Information processing system

Abstract

An information processing system disclosed herein includes one or more memories and one or more processors. The one or more processors acquire a data file to be analyzed, extract a plurality of pieces of data included in the data file, analyze each of the plurality of pieces of data using a generation model, and display a graph summarizing the analysis results.

Description

Information Processing System

[0001] The present disclosure relates to an information processing system.

[0002] When performing data analysis, the user designs what data analysis to perform according to the purpose of the data analysis.

[0003] Japanese Patent No. 7527581

[0004] According to the present disclosure, a technique for easily performing data analysis can be provided.

[0005] An information processing system according to an aspect of the present disclosure includes one or more memories and one or more processors. The one or more processors acquire a data file to be analyzed, extract a plurality of data included in the data file, analyze each of the plurality of data using a generation model, and display a graph obtained by aggregating the analysis results.

[0006] It is a diagram showing the flow of processing in an information processing system. It is a diagram showing an example of a page displayed on a web browser. It is a diagram showing an example of a page in which analysis data and the purpose of analysis are input. It is a diagram showing an example of a page in which a plurality of classification categories are displayed. It is a diagram showing an example of a page in which analysis results are displayed. It is a diagram showing another example of a page in which analysis results are displayed. It is a diagram showing an example of the purpose of analysis and text data. It is a diagram showing the flow of processing in an information processing system. It is a diagram showing an example of a screen presenting a document summary and label proposals. It is a diagram showing the flow of processing in an information processing system. It is a diagram showing an example of a screen presenting a document summary and label proposals. It is a diagram showing an example of a screen presenting a document summary and label proposals. It is a diagram showing the flow of processing in an information processing system. It is a diagram showing a display example of each item and each summary. It is a diagram showing a display example in which elements serving as keywords in each item and each summary are displayed so as to be selectable. It is a diagram showing a display example of a list of related texts associated with an element selected by a user operation. It is a block diagram showing an example of the hardware configuration of a computer.

[0007] Hereinafter, embodiments of this disclosure will be described with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant descriptions will be omitted.

[0008] An information processing device of the present disclosure is, as an example without limitation, a computer 7, as shown in Figure 17, which includes one or more memories (memories 72, 73) and one or more processors (processor 71). The memories are storage units for storing programs, and the processors are processing units that execute programs and perform the processing related to the present disclosure.

[0009] The information processing system of this disclosure comprises one or more information processing devices. An example of the information processing system according to the first to ninth embodiments of this disclosure includes a web browser that displays a screen, and displays the results of an analysis performed on a data file selected by the user in the web browser, in accordance with the purpose of analysis entered by the user in the web browser. The analysis may include classifying each data item in the data file into categories, performing sentiment analysis on each data item, or other analyses. The selection of data files in the web browser, the input of the purpose of analysis and classification categories in the web browser, and the display of analysis results are achieved using external devices 9B connected via the device interface 75, such as a keyboard, pointing device, display, touchscreen, microphone, etc. In this disclosure, "classification category (category)" may conceptually include classification items, tags, or labels assigned to data.

[0010] (First Embodiment) A first embodiment of the information processing system of the present disclosure will be described with reference to Figures 1 to 6. In this embodiment, the information processing system has one or more information processing devices as client devices and one or more information processing devices as server devices, and the client devices and server devices communicate with each other via a communication network such as the Internet using their respective network interfaces 74. The client devices may function as terminals operated by users, and the server devices may function as analysis devices that analyze data received from terminals.

[0011] The client device (one or more processors) accesses the server device to perform data analysis, and the server device (one or more processors) generates a page in response to the access and sends that page to the client device. The steps described below are examples, and their execution order may be changed as appropriate.

[0012] Figure 1 is a diagram showing the processing flow in the information processing system of this embodiment.

[0013] In step S101, the client device displays the received page in a web browser and accepts a data file containing the data to be analyzed from the user via that page. That is, the client device accepts a data file specified by the user. For example, the client device may accept the data file when the user enters the file path of the data file on the page, or when the user drags and drops the icon of the data file onto the page.

[0014] Figure 2 shows an example of a page displayed in the web browser of the client device in step S101. Page 300 includes an "analysis results" area 301 that displays the results of analyzing the data to be analyzed, an "analysis data" area 302 that accepts the data to be analyzed, and an "analysis purpose" area 303 that accepts the purpose of analyzing the data.

[0015] For example, the client device accepts a data file when the user selects the "Select File" button 304 displayed in the "Analysis Data" area 302 and selects a data file from the displayed file selection dialog, or when the user enters the file path of the data file. Alternatively, the client device accepts a data file when the user drags and drops the data file icon onto the "Analysis Data" area 302.

[0016] Let's return to Figure 1 for explanation. In step S102, the client device sends the received data file to the server device.

[0017] In step S103, the server device extracts and stores each data item to be analyzed from the received data file. For example, the server device may extract the first column of each record in a CSV (Comma Separated Values) format data file as each data item to be analyzed. Alternatively, the server device may extract the columns specified by the user in each record of a CSV format data file as each data item to be analyzed. Another example is that the server device may extract each row of a text format data file as each data item to be analyzed.

[0018] In step S104, the server device sends the data extraction results to the client device. The extraction results may include each piece of data separated into records.

[0019] The series of steps for data extraction from step S102 to step S104 may be performed on the client device instead of sending the data file to the server device. For example, each data point to be analyzed may be extracted on the client device by splitting the given data file into columns or rows.

[0020] In step S105, the client device displays the received extraction results on page 300, which is displayed in the web browser. The client device then accepts the purpose of analyzing the data (the purpose of analysis) via the "Purpose of Analysis" area 303 on page 300. The purpose of analysis may be expressed as text information, audio information, or other information. For example, the client device may accept text information entered by the user via keyboard, or text information entered by the user via voice. The client device may also accept operations from the user to modify, add, or delete data on the page displaying the received extraction results, and may update the data stored in the server device by sending the data after the addition or deletion operation is completed to the server device.

[0021] Figure 3 shows an example of a page displayed in the web browser of the client device in step S105. Page 310 displays the data extraction results 311 in the "Analysis Data" area 302 and the purpose of the analysis 312 received from the user in the "Purpose of Analysis" area 303. For example, the client device receives the purpose of analysis as, "I want to analyze the reviews of our company's data analysis application and identify the issues that need to be solved."

[0022] Let's return to Figure 1 for explanation. In step S106, the client device transmits the purpose of the analysis to the server device.

[0023] In step S107, the server device obtains candidate categories (classification categories) for each of the multiple categories to be classified, which were extracted in step S103, based on the purpose of the received analysis.

[0024] As an example, the server device may generate one or more prompts (input information) including at least one of "the purpose of the analysis" and "an instruction to propose several classification categories suitable for achieving the purpose of the analysis," input these prompts to a generative model, and obtain several candidate categories from the generative model's response. In this case, the server device may also specify to the generative model via the prompts the number of proposed categories, their upper limit, and / or the upper limit of the number of characters in the proposed categories, and the generative model may generate candidate categories in response to the input prompts and return a response containing them to the server device.

[0025] Here, a prompt containing information (including text and instructions) includes not only prompts that contain that information as is, but also prompts that contain information that has been processed to the extent that its content is not excessively altered. The same applies to other prompts described later.

[0026] The generative model may be, for example, a machine learning model, or a pre-trained language model. The generative model may be, for example, a Transformer-based model. The generative model may also be a language model such as a Large Language Model (LLM) or a Vision Language Model (VLM), which includes conversational AI (Artificial Intelligence) such as ChatGPT®. In the following, the generative model may be the same or different generative models. Furthermore, the generative model used by the information processing system may be provided by the information processing system itself or by an external device.

[0027] As another example, the server device may vectorize the "purpose of analysis" into a text vector, and based on the resulting text vector, select a pre-prepared set of categories, or search for a corresponding category from a pre-prepared set of categories to obtain multiple category candidates. For example, based on the purpose of analysis 312 shown in Figure 3, the server device may obtain five categories as classification category candidates: "functionality," "usability," "performance," "design," and "support."

