Information processing method, program, and information processing system

Abstract

An information processing method executed by an information processing device, the method including: a processing indication acquisition step for acquiring processing indication information that indicates processing; a use information acquisition step for acquiring use information needed for processing from an information source on the basis of the processing indication information; and a processing step for executing processing by using a trained model on the basis of the use information.

Description

Information processing method, program, and information processing system

[0001] The present disclosure relates to an information processing method, a program, and an information processing system.

[0002] Patent Literature 1 (JP 2023-69344 A) discloses an information processing device that improves the prediction accuracy of a machine learning model. In this information processing device, when training a reliability determination model for evaluating the reliability of a correct answer label used in training a machine learning model, additional information including the freshness of the information is collected and used to more accurately evaluate the reliability of the correct answer label.

[0003] It is known that the processing results output by a trained model change depending on the data (usage information) input to the trained model, but there is a need to control the usage information so that the desired processing results are obtained based on processing instruction information that instructs the processing.

[0004] The present disclosure aims to effectively obtain desired processing results.

[0005] According to one embodiment, the information processing method is executed by an information processing device, and includes a processing instruction acquisition step of acquiring processing instruction information that instructs processing, a usage information acquisition step of acquiring usage information required for processing from an information source based on the processing instruction information, and a processing step of executing processing using a trained model based on the usage information.

[0006] According to one embodiment of the program, an information processing device is caused to execute an information processing method including a processing instruction acquisition step of acquiring processing instruction information that instructs processing, a usage information acquisition step of acquiring usage information required for processing from an information source based on the processing instruction information, and a processing step of executing processing using a trained model based on the usage information.

[0007] According to one embodiment of the information processing system, a processing instruction acquisition step is performed to acquire processing instruction information that instructs processing, a usage information acquisition step is performed to acquire usage information required for processing from an information source based on the processing instruction information, and a processing step is performed to execute processing using a trained model based on the usage information.

[0008] According to one embodiment, desired processing results can be obtained effectively.

[0009] FIG. 1 is a diagram showing an example of the configuration of an information processing system according to a first embodiment. FIG. 2 is a diagram showing an example of the hardware configuration of a server according to the first embodiment. FIG. 3 is a diagram showing an example of the functional configuration of a server according to the first embodiment. FIG. 4 is a diagram showing an example of the processing flow of an information processing system according to the first embodiment. FIG. 5 is a diagram showing a pattern of index information of usage information in an information processing system according to the first embodiment. FIG. 6 is a diagram showing a pattern of index information of usage information in an information processing system according to the first embodiment. FIG. 7 is a diagram showing a pattern of index information of usage information in an information processing system according to the first embodiment. FIG. 8 is a diagram showing a pattern of index information of usage information in an information processing system according to the first embodiment. FIG. 9 is a diagram showing an example of the functional configuration of a server according to a second embodiment. FIG. 10 is a diagram showing an example of the processing flow of an information processing system according to the second embodiment.

[0010] (First embodiment) Hereinafter, a first embodiment of an information processing system according to the present disclosure will be described using Figures 1, 2, 3, 4, and 5A, 5B, 5C, 5D, and 5E. Note that the same reference numerals are used in each figure to designate identical or equivalent components and parts. Furthermore, the dimensional proportions in the drawings are exaggerated for the sake of explanation and may differ from the actual proportions.

[0011] (System Overview) First, an overview of the information processing system 10 according to this embodiment will be described. The information processing system 10 according to this embodiment is a system in which, when a prompt serving as processing instruction information instructing processing is input to a trained model such as a large-scale language model, the trained model generates a processing result corresponding to the prompt and outputs the processing result. That is, when a prompt is input by a user U, the information processing system 10 acquires usage information required for processing from a predetermined information source based on the prompt, performs processing using the trained model, and generates a processing result. Details will be described later.

[0012] (System Configuration) Fig. 1 is a diagram showing an example of the configuration of an information processing system 10 according to this embodiment. As shown in Fig. 1, the information processing system 10 according to this embodiment includes a server 12 as an information processing device and a user terminal 14, which are communicably connected to each other via a network N. The network N is, for example, a wired LAN (Local Area Network), a wireless LAN, the Internet, a public line network, a mobile data communication network, or a combination thereof.

[0013] The user terminal 14 is an example of an information processing device that is operated by a user U to input and display various information. The user terminal 14 may be a personal computer (PC), a smartphone, a tablet terminal, a server device, a microcomputer, a wearable device, or a combination thereof.

