Example of Search Engine SE] The search engine SE has a function for collecting records related to the question Qre from the database DB. For example, the search engine SE can collect records including summaries containing a character string included in the question Qre, or records including summaries in which a character string included in the question Qre appears frequently. It can also collect records including summaries highly similar to the question Qre. When record ID_1 includes an embedded representation of the original data OD_1, the embedded representation of the question Qre can be compared with the embedded representation of the original data OD_1 to determine the semantic similarity. Using the embedded representation, the semantic similarity between the question Qre and the original data OD_1 can be evaluated. This can absorb variations in expressions due to synonyms, etc. In particular, it can absorb variations in verb expressions. For example, the similarity can be calculated using the Euclidean distance of the embedded representations to be compared. Furthermore, the search engine SE can calculate the cosine similarity of the embedded representations to be compared to search for original data OD_1 similar to the question Qre.

[0089] The search engine SE also has a function of creating background information BI. For example, a predetermined number of records selected in descending order of relevance to the question Qre can be used as the background information BI.

[0090] [Example of Configuration of Background Information BI] The background information BI includes summaries included in the adopted records. For example, when summaries Sum_1, Sum_X, and Sum_Y are related to the query Qre, the background information BI includes summaries Sum_1, Sum_X, and Sum_Y (see FIG. 3B ).

[0091] <<Configuration Example of Subcomponent 21B>> The subcomponent 21B has a function of creating a directive Pt-I.

[0092] [Configuration Example of Directive Statement Pt-I] The directive statement Pt-I includes a question Qre, background information BI, and a command g-I( ) (see FIG. 4).

[0093] The instruction g-I() includes an instruction to generate an answer Ans to the question Qre by referring to the background information BI. For example, the following paragraph of text can be used for the instruction Pt-I.

[0094] Please use the information below to answer the question. ##Background Information {Background Information BI} ##Question {Question Qre} ##Answer

[0095] The above command "Please answer the question using the information as a reference" corresponds to command g-I(). The background information BI is copied into {background information BI}, and the question Qre is copied into {question Qre}. The final command "##Answer" prompts the large-scale language model LLM to output.

[0096] As a result, even if the original data OD_1 is written in an artificial language other than a natural language, such as a programming language or a hardware description language, the main points can be described in the summary Sum_1 using natural language. Furthermore, the summary Sum_1 can be searched for using natural language. Furthermore, records containing summaries related to the question Qre can be collected from the database DB. Furthermore, summaries related to the question Qre can be collected from the database DB and used as background information BI. Furthermore, the large-scale language model LLM can be made to refer to the background information BI to generate an answer Ans to the question Qre. Furthermore, original data related to the question Qre can be provided. As a result, a novel information processing system with excellent convenience, usefulness, and reliability can be provided.

[0097] <Configuration Example 2 of Component 30> The component 30 also has a function of receiving a registration request Reg and transmitting it to the component 21 (see FIG. 5 ). For example, a user of the information processing system inputs the registration request Reg to the component 30. Specifically, the user of the information processing system inputs the registration request Reg to the component 30 using an input device such as a keyboard, a mouse, or an eye-gaze input device.

[0098] The registration request Reg includes original data OD_2 (see FIG. 6A). For example, data written in a programming language or a hardware description language can be used as the original data OD_2. Also, data written in a native language or other natural language can be used as the original data OD_2.

[0099] <Configuration Example 2 of Component 20> The component 20 has a function of receiving a directive Pt-II and transmitting a summary Sum_2 to the component 21 (see FIG. 5 ). The component 20 also has a function of performing processing using a large-scale language model LLM.

[0100] <<Configuration Example 2 of Large-Scale Language Model LLM>> The large-scale language model LLM has a function of generating a summary Sum_2 in accordance with a directive Pt-II. The summary Sum_2 includes the gist of the original data OD_2. The large-scale language model LLM can also generate a distributed representation or an embedded representation of the original data OD_2.

[0101] <Configuration Example 2 of Component 21> The component 21 has a function of receiving a registration request Reg from the component 30 and sharing the registration request Reg within the component 21.

[0102] The component 21 also has a function of creating an instruction statement Pt-II and transmitting it to the component 20. Details of the instruction statement Pt-II will be described later.

[0103] The component 21 also has a function of receiving the summary Sum_2 from the component 20 and sharing it within the component 21 .

[0104] The component 21 also includes a subcomponent 21C (see FIG. 7).

[0105] <<Configuration Example 1 of Subcomponent 21C>> The subcomponent 21C has a function of extracting the original data OD_2 from the registration request Reg and sharing it within the component 21, and a function of creating the instruction statement Pt-II.