[0028] In step S108, the server device sends the acquired list of candidate categories to the client device.

[0029] The series of steps from step S106 to step S108, for obtaining classification categories based on the purpose of analysis, may be performed on the client device instead of sending the purpose of analysis to the server device. For example, the client device may use classification categories that have been pre-cached according to the purpose of analysis.

[0030] In step S109, the client device displays the received category candidates on the page shown in the web browser. The client device may also display the category candidates on the page in a modal window. The client device may also accept requests from the user to modify, add, or delete the category candidates displayed on the page.

[0031] Figure 4 shows an example of a page displayed in the web browser of the client device in step S109. Page 320 displays multiple classification categories received from the server device in a modal window, and accepts user requests to modify or delete these classification categories or add new classification categories through this modal window. For example, based on the analysis objective 312 shown in Figure 3, the client device displays the five categories "Functionality," "Usability," "Performance," "Design," and "Support," which the server device has acquired as candidate classification categories, in a modal window, and accepts user requests to modify or delete these categories or add new categories.

[0032] Let's return to Figure 1 for explanation. In step S110, when the client device receives an instruction from the user to start the analysis, it sends the multiple classification categories determined by the user, which have been modified, added, or deleted as needed, to the server device. For example, when the client device is displaying the five categories shown on page 320 of Figure 4 ("Functionality", "Usability", "Performance", "Design", and "Support"), and the user selects the "Start Analysis" button 321, it sends these five categories to the server.

[0033] In step S111, the server device classifies each data extracted in step S103 into one of the received classification categories. As an example, the server device may classify each data into one of the classification categories by generating one or more prompts that include the received "multiple classification categories," "each data," and "an instruction to classify each data into one of those multiple classification categories," inputting these prompts into a generative model, and obtaining the classification result for each data from the generative model's response.

[0034] In step S112, the server device performs sentiment analysis on each data extracted in step S103. As an example, the server device may perform sentiment analysis on each data by generating one or more prompts including "each data" and "instructions to classify each data as either positive or negative," inputting these prompts into a generative model, and obtaining the classification result for each data from the generative model's response. The sentiment analysis may classify each data as either positive, negative, or neutral.

[0035] In step S113, the server device sends the classification result from step S111 and the sentiment analysis result from step S112 to the client device as analysis results. The server device may format the analysis results and send the formatted analysis results to the client device. For example, the server device may group each data by classification category and format each data so that it is displayed by classification category on the client device. For example, the server device may link each data to the sentiment analysis result and format each data so that it is displayed on the client device along with information corresponding to the sentiment analysis result (icon (e.g., positive icon), text, numerical value, etc.).

[0036] The server device may aggregate the analysis results and send the aggregated analysis results to the client device. The server device may also generate a graph of the aggregated analysis results and send that graph to the client device. For example, the server device may generate a graph showing the number of data points in each classification category, or a graph showing the ratio of the number of data points in each classification category to the total number of data points, based on the classification results in step S111, and send that graph to the client device. For example, the server device may aggregate the number of data points for each classification based on the sentiment analysis results of each data point and send that number to the client device.

[0037] In step S114, the client device displays the received analysis results on a page shown in the web browser. The client device may display all the data to be analyzed for each classification category, or, if it receives a selection of a category to be displayed by the user, it may display the data for the selected category.

[0038] The client device may automatically display the graphs included in the analysis results (i.e., graphs summarizing the analysis results). Automatically displaying the graphs included in the analysis results conceptually includes displaying them automatically without any additional user operation, or displaying them automatically in response to the user's operation of the graph display button, meaning that the graphs are displayed without the user directly performing any aggregation operations on the analysis results. Aggregation operations include specifying the range of data to be graphed and selecting the data series (data type) to be graphed. Thus, in this embodiment, when the client device receives an analysis start instruction from the user in step S110, the information processing system can automatically present the data analysis results to the user in a visualized form using graphs, without requiring the user to perform any aggregation operations on the analysis results.

[0039] Figures 5 and 6 show examples of pages displayed in the web browser of the client device in step S114. Page 330, shown in Figure 5, displays the analysis results received from the server device in the "Analysis Results" area 301. Page 330 displays a pie chart 331 on the left side showing the number of data items classified into each classification category and their proportion to the total. On the right side of page 330, it displays category buttons 332 for displaying the data classified into each classification category, and a list 333 of the data classified into the classification category currently selected by the category button 332.

[0040] Page 330 may display the pie chart 331 in a manner that allows the user to identify which classification category is currently selected. The display manner is not limited to any particular manner, such as making the portion of the pie chart 331 corresponding to the currently selected classification category larger than other portions, making it blink, or surrounding it with a thick border.

[0041] Page 330 accepts the user's selection of a displayed classification category via the category button 332. Page 330 also displays the classification results and the sentiment analysis results for each data item classified into each category. The sentiment analysis results may include aggregated results for each classification category and results for each individual data item.

[0042] For example, as illustrated in Figure 5, the client device displays on page 330 the following as analysis results: a pie chart 331 showing the number and percentage of data classified into each of the five categories ("Functionality," "Usability," "Performance," "Design," and "Support") shown on page 320 of Figure 4; five category buttons 332 corresponding to each of the five categories ("Functionality" button, "Usability" button, "Performance" button, "Design" button, and "Support" button); each data 333 classified into the currently selected classification category ("Functionality" category); the aggregated sentiment analysis results 334 for each data classified into the currently selected classification category (11 positive opinions, 6 negative opinions); and the sentiment analysis results 335 for each data (the 1st to 3rd, 5th, and 6th data are positive opinions, and the 4th data is a negative opinion). As illustrated in Figure 5, the client device displays the section of the pie chart 331 corresponding to the "Functionality" category larger than the other categories so that the user can identify that the currently selected classification category is the "Functionality" category.

[0043] Also, for example, when the client device receives the selection of the "Support" button by the user, it selects the "Support" category as the displayed classification category. Then, for example, as illustrated in FIG. 6, the client device displays on page 340, as analysis results, a pie chart 341 showing the number and proportion of data classified into each of the five aforementioned categories, five category buttons 342 corresponding to each of the five categories, each data 343 classified into the currently selected classification category ("Support" category), the aggregated result 344 of the sentiment analysis of each data classified into the currently selected classification category (3 positive opinions and 9 negative opinions), and the result 345 of the sentiment analysis of each data (the 1st, 3rd, and 6th data are positive opinions, and the 2nd, 4th, and 5th data are negative opinions). As illustrated in FIG. 6, the client device displays the portion corresponding to the "Support" category in the pie chart 341 larger than the other categories so that the user can identify that the currently selected classification category is the "Support" category.

[0044] Returning to FIG. 1 for explanation. In step S115, the server device may save the analysis result transmitted to the client device in step S113 as an analysis history.

[0045] In the present embodiment, the server device may execute only one of the analysis (category classification) in step S111 and the analysis (sentiment analysis) in step S112, transmit the analysis result including the executed one analysis to the client device, and the client device may display the analysis result.

[0046] In the present embodiment, instead of transmitting the graph and the aggregated result of the data to the client device, the server device may transmit the data necessary for the client device to draw the graph and the data necessary for the client device to aggregate the data. Then, the client device may use these data to draw the graph or aggregate the data.

[0047] In this embodiment, the generation model used by the server device may be provided by the server device itself, or may be provided by another third device (information processing device) that is not the server device or the client device. When the server device uses the generation model, it may use an API (Application Programming Interface) for the generation model. Also, the generation model in this embodiment may be the same model or different models.

[0048] Hereinafter, the configurations according to the second to ninth embodiments will be described focusing on the differences from the first embodiment. Unless otherwise specified, the same processing as in the first embodiment is performed.