[0014] The server 12 is an example of an information processing device that acquires information input from the user terminal 14, processes the information, and outputs the results. The server 12 may be a PC (Personal Computer), a smartphone, a tablet terminal, a server device, a microcomputer, or a combination of these. The specific configuration and operation of the server 12 will be described later.

[0015] 2 is a block diagram showing the hardware configuration of the server 12. The server 12 includes a processor 120, a memory 122, a storage 124, a communication I / F 126, an input / output I / F 128, and a drive device 134, which are communicatively connected to each other via a bus B.

[0016] The processor 120 controls each component of the server 12 and realizes the functions of the server 12 by loading various programs stored in the storage 124 into the memory 122 and executing them. The programs executed by the processor 120 include, but are not limited to, an operating system (OS) and a program 220 described below. The execution of these programs by the processor 120 realizes part of the state visualization method according to this embodiment. The processor 120 is, for example, a central processing unit (CPU), a micro processing unit (MPU), a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a digital signal processor (DSP), or a combination thereof.

[0017] The memory 122 is, for example, a read-only memory (ROM), a random access memory (RAM), or a combination thereof. The ROM is, for example, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a combination thereof. The RAM is, for example, a dynamic random access memory (DRAM), a static random access memory (SRAM), a magnetoresistive random access memory (MRAM), or a combination thereof.

[0018] The storage 124 stores the OS, various programs described below, and various data. The storage 124 is, for example, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), a storage class memory (SCM), or a combination thereof.

[0019] The communication I / F 126 is an interface for connecting the server 12 to external devices, including the user terminal 14 and the image capture device 16, via the network N and for controlling communication. The communication I / F 126 is, for example, an adapter compliant with Bluetooth (registered trademark), Wi-Fi (registered trademark), ZigBee (registered trademark), Ethernet (registered trademark), or optical communication (e.g., Fibre Channel), but is not limited to these.

[0020] The input / output I / F 128 is an interface for connecting an input device 132 and an output device 130 to the server 12. The input device 132 is, for example, a mouse, a keyboard, a touch panel, a microphone, a scanner, a camera, various sensors, an operation button, or a combination thereof. The output device 130 is, for example, a display, a projector, a printer, a speaker, a vibrator, or a combination thereof.

[0021] The drive device 134 reads and writes data from and to the disk media 136. The drive device 134 is, for example, a magnetic disk drive, an optical disk drive, a magneto-optical disk drive, or a combination thereof. The disk media 109 is, for example, a compact disc (CD), a digital versatile disc (DVD), a floppy disc (FD), a magneto-optical disc (MO), a Blu-ray (registered trademark) disc (BD), or a combination thereof.

[0022] In this embodiment, the program may be written to memory 122 or storage 124 during the manufacturing stage of server 12, may be provided to server 12 via network N, or may be provided to server 12 via a non-transitory computer-readable recording medium such as disk media 136.

[0023] Furthermore, the hardware configuration of the user terminal 14 is substantially the same as the hardware configuration of the server 12 described above, and therefore a detailed description thereof will be omitted.

[0024] (Functional Configuration) Next, the functional configuration of the server 12 will be described. Fig. 3 is a diagram showing an example of the functional configuration of the server 12. When executing various programs, the server 12 uses the above-mentioned hardware resources to realize various functions. The server 12 has a communication unit 20, a storage unit 22, and a control unit 24 as the functional configuration realized by the server 12. Each functional configuration is realized when the processor 120 reads and executes a program 220 stored in the memory 122 or the storage 124.

[0025] The communication unit 20 is realized by the communication I / F 126. The communication unit 20 transmits and receives information to and from the user terminal 14 via the network N. The communication unit 20 receives information input from the user terminal 14. The communication unit 20 also transmits information to the user terminal 14 and receives requests from the user U from the user terminal 14.

[0026] The storage unit 22 is realized by the memory 122 and the storage 124. The storage unit 22 stores a program 220, a trained model 222, a prompt 224, usage information 226, and a processing result 228.

[0027] The trained model 222 is composed of at least one trained machine learning model. As an example, the trained model 222 is a large-scale language model trained on a large amount of text data from internet articles, books, websites, etc. When text data is input as a prompt, the trained model 222 generates and outputs text data corresponding to the prompt. The large-scale language model in the trained model 222 of this embodiment determines the direction of an output result based on the content of the prompt. The output result refers to an expected output result.