[0106] [Configuration Example of Directive Statement Pt-II] The directive statement Pt-II includes original data OD_2 and command g-II() (see FIG. 8A). The command g-II() includes a command to generate a summary Sum_2 written in natural language from the original data OD_2. For example, the following paragraph of text can be used for directive statement Pt-II:

[0107] "Please provide a brief summary of the following source data. The source data describes an electronic circuit using the Verilog hardware description language. ##Source Data {Source Data OD_2} ##Summary"

[0108] The above command "Please summarize the original data in a simple sentence" corresponds to command g-II(). The original data OD_2 is copied into {original data OD_2}, and the final "## summary" prompts the large-scale language model LLM to output the data.

[0109] Also, for example, the following paragraph of text can be used for instruction sentence Pt-II.

[0110] Please provide a brief summary of the following source data. The source data contains a program written in Python. ##Source Data {Source Data OD_2} ##Summary

[0111] The above "Please summarize the original data in a simple sentence" corresponds to command g-II(). Also, "The original data contains a program written in Python." is information that clearly identifies the type of original data. Furthermore, the original data OD_2 is copied into {original data OD_2}, and the final "## summary" prompts the large-scale language model LLM for output.

[0112] <<Configuration Example 2 of Subcomponent 21A>> The subcomponent 21A has a function of creating a record ID_2 and registering the record ID_2 in the database DB. The record ID_2 includes original data OD_2 and a summary Sum_2 shared within the component 21 (see FIG. 8B ).

[0113] As a result, even if the original data OD_2 is written in an artificial language other than a natural language, such as a programming language or a hardware description language, the large-scale language model LLM can describe the main points of the original data in the summary Sum_2 using natural language. Furthermore, the original data OD_2 written in an artificial language other than a natural language can be associated with the summary Sum_2 written in a natural language and registered in the database DB. As a result, a novel information processing system with excellent convenience, usefulness, and reliability can be provided.

[0114] <<Configuration Example 2 of Subcomponent 21C>> The subcomponent 21C has a function of identifying the type of original data OD_2. For example, it can identify whether the original data OD_2 is written in a programming language or a hardware description language. Specifically, it can perform inference using a large-scale language model.

[0115] The subcomponent 21C also has a function of selecting one instruction from a plurality of instructions according to the identified type and adopting it as the instruction g-II( ).

[0116] For example, if the original data OD_2 is data written in a programming language, "Please summarize what purpose the following original data OD_2 is a program for." can be used in command g-II().

[0117] Also, for example, if the original data OD_2 is data written in a hardware description language, "Please summarize what purpose the following original data OD_2 is an electronic circuit for." can be used in the instruction g-II().

[0118] Furthermore, for example, if the original data OD_2 is data written in one's native language or another natural language, "Please summarize the following original data OD_2." can be used in command g-II( ).

[0119] This allows useful content to be included in the summary Sum_2 depending on the type of original data OD_2, thereby providing a novel information processing system that is highly convenient, useful, and reliable.

[0120] <Configuration Example 3 of Component 20> The component 20 has a function of receiving a directive Pt-II(1) and transmitting a summary Sum_2(1) to the component 21 (see FIG. 5 ). It also has a function of receiving a directive Pt-II(2) and transmitting a summary Sum_2(2) to the component 21.

[0121] <<Configuration Example 3 of Large-Scale Language Model LLM>> The large-scale language model LLM has a function of generating a summary Sum_2(1) according to a directive Pt-II(1) and generating a summary Sum_2(2) according to a directive Pt-II(2). Details of the directives Pt-II(1) and Pt-II(2) will be described later.

[0122] <<Configuration Example 3 of Subcomponent 21C>> The subcomponent 21C has a preprocessing function. The preprocessing function includes a function for extracting chunk Cnk(1) from the beginning of the original data OD_2 and a function for extracting chunk Cnk(2) following chunk Cnk(1) from the original data OD_2 (see FIG. 6B ). For example, if the original data OD_2 is very long, the original data OD_2 can be divided into data of a length that can be processed by the large-scale language model LLM. Specifically, the preprocessing function divides the original data OD_2 into m chunks, namely chunks Cnk(1) to Cnk(m) (m is an integer greater than 1). Note that a chunk can be a data of a predetermined length, a block of data separated by a predetermined delimiter, or a block of data with a predetermined structure. Furthermore, if a netlist written in a hardware description language is composed of multiple subcircuits, the description of one subcircuit can be one chunk.

[0123] [Configuration Example of Directive Statement Pt-II(1) and Directive Statement Pt-II(2)] The subcomponent 21C also has a function to create directive statements Pt-II(1) and Pt-II(2). For example, if the original data OD_2 is divided into n chunks, m directives, Pt-II(1) to Pt-II(m), are created.

[0124] Directive Pt-II(1) includes chunk Cnk(1) and command g-II() (see FIG. 9A). Directive Pt-II(2) includes chunk Cnk(2) and command g-II() (see FIG. 9B). Note that, for example, the command g-II() can be the command to generate a summary Sum_2 written in natural language from the original data OD_2.