[0049] (Second Embodiment) In step S111, the server device may obtain data classification examples exemplified in the generation model for classification for each received category. A data classification example is an example of data and the classification category into which the data is classified. And in step S111, the server device generates one or more prompts including the received "plurality of classification categories", the obtained "each classification example", "each data", and "an instruction to classify each data into any one of the plurality of classification categories", inputs the prompt into the generation model, and may execute a series of processes of obtaining the classification result of each data from the response of the generation model.

[0050] As an example, the server device generates one or more prompts including the "purpose of analysis" received in step S106, the "plurality of classification categories" received in step S110, and "an instruction to exemplify data to be classified into each classification category together with the classification category", inputs the prompt into the generation model, and may obtain data classification examples from the response of the generation model thereto.

[0051] As another example, the server device may select a corresponding data classification example from a set of pre-prepared data classification examples based on the "purpose of analysis" received in step S106 and the "multiple classification categories" received in step S110, and acquire those selected data classification examples.

[0052] (Third Embodiment) In step S111, the server device may classify each data extracted in step S103 into two or more of the received classification categories. As an example, the server device may classify each data into one or more of the multiple classification categories (i.e., at least one of the multiple classification categories) by generating one or more prompts including the received "multiple classification categories", "each data", "an instruction to classify each data into one of those multiple classification categories", and "an instruction to classify the data into either of the two or more classification categories if the data can be classified into either of the two or more classification categories", inputting the prompts into the generation model, and obtaining the classification result for each data from the generation model's response.

[0053] (Fourth Embodiment) In step S107, the server device may, as an example, obtain a plurality of candidate classification categories by the method described above, and select a smaller number of classification categories from the obtained plurality of classification categories. In step S108, the server device may transmit the selected classification categories to the client device.

[0054] The selection of a smaller number of classification categories may be carried out as follows: For example, the server device may classify a portion of the sampled data extracted in step S103 into the acquired classification categories, and as a result of the classification, select a predetermined number or more classification categories. Alternatively, for example, the server device may classify a portion of the data extracted in step S103 (a predetermined number of sampled data, etc.) into the acquired classification categories, and as a result of the classification, select a predetermined number of classification categories in descending order of the number of classified data.

[0055] (Fifth Embodiment) In step S107, the server device may obtain candidate classification categories for each of the multiple classification categories to be classified based on the data to be analyzed.

[0056] As an example, the server device may generate one or more prompts that include some of the data extracted in step S103 (such as a predetermined number of sampled data) and instructions that propose a plurality of classification categories suitable for classifying that portion of the data, input these prompts into the generation model, and obtain candidate categories from the generation model's response.

[0057] As another example, the server device may vectorize some of the data extracted in step S103 (such as a predetermined number of sampled data) into text vectors, and based on the obtained text vectors, select a set of pre-prepared categories, or search for a corresponding category from the pre-prepared categories to obtain multiple category candidates. In step S107, the server device may obtain multiple category candidates by combining the acquisition of classification category candidates based on the "purpose of analysis" and the acquisition of classification category candidates based on the "data to be analyzed".

[0058] (Sixth Embodiment) After displaying the analysis results in step S114, the client device and server device may repeat the processing from step S105 to step S115. For example, in the second and subsequent steps of step S114, the client device may display the first analysis results and the second and subsequent analysis results side by side on a web browser page. By displaying multiple analysis results side by side for the same data subject, the user can be provided with data analysis from multiple perspectives.

[0059] (Seventh Embodiment) In step S111, the server device may sort the data classified into each category in descending order of importance, based on the "purpose of analysis" received in step S106 and the data classified into each category. In step S113, the server device may send the analysis results, including the data sorted in descending order of importance, to the client device. The client device may then display a list of data classified into the currently selected category, sorted in descending order of importance.

[0060] The server device may, for example, input one or more prompts into the generative model, including "the purpose of the analysis," the data, and "instructions to indicate the importance of the data in relation to the purpose of the analysis," obtain the importance of the data in response, and sort the data by classification category based on the importance of each data point.

[0061] (Eighth Embodiment) In step S111, the server device may acquire text data (summary text) that summarizes the content of data classified into the same classification category. In step S113, the server device may send the analysis results, including the summary text for each classification category, to the client device. The client device may then display the summary text for the currently selected classification category.

[0062] For example, the server device may input one or more prompts to the generative model, including all or more data classified into the same classification category and "instructions to summarize the content of the data," and obtain a summary text of that classification category as a response.

[0063] (Ninth Embodiment) After displaying the analysis results in step S114, the client device and server device may perform the processing from steps S105 to S115 on data that has been classified into one or fewer classification categories specified by the user from among a plurality of classification categories.

[0064] For example, in step S114 from the second time onward, the client device may display the results of the first analysis and the results of the second and subsequent analyses side by side on a web browser page. This allows the user to easily perform more detailed data (re)analysis on some of the data using the analysis results already performed.

[0065] An information processing system according to the tenth to eleventh embodiments of this disclosure, as an example, has a web browser that displays a screen, and displays the results of analyzing a data file selected by the user in the web browser on the web browser (display device). The analysis may be performed according to data entered by the user in natural language in the web browser for the purpose of data analysis. The analysis may be the assignment of classification labels to the text contained in the data file (including classification of text into classification labels), or other types of analysis. The selection of data files in the web browser, the input of the purpose of analysis and classification labels in the web browser, and the display of analysis results are realized using external devices 9B connected via the device interface 75, such as a keyboard, pointing device, display, touchscreen, microphone, etc. In this disclosure, "classification label (label)" may conceptually include the category to which the text is classified and the tag assigned to the text. Note that the text is an example of data.

[0066] (Tenth Embodiment) A tenth embodiment of the information processing system of the present disclosure will be described with reference to Figures 7 to 12. In this embodiment, the information processing system may generate one or more labels that can be assigned to each of the multiple sentences (natural language sentences) contained in the text data 102, based on those multiple sentences. The information processing system may also propose the generated one or more labels to the user. The information processing system may also receive feedback from the user regarding the labels in the form of text or a pull-down menu, and based on that feedback, generate one or more labels again and propose them to the user. Furthermore, the information processing system may select one or more labels from the generated one or more labels based on the sentences (natural language sentences) 101 written for the purpose of analyzing the text data 102, and assign them to each sentence contained in the text data 102.

[0067] Figure 8 is a diagram showing the processing flow in the information processing system of this embodiment. The information processing system may execute the processing flow shown in Figure 8 by having the processor execute a program.

[0068] The information processing system divides the multiple sentences contained in the text data 102 into multiple chunks and performs a mapping step to generate feature data representing the characteristics of each chunk. Subsequently, the information processing system performs a reduce step to summarize the entire set of multiple sentences contained in the text data 102 from the feature data of each chunk and generate the necessary labels.

[0069] A chunk may contain multiple sentences. The same sentence may be included in multiple chunks, or it may not be included in multiple chunks. The chunking of multiple sentences is not limited to a specific method.

[0070] The feature data generated by the mapping step for each chunk may be text written in natural language. In this embodiment, as an example of feature data, a summary of the sentences contained in the chunk and the labels to be prepared to be assigned to the sentences contained in the chunk are used. In other embodiments, only the summary may be used, or other data that can be processed in the reduce step may be used.

[0071] The amount of processing (input) per step can be controlled by controlling the chunk size (the amount of text contained in one chunk) in the mapping step. The total amount of processing (output) can be controlled by controlling the number of chunks (number of divisions) in the mapping step.

[0072] The map step may be executed multiple times, and a portion of the results from those multiple executions may be selected to execute the reduce step. The reduce step may be a multi-stage process of two or more stages. When this process is executed multiple times, in the second and subsequent executions, the map step may be omitted, and only the later reduce steps may be re-executed in accordance with the changes in the instruction. In other words, the results of the map step in the first execution may be reused to execute the reduce steps in the second and subsequent executions according to the instruction.