[0028] Here, the directionality indicates the nature or characteristics of the output result (hereinafter simply referred to as "nature"). The trained model 222 determines the following two properties, for example. The first property is a "strict, source-based" property (hereinafter simply referred to as "strict"), and output results according to this property are based on clear information extracted or derived from specific sources. On the other hand, the second property is a "broad and creative" property (hereinafter simply referred to as "creativity"), and output results according to this property are not limited to a specific source but draw insights from a wide range of information and provide new perspectives and interpretations. The trained model 222 determines the degree of "strictness" and "creativity," which are opposing properties of the output result, by determining the directionality. That is, if the prompt instructs "Please answer based on the contents of the previous management meeting," the trained model 222 determines that an output result with a strict nature based on a source such as the minutes of the previous management meeting is required, and determines the directionality of the output result to be "strict." On the other hand, if the prompt instructs the user to "take recent economic and technological trends into consideration when answering," the trained model 222 determines that an output result with a creative characteristic not limited to a specific information source is required, and determines the direction of the output result as "creativity." The above-described determination of "strictness" and "creativity" is merely an example, and it goes without saying that the direction can be determined appropriately for other prompts based on their content. The direction may also be determined based on other characteristics. Furthermore, rather than simply determining the presence or absence of each characteristic, the degree to which the characteristic is required may be determined numerically, such as "strictness 80, creativity 20." Furthermore, the output result may be divided into multiple parts, and the characteristic may be determined for each part. A specific example would be to divide the output result into a first part providing a fact-based answer and a second part providing consideration of the first part, and to determine the first part as "strictness" and the second part as "creativity."

[0029] Furthermore, the trained model 222 determines, based on the prompt, candidates for usage information required for processing. Specifically, the trained model 222 evaluates the characteristics of information required for processing based on the prompt (e.g., whether the information is based on common sense or specialized knowledge), and when it determines that the characteristics require acquisition from an external information source, the trained model 222 searches the information source and determines the information required for processing as a candidate for usage information. Note that the information source may be the memory unit 22 in which various types of information are stored, or may be an external information source such as a book or information on the Internet. Note that the usage information is various types of information including text information stored in these information sources, and will be described in detail below.

[0030] Furthermore, the trained model 222 generates at least a portion of index information for the candidate usage information. At least a portion of this index information is, for example, information freshness information, and this freshness information changes depending on the content of the usage information and the processing content. In other words, freshness information is the "value of information over time," and can be broadly divided into a pattern in which value decreases over time (hereinafter simply referred to as a "decreasing pattern"), a pattern in which value remains constant over time (hereinafter simply referred to as a "constant pattern"), and a pattern in which value increases over time (hereinafter simply referred to as an "increasing pattern"). Each pattern will be explained below. The "value of information" here refers to how much the information contributes to the processing result desired by user U, and when the value of information is high, it contributes greatly to outputting the desired processing result.

[0031] Decreasing patterns include a pattern in which value decreases at a substantially constant rate (see FIG. 5A ) and a pattern in which value decreases rapidly (see FIG. 5B ). Specific examples of information in which value decreases at a substantially constant rate include general news information, product information for which new models are periodically released, and information on the results of events, etc. These types of information are time-limited information whose value gradually decreases over time. Specific examples of information in which value decreases rapidly include stock market information, weather forecasts, and news information related to events with rapidly changing conditions, such as accidents and disasters.

[0032] Information with a fixed pattern (see FIG. 5C) specifically corresponds to information about historical facts, scientific principles, etc. This information has a relatively small change in value over time.

[0033] The rising pattern includes a pattern in which the value rises at a substantially constant rate (see FIG. 5D ) and a pattern in which the value rises rapidly (see FIG. 5E ). Information with a pattern in which the value rises at a substantially constant rate is, specifically, information such as historical data, classical literature, products produced in limited quantities, or old books, which may not initially be recognized as having much value but whose importance and value are rediscovered over time, causing its value to rise. Information with a pattern in which the value rises rapidly is, specifically, information about past artworks or collectibles, information about predictive analyses or predictions, information about specific technologies or theories, and information about vintage products, which initially did not attract much attention but whose value rises rapidly due to events such as the death of the creator or the subsequent success of past predictions. Note that the specific information for each pattern is not limited to those described above.

[0034] The three patterns described above are merely examples, and other patterns may be included, such as a pattern in which the value fluctuates periodically, such as information about periodic events. Furthermore, patterns that combine the above patterns may be included. That is, examples include information whose value increases but remains constant or decreases after a predetermined timing, information whose value was constant but increases or decreases after a predetermined timing, and information whose value decreases but remains constant or increases after a predetermined timing. In other words, the trained model 222 can also be expressed as "evaluating usage information whose value changes depending on the acquired prompt using an index called freshness information."

[0035] The prompt 224 is information indicating a prompt input by the user U. The prompt 224 is written in natural language and is stored in association with the user U's account information, for example.