[0125] <Configuration Example 3 of Component 21> The component 21 has a function of transmitting the directives Pt-II(1) and Pt-II(2) in a predetermined order to the component 20. For example, if the original data OD_2 is divided into n chunks, the component 21 transmits the n directives, from the directive including the first chunk of the original data OD_2 to the directive including the last chunk, in the order in which they appear in the original data OD_2.

[0126] The component 21 also has a function of accepting the summaries Sum_2(1) and Summarization Sum_2(2) in a predetermined order and sharing them within the component 21. For example, if n instruction sentences are sent in order, the component 21 accepts the n summaries in that order.

[0127] The component 21 has a function of creating a directive Pt-II. The directive Pt-II includes summaries Sum_2(1), Sum_2(2), and a command g-II() instead of the original data OD_2 and the command g-II(). For example, if n summaries are received in order, the directive Pt-II includes m summaries Sum_2(1) to Sum_2(m) and the command g-II() (see FIG. 9C ).

[0128] This allows the original data OD_2 to be divided and processed even if it is very long. Furthermore, even if the original data OD_2 is very long, a summary Sum_2 can be generated. As a result, a novel information processing system that is highly convenient, useful, and reliable can be provided.

[0129] <Configuration Example 2 of Information Processing System> The information processing system described in this embodiment includes a component 30, a component 21, and a component 20 (see FIGS. 1 and 5).

[0130] For example, an information processing system according to an embodiment of the present invention can be configured with an information processing device that performs the functions of component 30, an information processing device that performs the functions of component 21, and an information processing device that performs the functions of component 20. Note that the number of information processing devices that configure the information processing system according to an embodiment of the present invention is one or more. Furthermore, for example, the information processing system according to an embodiment of the present invention can be configured by connecting a plurality of information processing devices using a network 51.

[0131] When an information processing system according to one embodiment of the present invention is configured using a plurality of information processing devices, the load related to information processing can be distributed.

[0132] <Configuration Example 1 of Information Processing Device> Configuration Example 1 of the information processing device described in this embodiment can be used for the component 30. Configuration Example 1 of the information processing device can also be called a client computer, etc. For example, a desktop computer can be used for the component 30.

[0133] The information processing device according to the first exemplary configuration can accept data input by a user of the information processing system according to an embodiment of the present invention. The information processing device according to the first exemplary configuration can also provide the user with data output by the information processing system according to an embodiment of the present invention.

[0134] For example, dedicated application software, a web browser, etc., run on the component 30. A user of the information processing system according to an embodiment of the present invention can access the information processing system via either of these components, thereby enjoying services using the information processing system according to an embodiment of the present invention.

[0135] <Configuration Example 2 of Information Processing Apparatus> Configuration example 2 of the information processing apparatus described in this embodiment can be used for the component 21. For example, the component 21 can be a workstation, a server computer, a supercomputer, or the like.

[0136] Moreover, it is preferable that the information processing device in configuration example 2 has a function as a parallel computer. By using it as a parallel computer, it is possible to perform large-scale calculations necessary for learning and inference of artificial intelligence (AI), for example.

[0137] Furthermore, configuration example 2 of the information processing device can perform processing using a natural language model that uses AI.

[0138] For example, processing can be performed using natural language processing models such as GPT-3 (registered trademark), GPT-3.5, GPT-4 (registered trademark), LaMDA, Llama2, Llama3, and Codellama.

[0139] <Configuration Example 3 of Information Processing Device> For example, configuration example 3 of the information processing device described in this embodiment can be used for the component 20. Note that the component 20 is larger in scale and has higher computing power than the component 21. For example, a large computer such as a server computer or a supercomputer can be used for the component 20.

[0140] Moreover, it is preferable that the information processing device in configuration example 3 has a function as a parallel computer. By using it as a parallel computer, it is possible to perform large-scale calculations necessary for AI learning and inference, for example.

[0141] Furthermore, the information processing device according to the third exemplary configuration can perform processing using a natural language model that uses AI. In particular, the information processing device can perform processing using a general-purpose language model that can perform various natural language processing tasks.

[0142] For example, it is possible to perform processing using natural language models such as GPT-3 (registered trademark), GPT-3.5, GPT-4 (registered trademark), LaMDA, Llama2, Llama3, and Codellama. In particular, it is preferable to be able to perform processing using GPT-4 (registered trademark). For example, if it is possible to perform processing using a large-scale language model that is larger than conventional natural language models, it is possible to realize more natural document generation or dialogue.

[0143] Note that a person who provides a service using an information processing system according to one embodiment of the present invention does not necessarily have to own the information processing device of Configuration Example 3. For example, a service provider can use part of a service provided by another business or the like using the information processing device of Configuration Example 3.

[0144] <Configuration Example of Network 51> The network 51 that can be used in the information processing system of one embodiment of the present invention can connect multiple information processing devices. This allows the connected multiple information processing devices to transmit and receive data to and from each other. In addition, the load related to information processing can be distributed.

