Information processing system and information processing method

Abstract

To make it easy to create a diary. [Solution] An information processing system comprising: a conversation data acquisition unit that acquires conversation data with a user; an emotion identification unit that analyzes the conversation data to identify the user's emotions; and a diary generation unit that generates the user's diary data based on the conversation data and emotions.

Description

Technical Field

[0001] The present invention relates to an information processing system and an information processing method.

Background Art

[0002] In Patent Document 1, a system for supporting the creation of a diary has been proposed.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, creating a diary is laborious and difficult to continue.

[0005] The present invention has been made in view of such a background, and an object thereof is to provide a technology capable of easily creating a diary.

Means for Solving the Problems

[0006] The main invention of the present invention for solving the above problems is an information processing system, comprising: a conversation data acquisition unit that acquires conversation data with a user; an emotion identification unit that analyzes the conversation data to identify the emotion of the user; and a diary generation unit that generates diary data of the user based on the conversation data and the emotion.

[0007] Regarding other problems disclosed in the present application and their solutions, they will be clarified by the embodiments of the invention and the drawings.

Effects of the Invention

[0008] According to the present invention, a diary can be easily created. [Brief explanation of the drawing]

[0009] [Figure 1] This figure shows an example of the overall configuration of an information processing system. [Figure 2] This figure shows an example of the hardware configuration of management server 2. [Figure 3] This figure shows an example of the software configuration for management server 2. [Figure 4] This diagram illustrates the operation of management server 2. [Modes for carrying out the invention]

[0010] <System Overview> The following describes an information processing system according to one embodiment of the present invention. The information processing system of this embodiment assists users in creating diaries (also called journals, etc.). A diary is a record of daily events as well as the user's subjective feelings (emotions, thoughts, reactions, etc.) to deepen self-understanding, and is different from mere records such as meeting minutes or various logs that leave objective facts and data. Although it is called "days," it does not have to be on a daily basis, but may be for any length of time (which may be fixed or variable), such as hours, half days, weeks, two weeks, or months. It may also be created irregularly rather than regularly. Since diaries express the user's subjective feelings, by analyzing the diary, it is possible to estimate, judge, or diagnose the user's mental state, for example. Users are often unaware of their own subjective feelings, and there are quite a few users who have difficulty keeping diaries. The information processing system of this embodiment assists users in creating diaries through conversation.

[0011] Figure 1 shows an example of the overall configuration of an information processing system. The information processing system in this embodiment includes a management server 2. The management server 2 is connected to the user terminal 1 via a communication network. The communication network is, for example, the internet and is constructed using public telephone networks, mobile phone networks, wireless communication channels, Ethernet (registered trademark), etc.

[0012] User terminal 1 is a computer operated by the user. User terminal 1 can be, for example, a smartphone, a tablet computer, or a personal computer.

[0013] The management server 2 may be a general-purpose computer such as a workstation or personal computer, or it may be logically implemented through cloud computing.

[0014] <Management Server> Figure 2 shows an example of the hardware configuration of the management server 2. Note that the illustrated configuration is just one example, and other configurations are also possible. The management server 2 includes a CPU 201, memory 202, storage device 203, communication interface 204, input device 205, and output device 206. The storage device 203 stores various data and programs, such as a hard disk drive, solid-state drive, or flash memory. The communication interface 204 is an interface for connecting to a communication network, such as an adapter for connecting to Ethernet®, a modem for connecting to a public telephone network, a wireless communication device for wireless communication, or a USB (Universal Serial Bus) connector or RS232C connector for serial communication. The input device 205 is for inputting data, such as a keyboard, mouse, touch panel, button, or microphone. The output device 206 is for outputting data, such as a display, printer, or speaker. Furthermore, each functional unit of the management server 2, as described later, is realized by the CPU 201 reading programs stored in the storage device 203 into memory 202 and executing them, and each storage unit of the management server 2 is realized as part of the storage area provided by memory 202 and storage device 203.

[0015] Figure 3 shows an example of the software configuration of the management server 2. The management server 2 includes a conversation data acquisition unit 211, an emotion identification unit 212, a diary generation unit 213, a conversation data generation unit 214, an emotion history graph generation unit 215, a diary sharing unit 216, a user understanding unit 217, a content recommendation unit 218, a conversation data storage unit 231, and a diary data storage unit 232.