[0036] The usage information 226 is required for processing. As described above, this usage information 226 is information acquired from information sources such as books, the Internet, and the storage unit 22, and includes text information, image information, and the like. The usage information 226 includes unstructured data and structured data. Here, "unstructured data" refers to data that is stored without being structured in a predefined manner. On the other hand, "structured data" refers to data that exists in a specific field in a file or record. "Structured data" also includes "semi-structured data." The usage information 226 can be stored in a data lake (not shown) in the storage unit 22, and can be acquired from the data lake using the index information of the usage information 226 (described later).

[0037] The processing result 228 is information indicating the result of executing a process using the usage information based on the prompt. The processing result 228 is various information including text information and image information, and is stored in association with the prompt and the usage information.

[0038] The control unit 24 is realized by the processor 120 reading and executing the program 220 from the memory 122 (see FIG. 2 ) and working in cooperation with other hardware components. The control unit 24 includes an information acquisition unit 242, an information processing unit 244, and an output unit 246.

[0039] The information acquisition unit 242 acquires various information required for processing by the information processing unit 244. Specifically, it acquires a prompt 224 that instructs processing and usage information 226 that includes unstructured data required for processing. The information acquisition unit 242 also acquires index information for the usage information 226 (details will be described later). This acquired information may be information stored in the storage unit 22, or may be information acquired from an external data source or the like.

[0040] The information processing unit 244 determines the directionality of the output result based on the content of the prompt, acquires and generates index information of the usage information 226, determines the degree of match between the directionality of the output result and the usage information, and executes processing based on the prompt. The directionality of the output result is determined by using the trained model 222 to determine the properties of the output result expected in response to the prompt input by the user U, as described above.

[0041] The information processing unit 244 generates freshness information from the index information of the usage information 226 using the trained model 222 as described above. The information processing unit 244 is also able to use metadata and log data as index information of the usage information acquired by the information acquisition unit 242. Metadata is information that describes the characteristics and attributes of the data, such as the content, source, creator, and timestamp of the usage information 226. Log data is information recorded along with the time generated each time a certain operation is performed, such as the date and time when the usage information 226 was recorded or accessed, the executor who performed the recording or access, request and response information, and error information. These metadata and log data are linked to the usage information 226 in advance and are available for use. The information processing unit 244 uses at least one of the index information, freshness information, metadata, and log data to determine whether the usage information matches the directionality of the output result (details will be described later).

[0042] The information processing unit 244 determines the degree of match between the directionality of the output result and the usage information 226. That is, the information processing unit 244 determines the degree of match between the directionality and the usage information 226, based on the index information of the usage information 226, and whether the candidate in the usage information 226 matches the processing based on the prompt. As a specific example, if the directionality is determined to be "strict," the information processing unit 244 determines the degree of match between the usage information 226 and the candidate in the usage information 226 from the perspectives of whether the candidate in the usage information 226 is accurate, complete, up-to-date, reliable, clear, traceable, consistent, etc.

[0043] In other words, accuracy is determined based on criteria such as, for example, whether the update date and time of the usage information 226 in question is recent and whether that date and time has been recorded accurately from the metadata, whether the registrant is a highly trustworthy person or institution, whether the metadata conforms to a predefined format or standard, whether error messages or abnormal patterns are searched for in the log data to see if there are any signs of a system abnormality, and whether the freshness information shows an upward or downward pattern and the factors that led to that pattern are tracked to see if there are any inconsistencies with the content.

[0044] Completeness is determined based on criteria such as whether all necessary metadata elements are present, whether all log data that should be collected is present, and whether the freshness information contains important elements or bits that allow for a complete picture of fluctuations in information value, such as the specific time periods when an increase or decrease in value occurred and the circumstances under which this occurred.

[0045] Whether information is up to date or not is determined based on criteria such as whether it reflects the latest information from the metadata, whether the log data is real-time or the latest data, whether it reflects the latest fluctuations in information value from the freshness information, and whether the information value is above a predetermined value.

[0046] Whether a data is highly reliable or not is determined based on criteria such as whether the metadata indicates the reliability of the data source and the evaluation results of the metadata generation process, whether the log data is collected using reliable means and there is no risk of tampering, and whether the freshness information shows little fluctuation in the information value in any pattern.

[0047] Clarity is determined based on criteria such as whether the metadata is clearly explained and defined according to conventions, whether the log data is recorded in a consistent format and can be clearly interpreted, and whether the freshness information clearly indicates whether the information value is declining, remaining constant, or increasing.