[0145] When wireless communication is performed, communication standards such as the fourth generation mobile communication system (4G), fifth generation mobile communication system (5G), and sixth generation mobile communication system (6G), or specifications standardized by the IEEE such as Wi-Fi (registered trademark) and Bluetooth (registered trademark), can be used as communication protocols or communication technologies.

[0146] For example, a local network can be used for the network 51. Also, an intranet or an extranet can be used for the network 51. Also, a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), a global area network (GAN), etc. can be used for the network 51.

[0147] Furthermore, for example, a global network can be used for the network 51. Specifically, the Internet, which is the foundation of the World Wide Web (WWW), can be used.

[0148] Furthermore, a person who provides a service using the information processing system according to one embodiment of the present invention can provide the service using the information processing method according to one embodiment of the present invention via the network 51, for example.

[0149] When the information processing system according to an embodiment of the present invention is built within a local network, the possibility of confidential information leaking can be reduced, for example, compared to when the Internet is used.

[0150] <Configuration Example 4 of Information Processing Device> An information processing device that can be used in an information processing system of one embodiment of the present invention includes, for example, an input unit 110, a memory unit 120, a processing unit 130, an output unit 140, and a transmission path 150 (see FIG. 10).

[0151] In the drawings accompanying this specification, the components are classified by function and shown as independent blocks in the block diagrams. However, in reality, it is difficult to completely separate the components by function, and one component may be involved in multiple functions. For example, part of the processing unit 130 may function as the input unit 110. Also, one function may be involved in multiple components. For example, the processing performed by the processing unit 130 may be executed by different information processing devices depending on the processing.

[0152] <<Input Unit 110>> The input unit 110 can receive data from outside the information processing device. For example, the input unit 110 receives data via the network 51.

[0153] The input unit 110 supplies the received data to one or both of the storage unit 120 and the processing unit 130 via the transmission path 150 .

[0154] <<Storage Unit 120>> The storage unit 120 has a function of storing a program executed by the processing unit 130. The storage unit 120 can also have a function of storing data generated by the processing unit 130 (e.g., calculation results, analysis results, inference results), data accepted by the input unit 110, etc.

[0155] The storage unit 120 may have a database. Furthermore, the information processing device may have a database separate from the storage unit 120. The information processing device may have a function to retrieve data from a database that exists outside the storage unit 120, outside the information processing device, or outside the information processing system. Furthermore, the information processing device may have a function to retrieve data from both its own database and an external database.

[0156] Either or both of a storage and a file server can be used as the memory unit 120. Also, a database that records paths of files stored in a file server can be used as the memory unit 120.

[0157] The storage unit 120 includes at least one of a volatile memory and a non-volatile memory. Examples of the volatile memory include a dynamic random access memory (DRAM) and a static random access memory (SRAM). Examples of the non-volatile memory include a resistive random access memory (ReRAM), a phase change random access memory (PRAM), a ferroelectric random access memory (FeRAM), a magnetoresistive random access memory (MRAM), and a flash memory. The storage unit 120 may include at least one of NOSRAM (registered trademark) and DOSRAM (registered trademark). The storage unit 120 may include a recording media drive. Examples of the recording media drive include a hard disk drive (HDD) and a solid state drive (SSD).

[0158] NOSRAM is an abbreviation for "Nonvolatile Oxide Semiconductor Random Access Memory (RAM)." NOSRAM refers to a memory in which memory cells are two-transistor (2T) or three-transistor (3T) gain cells and transistors (also called OS transistors) that use metal oxide in their channel formation regions. OS transistors have an extremely small leakage current, i.e., a current that flows between the source and drain in an off state. NOSRAM can be used as a nonvolatile memory by retaining a charge corresponding to data in the memory cell using its extremely small leakage current characteristic. In particular, NOSRAM can read stored data without destroying it (nondestructive readout), making it suitable for arithmetic processing in which only data read operations are repeated a large number of times. NOSRAM can increase its data capacity by stacking layers, and therefore can be used as a large-scale cache memory, main memory, or storage memory to improve the performance of semiconductor devices.

[0159] DOSRAM is an abbreviation for "Dynamic Oxide Semiconductor RAM" and refers to a RAM having 1T (transistor) 1C (capacitor) type memory cells. DOSRAM is a DRAM formed using OS transistors, and is a memory that temporarily stores information sent from an external device. DOSRAM is a memory that takes advantage of the low off-state current of OS transistors.

[0160] In this specification and the like, a metal oxide refers to an oxide of a metal in a broad sense. Metal oxides are classified into oxide insulators, oxide conductors (including transparent oxide conductors), oxide semiconductors (also referred to as oxide semiconductors or simply as OSs), and the like. For example, when a metal oxide is used for a semiconductor layer of a transistor, the metal oxide may be referred to as an oxide semiconductor.