[0073] The information processing system generates one or more labels based on multiple sentences contained in the text data 102 and proposes them to the user, and generates and proposes other label proposals to the user in response to the user's feedback on this proposal. User feedback is not limited to specific content and may include, for example, instructions to further subdivide a particular label among the one or more proposed labels, or instructions to propose a new label considering a viewpoint specified by the user. Instructions may be given by the user through direct input of sentences written in natural language, or by the user through selection input from one or more options provided by the information processing system as a pull-down menu, and are not limited to a specific format.

[0074] The information processing system may provide these options as predetermined. The information processing system may have a generation model propose one or more feedback comments that a user might think of regarding the content and trends of multiple sentences included in the text data 102, and / or the last proposed label, and provide these as options. For the sake of explanation below, user feedback will be described as being provided by the user directly inputting text.

[0075] The processing flow shown in Figure 8 will be explained in more detail.

[0076] (Map step) The information processing system divides the text data 102 into multiple chunks (for example, texts A to Z). For each chunk, the information processing system generates a prompt that includes the multiple texts contained in that chunk, instructions to summarize the content of those texts, and instructions to suggest multiple labels to be prepared for labeling those texts.

[0077] The information processing system inputs the generated prompts into the generative model for each chunk, and in response, obtains from the generative model a summary of the text and several labels to be prepared for those texts. Here, each label consists of its name (label name) and a summary of the text in the chunk for that label (a summary of the text from the perspective of that label), and the prompt includes instructions to suggest them.

[0078] (Reduce Step) The information processing system generates a prompt that includes the chunk-by-chunk text summaries and labels obtained in the map step, instructions to summarize the content of the text summaries for all chunks, and instructions to suggest multiple labels to be prepared for the text.

[0079] The information processing system inputs the generated prompt into the generative model and, in response, obtains a summary and several labels to be prepared from the generative model. The generative model in the reduce step may be the same as or different from the generative model in the map step. Each label, as in the map step, consists of its name (label name) and a summary of the text contained in the chunk about that label (summary of the text from the perspective of that label), and the prompt includes instructions to suggest them.

[0080] Figure 9 shows an example of a screen that presents a document summary and proposed labels. As an example, the information processing system displays the screen 200 shown in Figure 9 based on the summary and labels obtained in the reduce step. The screen 200 displays the summary obtained in the reduce step in the summary field 201 as a summary of the multiple sentences contained in the document data 102. The screen 200 displays seven labels in the label field 202 as labels to be assigned to the multiple sentences contained in the document data 102. The information processing system can assign these labels to each of the multiple sentences contained in the document data 102.

[0081] The information processing system may display a button (not shown) on screen 200, and in response to the user's selection of the button, automatically select one or more labels from the proposed labels in the label field 202 based on the text 101 described for the purpose of analyzing the text data 102, and assign them to each text contained in the text data 102. The automatic assignment of labels to each text may be performed using, for example, the method disclosed in Patent Document 1, but is not limited to a specific method. The information processing system may display the label assignment results on a separate screen in the form of a table and / or graph.

[0082] Furthermore, screen 200 displays an input field 210 for receiving text written in natural language by the user (user comments) as feedback regarding the summary field 201 or the label field 202. Based on the feedback entered by the user in the input field 210, the information processing system may execute the processing flow shown in Figure 10.

[0083] The processing flow shown in Figure 10 will be explained below.

[0084] (Map step) The information processing system obtains the feedback text (user comments) entered in the input field 210 of the screen 200. The information processing system then generates a prompt for each chunk that includes multiple sentences contained in that chunk, instructions to summarize the content of those sentences, instructions to suggest multiple labels to be prepared for assigning labels to those sentences, and the user comments. The instructions to summarize the content of the sentences may also be instructions to summarize while taking the user comments into consideration. The instructions to suggest labels may also be instructions to suggest labels while taking the user comments into consideration.

[0085] The information processing system may include in the user comments not only the feedback text entered in the input field 210 this time, but also the "text summary and multiple labels" obtained from the generative model in the previously executed map step. The information processing system may also include in the user comments not only the feedback text entered in the input field 210 this time, but also the history of user comments obtained so far. The history of user comments may, for example, be a chat-formatted version of the "text summary and multiple labels" obtained from the generative model up to that point for that chunk, along with the text that the user has entered as feedback in the input field 210 up to that point.

[0086] The information processing system inputs the generated prompts into the generative model for each chunk, and in response, obtains a summary of the text and several labels to be prepared for those texts from the generative model.

[0087] In this way, the information processing system can obtain text summaries and labels that reflect user feedback, chunk by chunk.

[0088] (Reduce Step) The information processing system obtains the feedback text (user comment) entered in the input field 210. The information processing system then generates a prompt that includes the chunk-by-chunk text summary and labels obtained in the map step, instructions to summarize the content of the text summaries for all chunks, instructions to suggest multiple labels to be prepared for the text, and the user comment. The user comment may include not only the feedback text entered in the input field 210 this time, but also past user comments. The user comment may also include summaries and labels obtained in previous reduce steps.

[0089] The information processing system inputs the generated prompt into the generative model and, in response, obtains a summary and several labels to be prepared from the generative model. Each label, like in the map step, consists of its name (label name) and a summary of the text contained in the chunk about that label (summary of the text from the perspective of that label), and the prompt includes instructions to suggest them.

[0090] In this way, the information processing system can obtain new text summaries and labels that reflect user feedback on the previously proposed text summaries and labels. As a result, the information processing system can assign the newly obtained labels to each of the multiple texts contained in the text data 102.

[0091] Figure 11 shows an example of a screen that presents a document summary and proposed labels. As an example, the information processing system displays the screen 200 shown in Figure 11 based on the summary and labels obtained by the reduce step. The screen 200 shown in Figure 11 is a screen that adds a feedback field 203, a summary field 204, and a label field 205 to the screen 200 shown in Figure 9. The feedback field 203 displays the text entered in the input field 210. The summary field 204 displays the summary obtained by the reduce step that was re-executed, as a summary of the multiple sentences contained in the document data 102. The label field 205 displays the labels obtained by the reduce step that was re-executed, as labels to be assigned to the multiple sentences contained in the document data 102. The label field 205 shown in Figure 11 displays four labels.

[0092] The information processing system may display a button (not shown) on the screen 200 shown in Figure 11, and in response to the user's selection of the button, automatically select one or more labels from the labels (four labels as an example) proposed in the label field 205 based on the text 101 written as the purpose of analyzing the text data 102, and assign them to each text contained in the text data 102. The automatic assignment of labels to each text may be performed using, for example, the method disclosed in Patent Document 1, but is not limited to a specific method. The information processing system may display the label assignment results on a separate screen in the form of a table and / or graph.

[0093] Furthermore, screen 200 displays the input field 210 blank, allowing the user to submit new text (user comments) as feedback regarding the summary field 204 or the label field 205. Based on the feedback newly entered by the user in the input field 210, the information processing system may re-execute the processing flow shown in Figure 10, and as a result of this re-execution, it may display, for example, screen 200 shown in Figure 12.

[0094] Figure 12 shows an example of a screen that presents a document summary and suggested labels. The screen 200 shown in Figure 12 is the same as the screen 200 shown in Figure 11, but with the addition of a feedback field 206, a summary field 207, and a label field 208. The feedback field 206 displays the text entered in the input field 210 as user feedback to the summary field 204 or label field 205 presented by the information processing system. The summary field 207 displays a summary of the content of all texts included in the text data 102, obtained by the information processing system re-executing the map step and reduce step based on that feedback. The label field 208 displays suggested labels to be assigned to each text included in the text data 102, obtained by the system re-executing the map step and reduce step based on that feedback.

[0095] Even at this stage, the information processing system may display a button (not shown) on the screen 200 shown in Figure 12, and, in response to the user's selection of the button, automatically select one or more labels from the labels (four labels as an example) proposed in the label field 208 based on the text 101 described as the purpose of analyzing the text data 102, and assign them to each text contained in the text data 102. The automatic assignment of labels to each text may be performed using, for example, the method disclosed in Patent Document 1, but is not limited to a specific method. The information processing system may display the label assignment results on a separate screen in the form of a table and / or graph.