[0048] Whether or not something is traceable is determined based on criteria such as whether the source of the data is clearly indicated from the metadata and whether it is possible to query that source, whether each log entry can be traced over time from the log data (whether time information is included), and whether changes in the value of each piece of information can be traced over time from the freshness information, i.e., whether the time at which the change occurred is clearly indicated.

[0049] Consistency is determined based on criteria such as whether there is consistency between metadata and whether they comply with defined rules and standards, whether the log data is generated in a consistent format and pattern, and whether the freshness information shows consistent fluctuations in information value. Note that the above-mentioned criteria are merely examples, and other criteria may be used for determination.

[0050] If the degree of match between the directionality of the output result and the usage information 226 is equal to or greater than a predetermined value, the information processing unit 244 performs processing corresponding to the prompt based on the usage information by using the trained model 222. On the other hand, if the degree of match between the directionality of the output result and the usage information 226 is not equal to or greater than a predetermined value, the information processing unit 244 searches for the usage information 226 again and acquires, as a search result, candidates for the usage information 226 via the information acquisition unit 242.

[0051] The output unit 246 controls the information processing unit 244 so that the processing results are output to the user terminal 14 .

[0052] (Processing Executed by Information Processing System 10) Next, the operation of the information processing system 10 will be described. Fig. 4 is a flowchart showing an example of the flow of processing by the information processing system 10. The processor 120 reads out the program 220 stored in the storage 124, expands it in the memory 122, and executes it, thereby performing processing. Although not shown, when the processor 120 receives operation information to terminate the operation of the information processing system 10, or operation termination information from the user terminal 14 during the ongoing determination processing (these will be simply referred to as "termination operations"), the processor 120 terminates the processing based on the program 220 that is currently being processed.

[0053] The processor 120 determines whether a prompt input from the user U or the like has been acquired (step S100). If a prompt has not been input or cannot be acquired because the input is in progress (step S100: NO), the processor 120 repeats the processing of step S100. On the other hand, if a prompt has been acquired (step S100: YES), the processor 120 determines the direction of the result to be output based on the content of the prompt (step S102). Note that the above-mentioned step S100 corresponds to the "processing instruction acquisition step" recited in claim 1.

[0054] The processor 120 determines candidates for usage information required for processing based on the prompt and acquires index information for the candidate usage information (step S104). The processor 120 then determines whether or not there is usage information from the acquired index information that matches the directionality determined in step S102 (step S106). If there is no usage information with matching directionality (step S106: NO), the processor 120 excludes the usage information and re-determines candidates for usage information required for processing based on the prompt (step S107), and proceeds to step S104. On the other hand, if there is usage information with matching directionality (step S106: YES), the processor 120 acquires the usage information (step S108) and performs information processing using the usage information based on the prompt (step S110). The processor 120 then outputs the processing results (step S112) and terminates the processing. This output includes displaying the results on the display of the user terminal 14 and executing processing on external tools, sensors, data sources, etc. The above-mentioned step S108 corresponds to the "usage information acquisition step" and step S110 corresponds to the "processing step" in claim 1.

[0055] (Effects of First Embodiment) The information processing system 10 according to this embodiment executes a processing instruction acquisition step of acquiring a prompt (processing instruction information) that instructs processing, a usage information acquisition step of acquiring usage information 226 required for processing from an information source based on the processing instruction information, and a processing step of executing processing corresponding to the prompt 224 using the trained model 222 based on the usage information 226. As a result, processing is executed using the usage information 226 based on the prompt 224, and therefore a desired processing result can be obtained effectively.

[0056] Furthermore, in the usage information acquisition step, the information processing system 10 acquires the usage information 226 based on the output directionality determined from the prompt 224, and therefore can execute processing corresponding to the prompt 224 based on the appropriate usage information 226. This makes it possible to more effectively obtain desired processing results.

[0057] Furthermore, in the usage information acquisition step, the information processing system 10 uses the metadata of the usage information 226 when acquiring the usage information 226, and therefore can efficiently and appropriately select the usage information 226. This makes it possible to more effectively obtain the desired processing result.

[0058] Furthermore, in the usage information acquisition step, the information processing system 10 uses log data related to the usage information 226 when acquiring the usage information 226, so that the usage information 226 can be selected based on objective records, thereby enabling the desired processing results to be obtained more effectively.

[0059] Furthermore, in the usage information acquisition step, the information processing system 10 inputs information about the information source to the trained model 222 that has been trained to output index information for the input information, and uses the index information output from the trained model 222 when acquiring the usage information 226. This makes it possible to appropriately select the usage information 226 even for data that does not have metadata or log data. This makes it possible to obtain desired processing results more effectively.