[0161] The metal oxide contained in the channel formation region preferably contains indium (In). When the metal oxide contained in the channel formation region contains indium, the carrier mobility (electron mobility) of the OS transistor is increased. For example, indium oxide (InOx) or indium gallium zinc oxide (In—Ga—Zn oxide, also referred to as “IGZO”) can be used for the channel formation region. The metal oxide contained in the channel formation region is preferably an oxide semiconductor containing an element M. The element M is preferably at least one of aluminum (Al), gallium (Ga), and tin (Sn). Other elements that can be used for the element M include boron (B), silicon (Si), titanium (Ti), iron (Fe), nickel (Ni), germanium (Ge), yttrium (Y), zirconium (Zr), molybdenum (Mo), lanthanum (La), cerium (Ce), neodymium (Nd), hafnium (Hf), tantalum (Ta), and tungsten (W). However, the element M may be a combination of two or more of the above elements. The element M is, for example, an element having a high bond energy with oxygen. For example, the element M is an element having a higher bond energy with oxygen than indium. Furthermore, the metal oxide contained in the channel formation region is preferably a metal oxide containing zinc (Zn). Metal oxides containing zinc may be more likely to crystallize.

[0162] The metal oxide contained in the channel formation region is not limited to a metal oxide containing indium, but may be, for example, a metal oxide containing zinc but not indium, such as zinc tin oxide or gallium tin oxide, a metal oxide containing gallium, or a metal oxide containing tin.

[0163] <<Processing Unit 130>> The processing unit 130 has a function of performing processes such as calculation, analysis, and inference using data supplied from one or both of the input unit 110 and the storage unit 120. The processing unit 130 can supply generated data (e.g., calculation results, analysis results, and inference results) to one or both of the storage unit 120 and the output unit 140.

[0164] The processing unit 130 has a function of acquiring data from the storage unit 120. The processing unit 130 can also have a function of recording or registering data in the storage unit 120.

[0165] The processing unit 130 may include, for example, an arithmetic circuit. The processing unit 130 may include, for example, a central processing unit (CPU). The processing unit 130 may also include a graphics processing unit (GPU). The processing unit 130 may also include a neural processing unit / neural network processing unit (NPU).

[0166] The processing unit 130 may include a microprocessor such as a DSP (Digital Signal Processor). The microprocessor may be implemented by a PLD (Programmable Logic Device) such as an FPGA (Field Programmable Gate Array) or an FPAA (Field Programmable Analog Array). The processing unit 130 may also include a quantum processor. The processing unit 130 can perform various data processing and program control by interpreting and executing instructions from various programs using the processor. Programs that can be executed by the processor are stored in at least one of the memory area of ​​the processor and the storage unit 120.

[0167] The processing unit 130 may include a main memory. The main memory may include at least one of a volatile memory such as a RAM and a non-volatile memory such as a ROM (Read Only Memory). The main memory may also include at least one of the above-mentioned NOSRAM and DOSRAM.

[0168] The RAM may be, for example, a DRAM or an SRAM, and a virtual memory space is allocated and used as a working space for the processing unit 130. The operating system, application programs, program modules, program data, lookup tables, and the like stored in the storage unit 120 are loaded into the RAM for execution. The data, programs, and program modules loaded into the RAM are each directly accessed and operated by the processing unit 130.

[0169] The ROM can store a BIOS (Basic Input / Output System), firmware, etc., which do not require rewriting. Examples of ROM include mask ROM, OTPROM (One Time Programmable Read Only Memory), and EPROM (Erasable Programmable Read Only Memory). Examples of EPROMs include UV-EPROMs (Ultra-Violet Erasable Programmable Read Only Memories), which allow stored data to be erased by exposure to ultraviolet light, EEPROMs (Electrically Erasable Programmable Read Only Memories), and flash memories.

[0170] The processing section 130 can include one or both of an OS transistor and a transistor having silicon in a channel formation region (a Si transistor).

[0171] The processing unit 130 preferably includes an OS transistor. Because an OS transistor has an extremely small off-state current, using the OS transistor as a switch for retaining charge (data) flowing into a capacitor functioning as a memory element can ensure a long data retention period. By using this characteristic in at least one of the register and cache memory of the processing unit, the processing unit can be operated only when necessary, and can be turned off in other cases by saving information from the previous processing in the memory element. In other words, normally-off computing is possible, and the power consumption of the information processing system can be reduced.

[0172] It is preferable that the information processing device uses AI for at least some of its processing.

[0173] It is particularly preferable that the information processing device uses an artificial neural network (ANN, hereinafter simply referred to as a neural network). A neural network is realized by a circuit (hardware) or a program (software).

[0174] In this specification, a neural network refers to a general model that mimics the neural circuit network of a living organism, determines the connection strength between neurons through learning, and has problem-solving capabilities. A neural network has an input layer, an intermediate layer (hidden layer), and an output layer.

[0175] In this specification and the like, when discussing neural networks, determining the connection strengths (also called weighting coefficients) between neurons from existing information may be referred to as "learning."