[0096] According to the information processing system of this embodiment, the information processing system can re-propose new label proposals that reflect user feedback on label proposals previously suggested by the system, and can obtain preferred label proposals by analyzing the text data 102.

[0097] (Eleventh Embodiment) The information processing system of the tenth embodiment divides the multiple sentences contained in the text data 102 into multiple chunks, performs a mapping step which processes each chunk, and performs a reduce step which integrates the processing results of each chunk, thereby summarizing the content of all sentences contained in the text data 102 and proposing label proposals to be assigned to those sentences. This prevents a significant decrease in processing efficiency and / or accuracy when the amount of text in the text data 102 is enormous.

[0098] In the eleventh embodiment of the information processing system, the text data 102 may be processed all at once by a generative model without dividing the text into multiple chunks or separating it into a map step and a reduce step. In this case, the information processing system generates a prompt that includes all the text contained in the text data 102, instructions to summarize the content of those texts, and instructions to propose multiple labels to be prepared for those texts. The information processing system then inputs the generated prompt into the generative model and, as a response, obtains the summary and the multiple labels to be prepared from the generative model.

[0099] Furthermore, when receiving user feedback (user comments), the information processing system generates a prompt containing all the texts included in the text data 102, instructions to summarize the content of those texts, instructions to propose multiple labels to be assigned to those texts, and the user comments. In this case, the information processing system may also include the last label proposal made by the information processing system in the prompt.

[0100] Thus, the information processing system of this embodiment can also obtain new text summaries and labels that reflect user feedback on the previously proposed text summaries and labels. As a result, the information processing system can assign the newly obtained labels to each of the multiple texts contained in the text data 102.

[0101] (Twelfth Embodiment) An information processing system according to the twelfth embodiment of the present disclosure, as an example, has a web browser that displays a screen, and displays the results of analyzing a plurality of texts acquired via the web browser on the web browser (display device). The information processing system may analyze the plurality of texts, add annotations to the analysis results, and display the annotated analysis results. For example, if the analysis result is a summary of a plurality of texts, the information processing system analyzes the plurality of texts, summarizes the plurality of texts for each of the plurality of items, and displays each item and each summary on the web browser. At this time, the information processing system may divide each summary into elements that are keywords of the summary (keyword elements), link each of the plurality of texts to the keyword elements that are most relevant in the summary, and when a keyword element is selected, it may display a list of the keyword elements that are most relevant. The information processing system may also display a summary in which the number of keyword elements that are most relevant is added to the keyword elements in the summary. In this way, the information processing system may add annotations to each item and the summary sentence (summary) for each item, and display each item and each summary that has been annotated. User input within the summary text and the display of the summary are achieved using external devices 9B, such as a keyboard, pointing device, display, touchscreen, and microphone, which are connected via the device interface 75.

[0102] (Twelfth Embodiment) A twelfth embodiment of the information processing system of the present disclosure will be described with reference to Figures 13 to 16. In this embodiment, the information processing system analyzes multiple texts and displays a summary of the multiple texts as the analysis result.

[0103] Figure 13 is a diagram illustrating the processing flow in an information processing system.

[0104] In step S301, the information processing system acquires multiple texts. In this disclosure, "text" may refer to a single sentence or a collection of sentences. The text is an example of data.

[0105] The method for obtaining multiple texts is not limited to any particular method. An information processing system may obtain multiple texts by having the user manually input multiple texts or by having the user select a file containing multiple texts. For example, an information processing system may obtain a block of text that includes line breaks and obtain the text between line breaks as a single text.

[0106] In step S302, the information processing system uses a generative model to summarize multiple texts into one or more items. The information processing system may input a prompt written in natural language into the generative model and obtain a response from the generative model to the prompt input. At this time, the generative model may perform processing according to the input prompt and output the processing result as a response written in natural language.

[0107] The information processing system may obtain one or more predetermined items. The information processing system may obtain one or more items described in natural language by a user. The information processing system may input a prompt to a generative model that includes multiple texts and instructions to analyze those texts and suggest one or more items to summarize them, and obtain one or more items from the generative model as a response described in natural language. The information processing system may obtain one or more items by combining the methods described above or other methods.

[0108] One or more items may be structured hierarchically in a tree-like manner across multiple levels (major items, minor items, etc.), or they may not be hierarchical. Items may include concepts such as labels, categories, and classes, as examples.

[0109] An information processing system may input a prompt to a generative model that includes multiple texts and instructions to summarize those multiple texts with respect to each of the acquired items, and obtain a summary of the multiple texts for each item from the generative model as a response written in natural language.

[0110] An information processing system may input a prompt to a generative model that includes multiple texts, one or more items for analyzing those texts and summarizing them, and instructions for summarizing the texts from each of those one or more perspectives, and obtain from the generative model, as a response written in natural language, one or more items and summaries of the texts for each item.

[0111] An information processing system may input a prompt to a generative model that includes multiple texts and instructions to divide those texts into multiple items and summarize them, and may obtain from the generative model, as a response written in natural language, one or more items and summaries of the multiple texts for each item. The instructions may include instructions to output each item and a summary for each item.

[0112] In step S303, the information processing system displays one or more acquired items and a summary for each item on a display device. The information processing system may, for example, display one or more acquired items and a summary for each item on the screen 400 shown in Figure 14.

[0113] Figure 14 shows examples of how each item and its summary are displayed. In Figure 14, there are three items: "Location and Access," "Room Quality," and "Meal Quality." The text following each item is a summary of that item.

[0114] In step S304, the information processing system adds annotations to each item and a summary for each item, and displays the annotated items and summaries on screen 400.

[0115] The information processing system structures the acquired one or more items and a summary for each item. The information processing system extracts the key elements of the summaries from each summary for each item and acquires these elements as keywords. The key elements are strings of characters in the summaries. Depending on the context, a keyword may refer to the element in the summary that serves as the keyword, a string of characters that indicates the keyword, or both. The information processing system may extract keywords not limited to words, but in units that make sense. If multiple items are hierarchically structured in a tree, the information processing system may extract keywords from each summary for each of the lowest-level items.

[0116] The information processing system structures each item and each summary based on each item and the keywords extracted from each summary. The information processing system may link each item and the elements extracted as keywords from the summary of that item to each other. Elements linked to an item belong to that item. If multiple items are hierarchically structured in a tree, the information processing system may link the elements extracted as keywords from the summary of the lowest-level item to the item corresponding to the summary from which the elements were extracted, and to the higher-level item that contains that item. The information processing system may structure each item and each summary by linking each item and the elements extracted as keywords from the summary of that item. The information processing system may structure each item and each summary by other means.

[0117] An information processing system may use a generative model to extract key elements and / or structure each item and each summary. For example, the information processing system may input a prompt to a generative model that includes each summary for each item and an instruction to extract key elements from each summary and output data linking those elements to each item, and then structure each item and each summary from the response written in natural language.

[0118] For example, the information processing system could extract five elements as keywords from the summary of the item "Location and Access" shown in Figure 14: "The hotel is about a 20-minute walk from the station," "There is a shuttle bus service," "It can get crowded during busy periods," "There are few restaurants in the surrounding area," and "You need to go to the area around the station." The information processing system could then link each extracted element to the item "Location and Access" by assigning it to the item, thereby structuring the item "Location and Access" and its summary.

[0119] The information processing system identifies text associated with each element extracted as a keyword in the summary, and links that element with the associated text. For each element extracted as a keyword, the information processing system may identify text related to a combination from multiple texts based on the combination of that element and the item to which that element belongs, and the content of each of the multiple texts. For example, the information processing system may identify text related to a combination by analyzing the relationship between the combination of an element and an item and the content of each of the multiple texts using a generative model. The prompts input to the generative model are not limited to specific ones. For example, a prompt may include a combination of an element and an item, multiple texts, and an instruction to respond with the text that is most relevant to that combination from among the multiple texts. The information processing system may link an element with its associated text by assigning the identified text to that element.