[0060] Second Embodiment Next, an information processing system 40 according to a second embodiment of the present disclosure will be described with reference to Figures 6 and 7. The information processing system 40 according to the second embodiment has the same basic configuration as the first embodiment, and is characterized in that it executes a storage step of storing a processing result, and outputting the stored processing result when the relevance between the stored processing result and a new prompt 224 is equal to or greater than a predetermined standard. Note that the same components as those in the first embodiment are denoted by the same reference numerals, and their description will be omitted.

[0061] (Functional Configuration) The functional configuration of the server 50 as an information processing device in the information processing system 40 will be described. FIG. 6 is a diagram showing an example of the functional configuration of the server 50. When executing various programs, the server 50 realizes various functions using the same hardware resources as the server 12 of the first embodiment. The server 50 has a communication unit 20, a storage unit 60, and a control unit 70 as the functional configuration realized by the server 50. Each functional configuration is realized when the processor 120 reads and executes a program 600 stored in the memory 122 or the storage 124.

[0062] The control unit 70 is realized by the processor 120 reading and executing the program 600 from the memory 122 (see FIG. 2 ) and working in cooperation with other hardware components. The control unit 70 includes an information acquisition unit 242, an information processing unit 700, and an output unit 246.

[0063] Similar to the information processing unit 244 of the first embodiment, the information processing unit 700 determines the directionality of the output result based on the content of the prompt, acquires and generates index information for the usage information 226, determines the degree of match between the directionality of the output result and the usage information, and executes processing based on the prompt. In addition, the information processing unit 700 associates the processing result corresponding to the prompt 224 with the usage information 226, if available, and stores it in the storage unit 60.

[0064] Furthermore, when the information processing unit 244 acquires a new prompt 224, it outputs the processing result if the relevance between the prompt 224 and the processing result stored in the storage unit 60 is equal to or greater than a predetermined standard. For example, the predetermined standard is a standard for determining whether the new prompt 224 is similar to the prompt 224 linked to the processing result stored in the storage unit 60. Specifically, the predetermined standard is set to a value such as cosine similarity or Euclidean distance by vector transforming the new prompt 224 and the prompt 224 linked to the processing result stored in the storage unit 60. Here, a value equal to or greater than the predetermined standard indicates a higher degree of similarity. The predetermined standard may be other standards, such as word-based similarity or semantic similarity, in addition to the above-described standard. Furthermore, the predetermined standard may be a similarity determination based not only on a comparison between the new prompt 224 and the prompt 224 linked to the processing result stored in the storage unit 60 but also on a comparison between the new prompt 224 and the processing result itself stored in the storage unit 60. Furthermore, the predetermined criteria may be not only those that are judged to be similar, but also those that are judged to be connected, have something in common, or have an influencing relationship, or may be something else.

[0065] (Processing Executed by Information Processing System 40) Next, the operation of the information processing system 40 will be described. Fig. 7 is a flowchart showing an example of the flow of processing by the information processing system 40. The processor 120 reads out the program 600 stored in the storage 124, expands it in the memory 122, and executes it, thereby performing processing. Note that the same processes as those in the first embodiment are denoted by the same reference numerals, and descriptions thereof will be omitted.

[0066] After processing step S100, processor 120 determines the relevance between the acquired prompt 224 and the prompt 224 linked to the processing result stored in storage unit 60 (step S200), and determines whether the determined relevance meets or exceeds a predetermined standard (step S202). If the relevance does not meet or exceed the predetermined standard (step S202: NO), processor 120 proceeds to step S102. On the other hand, if the relevance meets or exceeds the predetermined standard (step S202: YES), processor 120 acquires the processing result corresponding to the prompt 224 linked to the processing result stored in storage unit 60 (step S204), and proceeds to step S112.

[0067] After the process of step S110, the processor 120 associates the process result with the corresponding prompt 224 and, if any, the usage information 226, with the usage information 226 in the storage unit 60 (step S206), and then proceeds to step S112. The process of step S206 corresponds to the "storing step" of claim 6.

[0068] (Effects of the Second Embodiment) The information processing system 40 according to this embodiment has the same configuration as the first embodiment, except for the execution of a storage step for storing a processing result and an output of the stored processing result when the relevance between the stored processing result and a new prompt 224 is equal to or greater than a predetermined standard. Therefore, the same effects as those of the first embodiment can be obtained. Furthermore, the information processing system 40 includes a storage step for storing the processing results of the processing steps in the memory unit 60 of the server 50, and outputs the processing result when the relevance between a new prompt 224 and the processing result stored in the memory unit 60 is equal to or greater than a predetermined standard when the new prompt 224 is acquired in the processing instruction acquisition step. Therefore, when a new prompt 224 that is relevance to a previously processed prompt 224 is input, the execution of processing using the trained model 222 can be avoided. This makes it possible to stabilize the processing results and reduce processing costs due to the use of the trained model 222.