[0176] In this specification and the like, the act of constructing a neural network using connection strengths obtained by learning and deriving a new conclusion from it may be referred to as "inference."

[0177] <<Output Unit 140>> The output unit 140 can output at least one of the calculation results, analysis results, and inference results of the processing unit 130 to the outside of the information processing device. For example, the output unit 140 can transmit data via the network 51. Specifically, a device such as a personal computer equipped with a communication port or a communication function can be used. Furthermore, a device equipped with a communication function may be used for the input unit 110 and the output unit 140.

[0178] <<Transmission Path 150>> The transmission path 150 has a function of transmitting data. Data can be transmitted and received between the input unit 110, the storage unit 120, the processing unit 130, and the output unit 140 via the transmission path 150. Specifically, an external bus, a LAN, or the Internet can be used as the transmission path 150.

[0179] Note that this embodiment mode can be appropriately combined with other embodiment modes described in this specification.

[0180] Embodiment 2 In this embodiment, an information processing method according to one embodiment of the present invention will be described with reference to FIGS.

[0181] FIG. 11 is a diagram illustrating an information processing method according to one embodiment of the present invention.

[0182] FIG. 12 is a diagram illustrating an information processing method according to one embodiment of the present invention.

[0183] <Example 1 of Information Processing Method> An information processing method according to one embodiment of the present invention includes steps S1 to S8 (see FIG. 11).

[0184] <<Step S1>> In step S1, the component 30 accepts a question Qre and transmits it to the component 21. The question Qre is written in a natural language. For example, a user of the information processing system inputs the question Qre.

[0185] <Step S2> In step S2, the component 21 receives the question Qre and shares it within the component 21. The component 21 includes a subcomponent 21A and a subcomponent 21B.

[0186] <Step S3> In step S3, the search engine SE of the subcomponent 21A collects records related to the question Qre from the database DB and creates background information BI.

[0187] The database DB includes a record ID_1, which includes original data OD_1 and a summary Sum_1.

[0188] The original data OD_1 is written in a programming language or a hardware description language, and the summary Sum_1 includes the gist of the original data OD_1 and is written in a natural language.

[0189] When a summary Sum_1 is relevant to the question Qre, the background information BI includes the summary Sum_1.

[0190] <<Step S4>> In step S4, the subcomponent 21B creates an instruction Pt-I and transmits it to the component 20. The instruction Pt-I includes a question Qre, background information BI, and a command g-I(). The command g-I() includes a command to generate an answer Ans to the question Qre by referring to the background information BI.

[0191] <Step S5> In step S5, the component 20 receives the instruction sentence Pt-I and generates an answer Ans using the large-scale language model LLM.

[0192] <Step S6> In step S6, the component 20 transmits the answer Ans to the component 21.

[0193] <Step S7> In step S7, the component 21 receives the answer Ans and transmits it to the component 30.

[0194] <Step S8> In step S8, the component 30 receives and provides the answer Ans.

[0195] As a result, even if the original data OD_1 is written in an artificial language other than a natural language, such as a programming language or a hardware description language, the main points can be described in the summary Sum_1 using natural language. Furthermore, the summary Sum_1 can be searched for using natural language. Furthermore, records containing summaries related to the question Qre can be collected from the database DB. Furthermore, summaries related to the question Qre can be collected from the database DB and used as background information BI. Furthermore, the large-scale language model LLM can be made to refer to the background information BI to generate an answer Ans to the question Qre. Furthermore, original data related to the question Qre can be provided. As a result, a novel information processing method that is highly convenient, useful, and reliable can be provided.

[0196] <Example 2 of Information Processing Method> An information processing method according to an aspect of the present invention includes a phase PH1 (see FIG. 12). Phase PH1 is a method of registering record ID_2, which includes original data OD_2 and summary Sum_2, in database DB.

[0197] <Example of Phase PH1> Phase PH1 includes steps S1 to S8.

[0198] <<Step S1>> In step S1 of phase PH1, the component 30 has a function of accepting a registration request (Reg) and transmitting it to the component 21. The registration request (Reg) includes original data OD_2, which is described in a programming language or a hardware description language. For example, a user of the information processing system inputs the registration request (Reg).

[0199] <Step S2> In step S2 of phase PH1, the component 21 accepts the registration request Reg and shares it within the component 21. The component 21 includes a subcomponent 21A and a subcomponent 21C.

[0200] <Step S3> In step S3 of phase PH1, the subcomponent 21C extracts the original data OD_2 from the registration request Reg and shares it within the component 21.

[0201] Step S4 In step S4 of phase PH1, the subcomponent 21C creates an instruction Pt-II and transmits it to the component 20. The instruction Pt-II includes the original data OD_2 and a command g-II(), and the command g-II() includes a command to generate a summary Sum_2 written in natural language from the original data OD_2.

[0202] <Step S5> In step S5 of phase PH1, the component 20 receives the directive Pt-II and generates a summary Sum_2 using the large-scale language model LLM.