[0016] <Storage section> The conversation data storage unit 231 stores data indicating the conversation with the user (hereinafter referred to as conversation data). Conversation data may include, for example, text data, audio data, or video data. The conversation data storage unit 231 can store conversation data over time. For example, the conversation data storage unit 231 can store conversation data along with time information. This time information may be a timestamp such as the date and time.

[0017] The diary data storage unit 232 stores data representing the user's diary (hereinafter referred to as diary data). The diary data can be, for example, text data. The diary data can also include still image data, audio data, video data, etc. The diary data storage unit 232 can accumulate diary data over time. The diary data storage unit 232 can store diary data in association with information that identifies the user (for example, user ID) and time information. The time information can be a timestamp such as date and time.

[0018] Diary data may consist of, for example, metadata, body text, and a summary. Metadata can include information that identifies the user (user ID), date, day of the week, weather, etc. The body text can include time (time of day when the event occurred (morning, noon, night, etc.)), place (location where the event occurred (home, workplace, store name, etc.)), people (people involved in the event (friends, colleagues, family, etc.)), actions (verb phrases describing the content of the event (ate a meal, attended a meeting, etc.)), emotions (the user's emotions regarding the event (joy, sadness, anger, etc.)), and details (a detailed description of the event, the user's thoughts and opinions, etc.). Details can include what events evoked emotions and how the user felt. In this way, diary data can record the events of the day and the user's emotions, thoughts, and reflections on them in a chronological order.

[0019] In addition to text data, diary data can also include image data and audio data. For example, photos and videos related to each entry can be taken and attached to the diary data. Also, by including audio data in the diary data, the user's raw voice and environmental sounds can be recorded, and a more immersive diary can be created. Furthermore, the diary data can include additional data such as the user's health information and environmental information. For example, by associating information such as the number of steps, heart rate, and sleep time obtained from a wearable device, and location information, temperature, humidity, etc. obtained from a smartphone with the diary data, it becomes possible to analyze the relationship between the user's physical and environmental conditions and emotions. In addition, the diary data can include memos and tags manually input by the user. For example, the user can assign keywords and categories related to the content of the diary, or add free-form comments.

[0020] <Functional unit> The conversation data acquisition unit 211 has a function of acquiring conversation data with the user. The conversation data acquisition unit 211 can receive and acquire conversation data such as text data, audio data, and video data transmitted from the user terminal 1. The conversation data acquisition unit 211 stores the acquired conversation data in the storage device 203 together with time information and accumulates it as conversation data over time.

[0021] <> The emotion identification unit 212 has a function of analyzing conversation data to identify the user's emotions. The emotion identification unit 212 analyzes the conversation data acquired by the conversation data acquisition unit 211 using technologies such as natural language processing and machine learning, and can estimate the user's emotions from text, voice, expressions, etc. The emotion identification unit 212 not only identifies the types of emotions such as joy, anger, sorrow, and happiness, but also quantifies and identifies the degree of each emotion.

[0022] The emotion identification unit 212 can score the emotional meaning of words in each sentence or paragraph by, for example, performing text mining or other techniques on the text data included in conversation data using a pre-prepared emotion dictionary. The emotion dictionary can store emotion labels such as "positive," "negative," and "neutral" for each word, as well as emotion values ​​that represent the degree of emotion in a numerical range such as -1 to +1.

[0023] The emotion identification unit 212 can estimate the speaker's emotions by analyzing features such as pitch, volume, tempo, and timbre of the speech data included in the conversation data. For example, by inputting acoustic features such as fundamental frequency (pitch), loudness, spectrum, and formant into a machine learning model, it can output emotion labels such as "anger," "sadness," and "joy" from the speech, along with the degree of each emotion in a numerical range such as 0 to 1.

[0024] The emotion identification unit 212 can estimate the speaker's emotions by detecting changes in facial expressions and body movements from the video data included in the conversation data. For example, by quantifying the coordinates and movements of feature points such as the degree of eye and mouth opening, eyebrow movements, cheek redness, and gestures, and inputting them into a machine learning model, it can output emotion labels such as "surprise," "disgust," and "fear," along with the degree of each emotion in a numerical range such as 0 to 1, from the video.