[0069] In the first and second embodiments described above, at least one of metadata, log data, and freshness information is used as the index information for the usage information 226. However, other data, such as annotation data, may also be used as the index information. Furthermore, the trained model 222 is configured to generate freshness information for the usage information 226 from the content of the usage information 226. However, the freshness information may be generated using not only the content of the usage information 226 but also information such as metadata and log data. Furthermore, the trained model 222 may generate not only freshness information as index information, but also information for determining information necessary for processing, such as a summary of the content of the usage information 226. The information processing systems 10 and 40 may be configured to determine this information using at least one of the rule base and the trained model 222 and acquire the usage information 226 necessary for processing from the information source.

[0070] Furthermore, while the freshness information of the usage information 226 is assumed to be generated by the trained model 222 that performs prompt-based processing, this is not limiting and the index information including the freshness information may be generated by a trained model other than the trained model 222 that performs prompt-based processing. In this configuration, a trained model optimized for generating index information can be used, thereby enabling more accurate acquisition of the usage information 226. The trained model that outputs the freshness information is, for example, a model that has undergone machine learning using a dataset that includes at least one of time-stamped information, origin information, update frequency information, user response information, and topic / field property information. Here, the time-stamped information is information to which a timestamp indicating when the target information was generated is attached, allowing the freshness of the information to be directly evaluated. The origin information is information indicating where the target information originates. For example, newspaper articles are most fresh on the day of publication and their freshness decreases over time, whereas historical data and cultural data have increasing information value over time, allowing the freshness of information to be evaluated. Update frequency information indicates how frequently the target information is updated, and allows for evaluation of the freshness of information, for example, that information that is updated frequently has high value. User reaction information is information that indicates how users reacted to the target information, such as the number of clicks, display time, number of shares, and comments, and allows for evaluation of the freshness of information, for example, that information with a high number of reactions has high information value. Topic / field characteristic information allows for evaluation of the freshness of information taking into account the background of each field, for example, that in the field of science and technology, information is updated every time new research or publication is made, but in fields such as literature and history, information is not updated as frequently.

[0071] Furthermore, in the processing performed by processor 120 in step S106 in the first and second embodiments described above, the presence or absence of usage information that matches the directionality determined in step S102 from the acquired index information is determined. However, the present invention is not limited to this, and the configuration may be such that the degree of match with the directionality in the acquired index information is calculated, and usage information corresponding to the index information in descending order of degree of match is obtained and used for processing, or the determination based on this degree of match may be combined with the determination of the presence or absence of the usage information described above.

[0072] (Variation 1) Furthermore, when the above-described information processing system is viewed from a different perspective, the problem (purpose) that the information processing system according to this embodiment is trying to solve can also be viewed as "reducing processing costs."

[0073] If the problem is understood as described above, an invention as a means for solving the problem can be, for example, as follows: "An information processing method executed by an information processing device, comprising: a processing instruction acquisition step of acquiring processing instruction information that instructs processing; a processing step of executing processing corresponding to the processing instruction information by utilizing a trained model based on the processing instruction information; and a storage step of storing a processing result from the processing step in a storage unit of the information processing device, wherein when new processing instruction information is acquired in the processing instruction acquisition step, the processing result is output if the relevance between the processing instruction information and the processing result stored in the storage unit is equal to or greater than a predetermined standard."

[0074] According to the above configuration, the information processing device includes a storage step for storing the processing results of the processing step in a memory unit of the information processing device, and when new processing instruction information (i.e., a prompt) is acquired in the processing instruction acquisition step, the processing result is output if the correlation between the new processing instruction information and the processing result stored in the memory unit is equal to or greater than a predetermined standard. Therefore, when new processing instruction information that is correlated with processing instruction information processed in the past is input, it is possible to avoid executing processing using a trained model. This makes it possible to stabilize the processing results and reduce processing costs due to the use of a trained model.

[0075] <Additional Notes> This embodiment includes the following disclosure.

[0076] (Supplementary Note 1) An information processing method executed by an information processing device, comprising: a processing instruction acquisition step of acquiring processing instruction information that instructs processing; a usage information acquisition step of acquiring usage information required for processing from an information source based on the processing instruction information; and a processing step of executing processing corresponding to the processing instruction information using a trained model based on the usage information.