[0120] An information processing system may identify one or more combinations related to each of several texts based on the content of the text and each combination of elements extracted as keywords and the items to which those elements belong. For example, an information processing system may identify one or more combinations related to a text by analyzing the relationship between the text and each combination of elements and items using a generative model. The prompts input to the generative model are not limited to specific ones. For example, the prompts may include each combination, the text, and instructions to respond with one or more combinations that the text is closely related to. An information processing system may link a text with related elements by assigning the elements included in the identified combinations to the text.

[0121] Furthermore, the process of linking text with elements extracted as keywords can also be referred to as assigning or classifying text to (one or more) elements.

[0122] The information processing system displays each item and its summary on a display device, based on structured items and summaries, and multiple texts associated with elements in the summaries, so that users can select each item and each keyword element in the summaries through user operations such as clicking or tapping.

[0123] The information processing system may count the number of texts associated with each element in the summary and display the counted number on a display device. The information processing system may count the total number of texts associated with each element belonging to each item and display the counted number on a display device as the number of related texts associated with that item. If the items are hierarchical, the information processing system may count the total number of texts associated with each element belonging to each of the lower-level items of a higher-level item and display the counted number on a display device as the number of related texts associated with that higher-level item.

[0124] When an element is selected in the summary, the information processing system may display a list of text associated with the selected element on the display device. When an item is selected, the information processing system may display a list of text associated with each element belonging to the selected item as a list of related texts associated with that item on the display device. When a higher-level item is selected in a hierarchical structure, the information processing system may display a list of text associated with each element belonging to each of the lower-level items of the selected item as a list of related texts associated with that higher-level item on the display device.

[0125] Figure 15 shows an example of a display that shows each item and its summary, allowing the user to select keywords within each item and summary. The information processing system displays, for example, three items—"Location and Access," "Room Quality," and "Meal Quality"—and a summary for each item, as shown in screen 410 in Figure 15, allowing the user to select keywords within each item and summary. For example, for the item "Location and Access," the information processing system displays the item and five elements belonging to it—"The hotel is about a 20-minute walk from the station," "There is a shuttle bus service," "It can get crowded during busy periods," "There are few restaurants in the surrounding area," and "It is necessary to go to the station area"—as selectable by the user.

[0126] The information processing system displays, for example, "44" next to the item "Location and Access" as the number of related texts associated with the item "Location and Access". The information processing system also displays, for example, "29" next to the element "The hotel is about a 20-minute walk from the station" as the number of related texts associated with the element "There are few restaurants in the surrounding area" which belongs to the item "Location and Access". The same applies to other items and elements.

[0127] For example, when the item "Location and Access" is selected by the user, the information processing system displays a list of related texts associated with that item. This list is a list of 44 related texts, which corresponds to the union of the related texts associated with each of the five elements belonging to the item "Location and Access".

[0128] Figure 16 shows an example of a display of a list of related texts associated with a keyword element in a summary, which is shown when the user selects that element. When the element "There are few restaurants in the vicinity" is selected by the user, the information processing system displays a list of 29 related texts associated with that element, for example, as shown in screen 420 in Figure 16.

[0129] The information processing system may, after displaying each item and its summary on a display device, accept additional instructions from the user in natural language and perform further text analysis in accordance with those instructions. In this further text analysis, the information processing system may also perform the aforementioned structuring of items and summaries and classification of text.

[0130] In the example above, the information processing system analyzed multiple texts, displayed the analysis results with annotations, and displayed an annotated summary of the multiple texts as the analysis result. Other examples of analysis results are described below. The information processing system may input a prompt containing multiple texts and instructions corresponding to the analysis content into a generative model, and obtain the analysis results from the generative model as its response.

[0131] An information processing system may analyze multiple texts and display an annotated text containing a proposal based on those texts, along with the reasons for that proposal. The information processing system may link one or more texts related to the reason from among the multiple texts to the part of the text that states the reason as an annotation. For example, the text could be "How about A? Reason: There is opinion B." The information processing system may display a text in which the related text is linked to the phrase "opinion B" in the text, and when that phrase is selected by the user, it may display the related text.

[0132] An information processing system may analyze multiple texts and display annotated sentences containing a condition and the number of texts that match that condition. The information processing system may associate one or more texts that are counted in the count with annotations in the part of the sentence where the count is stated. For example, the sentence could be "There are N data items that match the condition...", and the information processing system could display a sentence in which the related texts that were counted in the N items are associated with the "N items" statement in the sentence, and when that statement is selected by the user, the related texts can be displayed.

[0133] An information processing system may analyze multiple texts and display annotated sentences containing a condition and a statement indicating that no texts match that condition. The information processing system may associate the text that is most relevant to the condition with the entire sentence as an annotation. For example, the sentence could be "No data was found that matched the condition...", and the information processing system could display the sentence to which the text most relevant to the condition is associated, and if that sentence is selected by the user, it could display the related text.

[0134] (Other Embodiments) Embodiments may be formed by appropriately combining some of the embodiments described above.

[0135] For example, the information processing systems according to the first to ninth embodiments may be combined with the information processing systems according to the tenth to eleventh embodiments. That is, the information processing system may generate one or more labels that can be assigned to each of the multiple data contained in a data file, based on those multiple data, and classify the multiple data by assigning the generated one or more labels to each of the multiple data, and display the classification results of the multiple data.

[0136] For example, the information processing systems according to the first to ninth embodiments may be combined with the information processing system according to the twelfth embodiment. That is, the information processing system may add annotations to summary sentences that summarize each of the multiple data contained in a data file, and display a graph that aggregates the multiple data based on the annotated summary sentences.

[0137] For example, the information processing systems according to the 10th to 11th embodiments may be combined with the information processing system according to the 12th embodiment. That is, the information processing system may generate one or more labels that can be assigned to each of the multiple data contained in a data file, based on those multiple data, add annotations by assigning labels to summary sentences that summarize each of the multiple data, and display the annotated summary sentences.

[0138] In the embodiments described above, some or all of the devices (client devices, server devices, information processing devices, information processing systems) may be composed of hardware, or they may be composed of information processing by software (programs) executed by a CPU (Central Processing Unit), GPU (Graphics Processing Unit), etc. If the information processing is composed of software, the software that realizes at least some of the functions of each device in the embodiments described above may be stored on a non-temporary storage medium (non-temporary computer-readable medium) such as a CD-ROM (Compact Disc-Read Only Memory) or USB (Universal Serial Bus) memory, and the information processing of the software may be executed by loading it into a computer. Alternatively, the software may be downloaded via a communication network. Furthermore, all or part of the processing of the software may be implemented in a circuit such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array), so that the information processing by the software is executed by hardware.

[0139] The storage medium for the software may be a removable medium such as an optical disc, or a fixed storage medium such as a hard disk or memory. Furthermore, the storage medium may be located inside the computer (main memory, auxiliary storage, etc.) or outside the computer.

[0140] Figure 17 is a block diagram showing an example of the hardware configuration of each device (client device, server device, information processing device, information processing system) in the embodiment described above. Each device may be implemented as a computer 7, for example, comprising a processor 71, main memory 72 (memory), auxiliary memory 73 (memory), network interface 74, and device interface 75, which are connected via a bus 76.

[0141] The computer 7 in Figure 17 has one of each component, but it may have multiple identical components. Also, although Figure 17 shows one computer 7, the software may be installed on multiple computers, and each of these multiple computers may execute the same or different parts of the software's processing. In this case, it may be a distributed computing configuration in which each computer communicates via a network interface 74 or the like to execute processing. In other words, each device (client device, server device, information processing device, information processing system) in the above-described embodiment may be configured as a system that realizes its function by one or more computers executing instructions stored in one or more storage devices. Alternatively, it may be configured so that information transmitted from a terminal is processed by one or more computers located on the cloud, and the processing results are transmitted to the terminal.