[0203] Step S6 In step S6 of phase PH1, the component 20 transmits the summary Sum_2 to the component 21.

[0204] <Step S7> In step S7 of phase PH1, the component 21 receives the summary Sum_2 and shares it within the component 21.

[0205] <Step S8> In step S8 of phase PH1, the subcomponent 21A creates record ID_2 and registers it in the database DB. Note that record ID_2 includes original data OD_2 and summary Sum_2.

[0206] As a result, even if the original data OD_2 is written in an artificial language other than a natural language, such as a programming language or a hardware description language, the large-scale language model LLM can describe the main points of the original data in the summary Sum_2 using natural language. Furthermore, the original data OD_2 written in an artificial language other than a natural language can be associated with the summary Sum_2 written in a natural language and registered in the database DB. As a result, a novel information processing method that is highly convenient, useful, and reliable can be provided.

[0207] <Example 3 of Information Processing Method> An information processing method according to one aspect of the present invention includes phases PH1 and PH2 (see FIG. 12). Phase PH2 is a method in which, for example, a user of the information processing system checks and, if necessary, modifies summaries Sum_2 of records registered in the database DB.

[0208] <Example of Phase PH2> Phase PH2 follows the above-described phase PH1. Phase PH2 includes steps S1 to S6.

[0209] <Step S1> In step S1 of phase PH2, the component 21 transmits record ID_2 to the component 30. In particular, the component 21 transmits the summary Sum_2 included in record ID_2 to the component 30.

[0210] <Step S2> In step S2 of phase PH2, the component 30 provides record ID_2 and waits for input. For example, the component 30 outputs record ID_2 to the output unit and waits for input from a user of the information processing system to the input unit. Note that record ID_2 includes original data OD_2 and summary Sum_2. Furthermore, the original data OD_2 is written in an artificial language other than a natural language, and summary Sum_2 is written in a natural language. This allows, for example, a user of the information processing system to understand the provided summary Sum_2. Furthermore, if necessary, the component 30 can revise summary Sum_2 to create a revised document Sum_2Rev (see FIG. 8C ).

[0211] <Step S3> In step S3 of phase PH2, component 30 ends the process when approval is input, and proceeds to step S4 of phase PH2 when revised document Sum_2Rev is input (see FIG. 12). For example, a user of the information processing system inputs approval or revised document Sum_2Rev (see FIG. 5).

[0212] <Step S4> In step S4 of phase PH2, the component 30 replaces the summary Sum_2 of record ID_2 with the revised document Sum_2Rev to create record ID_2Rev, and sends it to the component 21 (see FIG. 12).

[0213] <Step S5> In step S5 of phase PH2, the component 21 receives the record ID — 2Rev and shares it within the component 21.

[0214] <Step S6> In step S6 of phase PH2, the subcomponent 21A replaces record ID_2 with record ID_2Rev.

[0215] This allows, for example, a user of the information processing system to check whether the summary Sum_2 generated by the large-scale language model LLM is valid. Furthermore, for example, a user of the information processing system can correct the summary Sum_2 generated by the large-scale language model LLM. As a result, a novel information processing method that is highly convenient, useful, and reliable can be provided.

[0216] Note that this embodiment mode can be appropriately combined with other embodiment modes described in this specification.

[0217] Ans: Answer, BI: Background Information, DB: Database, ID_1: Record, ID_2: Record, ID_2Rev: Record, ID_n: Record, LLM: Large Scale Language Model, OD_1: Original Data, OD_2: Original Data, OD_n: Original Data, Pt-I: Instruction, Pt-II: Instruction, Qre: Question, Reg: Registration Request, SE: Search Engine, Sum_1: Summary, Sum_2: Summary, Sum_2Rev: Revised Document, Sum_X: Summary, Sum_Y: Summary, 20: Component, 21: Component, 21A: Subcomponent, 21B: Subcomponent, 21C: Subcomponent, 30: Component, 51: Network, 110: Input Unit, 120: Storage Unit, 130: Processing Unit, 140: Output Unit, 150: Transmission Path