[0077] (Supplementary Note 2) The information processing method according to Supplementary Note 1, wherein in the usage information obtaining step, the usage information is obtained based on an output directionality determined from the processing instruction information.

[0078] (Supplementary Note 3) The information processing method according to Supplementary Note 1, wherein in the usage information acquisition step, metadata of the usage information is used when acquiring the usage information.

[0079] (Supplementary Note 4) The information processing method according to Supplementary Note 1, wherein in the usage information acquisition step, log data related to the usage information is used when the usage information is acquired.

[0080] (Supplementary Note 5) In the information processing method according to Supplementary Note 1, in the usage information acquisition step, information about the information source is input to a trained model that has been trained to output index information for the information when the information is input, and the index information output from the trained model is used when acquiring the usage information.

[0081] (Supplementary Note 6) An information processing method as described in Supplementary Note 1, comprising a storage step of storing the processing results of the processing step in a memory unit of the information processing device, and when new processing instruction information is acquired in the processing instruction acquisition step, outputting the processing results if the correlation between the processing instruction information and the processing results stored in the memory unit is equal to or greater than a predetermined standard.

[0082] (Supplementary Note 7) A program for causing an information processing device to execute an information processing method including: a processing instruction acquisition step of acquiring processing instruction information that instructs processing; a usage information acquisition step of acquiring usage information required for processing from an information source based on the processing instruction information; and a processing step of executing processing using a trained model based on the usage information.

[0083] (Supplementary Note 8) An information processing system executed by an information processing device, comprising: a processing instruction acquisition step of acquiring processing instruction information that instructs processing; a usage information acquisition step of acquiring usage information required for processing from an information source based on the processing instruction information; and a processing step of executing processing using a trained model based on the usage information.

[0084] The embodiments disclosed herein are illustrative in all respects and should not be considered limiting. The scope of the present disclosure is defined by the claims, not by the meaning described above, and is intended to include all modifications within the meaning and scope of the claims. Furthermore, the present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present disclosure.

[0085] The disclosure of Japanese Patent Application No. 2023-211291, filed on December 14, 2023, is incorporated herein by reference in its entirety. All documents, patent applications, and technical standards described herein are incorporated herein by reference to the same extent as if each individual document, patent application, and technical standard was specifically and individually indicated to be incorporated by reference.

Claims

1. An information processing method executed by an information processing device, comprising: a processing instruction acquisition step of acquiring processing instruction information that instructs processing; a usage information acquisition step of acquiring usage information required for processing from an information source based on the processing instruction information; and a processing step of executing processing corresponding to the processing instruction information based on the usage information using a trained model that learns text data and outputs a processing result corresponding to the processing instruction information when the processing instruction information is input, wherein in the usage information acquisition step, the usage information is acquired based on an output directionality determined from the processing instruction information.

2. The information processing method according to claim 1, wherein in said usage information acquisition step, metadata of said usage information is used when acquiring said usage information.

3. The information processing method according to claim 1, wherein in said usage information acquisition step, log data related to said usage information is utilized when said usage information is acquired.

4. The information processing method of claim 1, wherein, in the usage information acquisition step, when information is input, index information for determining whether or not the information corresponds to the usage information is obtained, and the information of the information source is input to the trained model or a trained model other than the trained model that has been trained to output index information for the information when information is input, and the output index information is used when acquiring the usage information.

5. An information processing method as described in claim 1, further comprising a storage step of storing the processing result in the processing step and the processing instruction information linked to the processing result in a memory unit of the information processing device, and when new processing instruction information is acquired in the processing instruction acquisition step, outputting the processing result if the correlation between the processing instruction information and the processing instruction information linked to the processing result stored in the memory unit is equal to or greater than a predetermined standard.

6. A program for causing an information processing device to execute an information processing method, comprising: a processing instruction acquisition step for acquiring processing instruction information that instructs processing; a usage information acquisition step for acquiring usage information required for processing from an information source based on the processing instruction information; and a processing step for executing processing based on the usage information using a trained model that learns text data and outputs a processing result corresponding to the processing instruction information when the processing instruction information is input, wherein in the usage information acquisition step, the usage information is acquired based on an output directionality determined from the processing instruction information.

7. An information processing system executed by an information processing device, comprising: a processing instruction acquisition step of acquiring processing instruction information that instructs processing; a usage information acquisition step of acquiring usage information required for processing from an information source based on the processing instruction information; and a processing step of executing processing based on the usage information using a trained model that learns text data and outputs a processing result according to the processing instruction information when the processing instruction information is input, wherein in the usage information acquisition step, the usage information is acquired based on an output directionality determined from the processing instruction information.

Images