[0142] The various calculations performed by each device (client device, server device, information processing device, information processing system) in the embodiments described above may be performed in parallel using one or more processors, or using multiple computers via a network. Alternatively, the various calculations may be distributed to multiple processing cores within a processor and performed in parallel. Furthermore, some or all of the processing and means of this disclosure may be implemented by at least one of a processor and a storage device located on a cloud that can communicate with computer 7 via a network. Thus, each device in the embodiments described above may be in the form of parallel computing using one or more computers.

[0143] The processor 71 may be at least an electronic circuit (processing circuit, processing circuitry, CPU, GPU, FPGA, ASIC, etc.) that performs either computer control or calculations. The processor 71 may also be a general-purpose processor, a dedicated processing circuit designed to perform specific calculations, or a semiconductor device including both a general-purpose processor and a dedicated processing circuit. Furthermore, the processor 71 may include optical circuits or quantum computing-based calculation functions.

[0144] The processor 71 may perform calculations based on data and software input from various devices within the computer 7, and may output calculation results and control signals to these devices. The processor 71 may also control the various components of the computer 7 by executing the computer 7's OS (Operating System) or applications.

[0145] Each of the devices in the embodiments described above (client device, server device, information processing device, information processing system) may be implemented by one or more processors 71. Here, the processor 71 may refer to one or more electronic circuits arranged on one chip, or one or more electronic circuits arranged on two or more chips or two or more devices. When multiple electronic circuits are used, each electronic circuit may communicate by wire or wireless.

[0146] The main memory 72 may store instructions executed by the processor 71 and various data, and the information stored in the main memory 72 may be read by the processor 71. The auxiliary memory 73 is a memory device other than the main memory 72. These memory devices refer to any electronic component capable of storing electronic information, and may be semiconductor memory. The semiconductor memory may be either volatile memory or non-volatile memory. In each of the devices (client device, server device, information processing device, information processing system) in the above-described embodiment, the memory device for storing various data may be implemented by the main memory 72 or the auxiliary memory device 73, or by the built-in memory of the processor 71. For example, each storage unit in the above-described embodiment may be implemented by the main memory 72 or the auxiliary memory device 73.

[0147] In the embodiments described above, if each device (client device, server device, information processing device, information processing system) consists of at least one storage device (memory) and at least one processor connected to (coupled with) this at least one storage device, then at least one processor may be connected to one storage device. Also, at least one storage device may be connected to one processor. Furthermore, the configuration may include at least one processor among a plurality of processors being connected to at least one storage device among a plurality of storage devices. This configuration may also be realized by storage devices and processors included in a plurality of computers. Moreover, the configuration may include a storage device integrated with a processor (for example, a cache memory including an L1 cache and an L2 cache).

[0148] The network interface 74 is an interface for connecting to the communication network 8 wirelessly or via a wired connection. The network interface 74 can be any appropriate interface, such as one conforming to existing communication standards. Information may be exchanged between the computer 7 and an external device 9A connected via the communication network 8 through the network interface 74. The communication network 8 may be a WAN (Wide Area Network), LAN (Local Area Network), PAN (Personal Area Network), or a combination thereof, as long as information is exchanged between the computer 7 and the external device 9A. Examples of WANs include the Internet, examples of LANs include IEEE 802.11 and Ethernet (registered trademark), and examples of PANs include Bluetooth (registered trademark) and NFC (Near Field Communication).

[0149] The device interface 75 is an interface such as USB that connects directly to the external device 9B.

[0150] External device 9A is a device connected to computer 7 via a network. External device 9B is a device directly connected to computer 7.

[0151] External device 9A or external device 9B may, for example, be an input device. The input device may be a camera, microphone, motion capture device, various sensors, keyboard, mouse, touch panel, etc., and provides the acquired information to the computer 7. Alternatively, it may be a device equipped with an input unit, memory, and processor, such as a personal computer, tablet terminal, or smartphone.

[0152] Furthermore, the external device 9A or external device 9B may, for example, be an output device. The output device may be a display device such as an LCD (Liquid Crystal Display) or an organic EL (Electro Luminescence) panel, or it may be a speaker that outputs sound, etc. It may also be a device equipped with an output unit, memory, and processor, such as a personal computer, tablet terminal, or smartphone.

[0153] Furthermore, external devices 9A and 9B may be storage devices (memory). For example, external device 9A may be network storage, and external device 9B may be storage such as an HDD.

[0154] Furthermore, the external device 9A or external device 9B may be a device that has some of the functions of the components of each device (client device, server device, information processing device, information processing system) in the embodiments described above. In other words, the computer 7 may transmit some or all of the processing results to the external device 9A or external device 9B, or may receive some or all of the processing results from the external device 9A or external device 9B.

[0155] Where the expression "at least one of a, b, and c" or "at least one of a, b, or c" (including similar expressions) is used in this specification (including the claims), it includes any of a, b, c, ab, ac, bc, or abc. It also includes multiple instances of any element, such as aa, abb, aabbcc, etc. Furthermore, it includes adding other elements other than the enumerated elements (a, b, and c), such as abcd having d.

[0156] In this specification (including the claims), when expressions such as "using data as input / based on data / according to / in accordance with data" (including similar expressions) are used, unless otherwise specified, this includes using the data itself or using data that has been processed in some way (e.g., data with added noise, normalized data, features extracted from the data, intermediate representations of the data, etc.). Furthermore, when it is stated that some result is obtained "using data as input / based on data / according to / in accordance with data" (including similar expressions), unless otherwise specified, this includes cases where the result is obtained based solely on the data in question or where the result is also influenced by other data, factors, conditions, and / or states other than the data in question. Furthermore, when it is stated that "data is output" (including similar expressions), unless otherwise specified, this includes cases where the data itself is used as output or where data that has been processed in some way (e.g., data with added noise, normalized data, features extracted from the data, intermediate representations of various types of data, etc.) is used as output.

[0157] In this specification (including the claims), the terms “connected” and “coupled” are intended to be non-restrictive terms that include any direct connection / coupling, indirect connection / coupling, electrical connection / coupling, communicative connection / coupling, operational connection / coupling, or physical connection / coupling. The terms should be interpreted as appropriate in the context in which they are used, but any form of connection / coupling that is not intentionally or naturally excluded should be interpreted non-restrictively as being included in the terms.

[0158] In this specification (including the claims), when the expression "A configured to B" is used, it may include that the physical structure of element A has a configuration capable of performing operation B, and that the permanent or temporary setting / configuration of element A is configured to actually perform operation B. For example, if element A is a general-purpose processor, it is sufficient that the processor has a hardware configuration capable of performing operation B, and that it is configured to actually perform operation B by the setting of a permanent or temporary program (instruction). Furthermore, if element A is a dedicated processor, dedicated arithmetic circuit, etc., it is sufficient that the circuit structure of the processor is implemented to actually perform operation B, regardless of whether control instructions and data are actually attached.

[0159] Wherever terms meaning "comprising" or "having" are used in this specification (including the claims), they are intended to be open-ended terms, including cases where the subject matter of such terms is not the object of the term. Where the object of such terms meaning "comprising" or "having" is an expression that does not specify a quantity or suggests a singular number (an expression with the article a or an), such expression should be interpreted as not being limited to a specific number.

[0160] In this specification (including the claims), even if expressions such as "one or more" or "at least one" are used in some places, and expressions that do not specify a quantity or suggest singularity (expressions using the articles a or an) are used in other places, the latter expressions are not intended to mean "one." In general, expressions that do not specify a quantity or suggest singularity (expressions using the articles a or an) should not necessarily be interpreted as not being limited to a specific number.

[0161] In this specification, if a particular configuration of an embodiment is described as providing a specific advantage or result, it should be understood, unless otherwise stated, that the same advantage or result can also be obtained from one or more other embodiments having that configuration. However, it should be understood that the presence or absence of such advantage or result generally depends on various factors, conditions, and / or states, and that the configuration does not necessarily guarantee that the advantage or result can be obtained. The advantage or result can only be obtained from the configuration described in the embodiment when various factors, conditions, and / or states are met, and the advantage or result cannot necessarily be obtained in the claimed invention that defines that configuration or a similar configuration.