Claims

a first component; and a second component; and a third component, the first component has a function of receiving a question and sending it to the third component, and a function of receiving an answer and providing it; the second component has a function of receiving a first instruction sentence and transmitting the answer to the third component, and a function of performing processing using a large-scale language model; the large-scale language model has a function of generating the answer in accordance with the first instruction sentence; the third component has a function of receiving the question and sharing it within the third component, a function of creating the first instruction sentence and sending it to the second component, and a function of receiving the answer and sending it to the first component; the third component comprises a first subcomponent and a second subcomponent; the first subcomponent comprises a database and a search engine; the database comprises a first record; the first record includes first raw data and a first summary; the first original data is written in a programming language or a hardware description language; the first summary includes a gist of the first original data; the first summary is written in a natural language; The search engine has a function of collecting records related to the query from the database and a function of creating background information; when the first summary is relevant to the question, the background information includes the first summary; the second subcomponent has a function of creating the first instruction statement; the first directive includes the question, the background information, and a first command; The first instructions include instructions for generating the answer to the question by referring to the background information. 　 the first component has a function of receiving a registration request and transmitting it to the third component; the registration request includes second original data; the second original data is written in a programming language or a hardware description language; the second component has a function of receiving a second instruction statement and transmitting a second summary to the third component; the large-scale language model is operable to generate the second summary in accordance with the second instruction; the second summary includes a gist of the second original data; the third component has a function of receiving the registration request and sharing the registration request within the third component, a function of creating the second instruction sentence and sending it to the second component, and a function of receiving the second summary and sharing it within the third component; the third component comprises a third subcomponent; the third subcomponent has a function of extracting the second original data from the registration request and sharing the second original data within the third component, and a function of creating the second instruction statement; the second directive includes the second original data and a second command; the second instructions include instructions for generating the second summary written in a natural language from the second original data; the first subcomponent has a function of creating a second record and registering the second record in the database; The information processing system of claim 1 , wherein the second record includes the second raw data and the second summary.

3. The information processing system of claim 2, wherein the third subcomponent has a function of identifying a type of the second original data, and a function of selecting one instruction from a plurality of instructions and adopting it as the second instruction depending on the identified type. 　 the second component has a function of receiving a third instruction statement and transmitting a third summary to the third component, and a function of receiving a fourth instruction statement and transmitting a fourth summary to the third component; the large-scale language model is operable to generate the third summary in accordance with the third instruction sentence and to generate the fourth summary in accordance with the fourth instruction sentence; the third subcomponent comprises a pre-processing function; The pre-processing function includes: a function of extracting a first chunk from the beginning of the second original data; a function of extracting a second chunk following the first chunk from the second original data; a function of creating the third directive and the fourth directive; the third directive includes the first chunk and the second instruction; the fourth directive includes the second chunk and the second instruction; the third component has a function of transmitting the third instruction statement and the fourth instruction statement to the second component in a predetermined order, and a function of receiving the third summary and the fourth summary in a predetermined order and sharing them within the third component; 4. The information processing system according to claim 2, wherein the second directive includes the third summary, the fourth summary, and the second command instead of the second original data and the second command. 　 An information processing method having first to eighth steps, In a first step, a first component receives a query and sends it to a second component; the question is written in natural language; In a second step, the second component receives the question and shares it within the second component; the second component comprises a first subcomponent and a second subcomponent; In a third step, the search engine of the first subcomponent collects records related to the query from a database to generate background information; the database comprises records; the record includes raw data and a summary; the original data is written in a programming language or a hardware description language; the summary includes a gist of the original data; the abstract is written in natural language; when the abstract is relevant to the question, the background information includes the abstract; In a fourth step, the second subcomponent creates and sends an instruction to a third component; the instructions include the question, the background information, and an instruction; the instructions include instructions for generating an answer to the question with reference to the background information; In a fifth step, the third component receives the instruction sentence and generates the answer using a large-scale language model; In a sixth step, the third component sends the response to the second component; In a seventh step, the second component accepts the response and sends it to the first component; In an eighth step, the first component accepts and provides the answer.

1. An information processing method having a first phase, comprising: The first phase includes first to eighth steps, In the first step of the first phase, the first component has a function of accepting a registration request and transmitting the request to the second component; The registration request includes original data, the original data is written in a programming language or a hardware description language; In a second step of the first phase, the second component receives the registration request and shares it within the second component; the second component comprises a first subcomponent and a second subcomponent; In a third step of the first phase, the second subcomponent extracts the original data from the registration request and shares it within the second component; In a fourth step of the first phase, the second subcomponent creates and sends an instruction statement to a third component; the instruction statement includes the original data and a command; the instructions include instructions for generating a summary written in natural language from the original data; In a fifth step of the first phase, the third component receives the instruction sentence and generates the summary using a large-scale language model; In a sixth step of the first phase, the third component transmits the summary to the second component; In a seventh step of the first phase, the second component receives the summary and shares it within the second component; In an eighth step of the first phase, the first subcomponent creates a first record and registers the first record in a database; The information processing method, wherein the first record includes the original data and the summary. 　 An information processing method having the first phase and the second phase, the second phase follows the first phase; the second phase includes first to sixth steps, In a first step of the second phase, the second component transmits the first record to the first component; In a second step of the second phase, the first component provides the first record and waits for input; In the third step of the second phase, the first component terminates the process when an acceptance is entered, and advances the process to the fourth step of the second phase when a revised document is entered; In a fourth step of the second phase, the first component replaces the summary of the first record with the modified document to create a second record and sends it to the second component; In a fifth step of the second phase, the second component receives the second record and shares it within the second component; The information processing method according to claim 6 , wherein in a sixth step of the second phase, the first subcomponent replaces the first record with the second record.

Images