[0162] In this specification (including the claims), when multiple hardware components perform a predetermined process, each component may cooperate to perform the predetermined process, or some components may perform all of the predetermined process. Alternatively, some components may perform part of the predetermined process, while other components perform the remainder. In this specification (including the claims), when expressions such as "one or more hardware components perform a first process, and the one or more hardware components perform a second process" (including similar expressions) are used, the hardware component performing the first process and the hardware component performing the second process may be the same or different. In other words, it is sufficient that the hardware component performing the first process and the hardware component performing the second process are included in the one or more hardware components. Hardware may include electronic circuits, devices containing electronic circuits, etc.

[0163] In this specification (including the claims), when multiple storage devices (memories) store data, each of the multiple storage devices may store only a portion of the data or the entire data. Furthermore, a configuration in which some of the multiple storage devices store data is also included.

[0164] In this specification (including the claims), terms such as "first," "second," etc., are used merely as a way of distinguishing between two or more elements and are not necessarily intended to impose technical meanings such as temporal, spatial, order, or quantity on the subject. Therefore, for example, references to a first element and a second element do not necessarily mean that only two elements can be employed therein, that the first element must precede the second element, or that the first element must exist for the second element to exist.

[0165] While embodiments of this disclosure have been described in detail above, this disclosure is not limited to the individual embodiments described above. Various additions, modifications, substitutions, and partial deletions are possible, provided that they do not depart from the conceptual idea and spirit of the present invention derived from the claims and their equivalents. For example, where numerical values ​​or mathematical formulas are used in the description of the embodiments described above, these are provided for illustrative purposes only and do not limit the scope of this disclosure. Similarly, the sequence of operations shown in the embodiments is also illustrative and does not limit the scope of this disclosure.

[0166] This application claims priority to Provisional Application No. 63 / 696,501 filed with the United States Patent and Trademark Office on 19 September 2024, Provisional Application No. 63 / 751,941 filed with the United States Patent and Trademark Office on 31 January 2025, and Provisional Application No. 63 / 791,886 filed with the United States Patent and Trademark Office on 21 April 2025, which are incorporated herein by reference to their entire contents.

[0167] 7: Computer 71: Processor 72: Main memory 73: Secondary memory 74: Network interface 75: Device interface 8: Communication network 9A, 9B: External devices

Claims

1. An information processing system comprising one or more memory and one or more processors, wherein the one or more processors acquire a data file to be analyzed, extract multiple data contained in the data file, analyze each of the multiple data using a generative model, and display a graph summarizing the results of the analysis.

2. The information processing system according to claim 1, wherein the data file is a data file specified by the user, and one or more processors display the graph that aggregates the results of the analysis without the user performing any aggregation operations on the results of the analysis.

3. The information processing system according to claim 1, wherein one or more processors input prompts to the generative model and cause the generative model to perform an analysis of each of the plurality of data according to the instructions of the prompts, thereby obtaining the results of the analysis.

4. The information processing system according to claim 3, wherein the prompt includes an instruction to classify the data, and one or more processors obtain the classification results of the plurality of data in accordance with the instructions of the prompt as the result of the analysis, and automatically display the graph that aggregates the classification results.

5. An information processing system comprising one or more memories and one or more processors, wherein the one or more processors acquire a data file to be analyzed, extract a plurality of data contained in the data file, acquire a purpose for analyzing the plurality of data, acquire a plurality of labels to be classified based on the purpose, classify each of the plurality of data into at least one of the plurality of labels, and display the results of the classification of the plurality of data.

6. The information processing system according to claim 5, wherein one or more processors perform sentiment analysis on each of the plurality of data, and display the results of the classification of the plurality of data and the results of the sentiment analysis on the plurality of data.

7. The information processing system according to claim 5, wherein one or more processors display the plurality of data for each label as a result of the classification of the plurality of data.

8. The information processing system according to claim 7, wherein one or more processors display a plurality of buttons corresponding to each of the plurality of labels, and display the data classified under the label corresponding to the button selected from among the plurality of buttons.

9. The information processing system according to claim 5, wherein one or more processors display a graph summarizing the results of the classification of the plurality of data.

10. The information processing system according to any one of claims 1, 2, 3, 4, or 9, wherein the graph is a pie chart.

11. An information processing system comprising one or more memories and one or more processors, wherein the one or more processors generate a prompt including an instruction to propose labels for classifying a plurality of data based on a plurality of data, input the prompt into a generation model and obtain, as a response, a draft of a first label for classifying the plurality of data, receive feedback from a user regarding the obtained draft of the first label, generate a prompt including an instruction to propose labels for classifying the plurality of data at least based on the received feedback, input the prompt into a generation model and obtain, as a response, a draft of a second label for classifying the plurality of data that reflects the feedback from the user.

12. The information processing system according to claim 11, wherein one or more processors divide the plurality of data into a plurality of chunks; for each chunk, generate a first prompt including an instruction to generate feature data that indicates the characteristics of the data contained in the chunk based on the data contained in the chunk; for each chunk, input the first prompt for the chunk to a first generative model and, as a response, obtain the feature data for the chunk; generate a second prompt including an instruction to propose labels for classifying the plurality of data based on the feature data for each chunk; input the second prompt to a second generative model and, as a response, obtain proposed first labels for classifying the plurality of data.

13. The information processing system according to claim 12, wherein the first generative model and the second generative model are the same generative model.

14. The information processing system according to claim 12 or 13, wherein one or more processors generate a third prompt for each chunk, which includes an instruction to generate feature data that describes the characteristics of the data contained in the chunk, based on the data contained in the chunk and the feedback received; input the third prompt for the chunk to the first generative model for each chunk, and in response obtain the feature data for the chunk; generate a fourth prompt, which includes an instruction to propose labels to be assigned to the plurality of data, based on the feature data for each chunk and the feedback received; input the fourth prompt to the second generative model, and in response obtain the second proposal for classifying the plurality of data.

15. The information processing system according to any one of claims 12 to 14, wherein the feature data of the chunk includes at least a summary of the contents of the data contained in the chunk, and a plurality of labels for the data contained in the chunk.

16. The information processing system according to any one of claims 11 to 15, wherein the feedback includes at least a sentence written in natural language by the user, or a choice selected by the user from one or more options.

17. An information processing system comprising one or more memories and one or more processors, wherein the one or more processors acquire multiple data, analyze the multiple data to generate analysis results, add annotations to the analysis results, and display the analysis results with the annotations added.

18. The information processing system according to claim 17, wherein one or more processors generate summaries of the plurality of data as analysis results, add annotations to the summaries, and display the annotated summaries as analysis results to which the annotations have been added.

19. The information processing system according to claim 18, wherein one or more processors generate the summary using a generative model.

20. The information processing system according to claim 18 or 19, wherein one or more processors, upon receiving an operation to select an element constituting the summary to which the annotation has been added, displays a list of data from the plurality of data related to that element.

21. The information processing system according to any one of claims 18 to 20, wherein one or more processors extract key elements from the summary as keyword elements, and annotate the summary by associating the keyword elements extracted from the summary with data from the plurality of data related to the keyword elements.

22. The information processing system according to claim 21, wherein one or more processors extract a plurality of keyword elements from the summary and perform the association for each of the plurality of keyword elements extracted from the summary, thereby adding annotations to the summary.

23. The information processing system according to claim 21 or 22, wherein one or more processors perform at least the extraction of keyword elements or the association of keyword elements with related data using a generative model.

24. The information processing system according to any one of claims 18 to 23, wherein one or more processors annotate the summary by adding the number of data related to an element constituting the summary to that element, and displays the summary with the number of related data added to the element.

25. The information processing system according to any one of claims 18 to 24, wherein one or more processors set a plurality of items for summarizing the plurality of data based on the contents of the plurality of data, generate a summary of the plurality of data for each of the plurality of items, add annotations to the summary of each of the plurality of items, and display the summary of each of the plurality of items to which the annotations have been added.

Images