[0025] The emotion identification unit 212 can estimate the speaker's emotions and their degree based on features detected from conversation data of multiple modalities, such as text, audio, and video. For example, the emotion identification unit 212 can analyze conversation data of multiple modalities, quantify each feature, and input it into a machine learning model to infer the speaker's emotions from both label and numerical perspectives. For example, the emotion identification unit 212 can obtain more reliable emotion estimation results by determining the emotion labels obtained for each modality by majority vote, or by weighting and averaging the emotion values ​​for each modality.

[0026] The diary generation unit 213 has the function of generating user diary data based on conversation data and emotions. The diary generation unit 213 extracts conversation data corresponding to a predetermined emotion identified by the emotion identification unit 212, where the degree of that emotion is above a predetermined threshold. The diary generation unit 213 generates a prompt to output the extracted conversation data, information indicating the identified emotion, and events estimated from the conversation data in diary format. The diary generation unit 213 can input this prompt into a large-scale language model to generate diary data. The generated diary data is either sent to the user terminal 1 or stored in the storage device 203.

[0027] The diary generation unit 213 can select conversation data to be used to generate diary data based on the type and degree of emotion identified by the emotion identification unit 212.

[0028] For example, the diary generation unit 213 can extract conversation data that has been assigned specific emotion labels such as "joy" or "sadness," and whose emotion value is above a predetermined threshold (e.g., 0.7 or higher), and use it to generate diary data. This allows the user to generate a diary centered around events that were impressive or memorable to them.

[0029] For example, the diary generation unit 213 may prioritize extracting conversation data that has been labeled with negative emotions such as "anger" or "disgust," regardless of the absolute magnitude of the emotion value, and include advice on how to deal with them in the diary data. This makes it possible to generate diaries that support the user's mental health and suggest ways to cope with negative emotions.

[0030] The diary generation unit 213 can highlight conversation data that has been assigned positive emotion labels such as "joy" or "fun," or automatically attach related photos, if the emotion value exceeds a threshold (e.g., 0.8 or higher).

[0031] The diary generation unit 213 can generate diary data with appropriate content and expressions according to the user's emotional state by utilizing the emotion labels and emotion values ​​assigned to the conversation data by the emotion identification unit 212.

[0032] The diary generation unit 213 can use natural language processing techniques to estimate events from conversation data. Specifically, by combining morphological analysis, syntactic analysis, named entity recognition, etc., it can detect words and phrases representing events from conversation data and organize them in chronological order.

[0033] The diary generation unit 213 can, for example, perform morphological analysis on each sentence contained in the conversation data to divide the sentence into words, determine the part of speech (noun, verb, adjective, etc.) of each word, analyze the syntactic structure of each sentence through syntactic analysis, and identify sentence components such as subject, predicate, and object. The diary generation unit 213 can also extract named entities such as names of people, places, organizations, and dates from the conversation data using named entity recognition techniques. For example, it can estimate the date, time, and location of an event using expressions that indicate time and place as clues. The diary generation unit 213 can, for example, integrate the information obtained through the above processes to extract words and phrases that represent events. The diary generation unit 213 can more accurately estimate events by arranging the conversation data chronologically and considering the surrounding context. For example, it can estimate the temporal order of multiple events using time expressions such as "last night" and "this afternoon," and conjunctions such as "then" and "next."

[0034] The diary generation unit 213 can dynamically generate prompts to be given to a large-scale language model in order to generate diary data based on conversation data and sentiment information. Prompts are instructional or illustrative sentences written in natural language and are used to specify the format and content of the sentences that the large-scale language model should output.

[0035] The diary generation unit 213 can generate prompts in the following steps, for example. (1) The extracted events are fitted into standard phrases such as "<time>, <place>, <person> and <action> occurred" to generate sentences that describe the events. (2) Using the emotion labels associated with the event and their corresponding emotion values, generate emotional expressions such as "I was very happy" or "I felt a little sad." (3) Combine the description of the event with the expression of emotion to generate each diary entry in the format of "<Event>. <Expression of emotion> upon experiencing it." (4) Arrange each entry chronologically, add appropriate conjunctions and adverbs, and adjust them to form a coherent sentence. (5) Add standard phrases such as "Diary entry for <date>" and "Thank you for your hard work today" to the beginning and end of your diary. (6) The diary entries generated in the manner described above are compiled into prompts along with instructions for the large-scale language model (such as "Please rewrite the following sentences into natural and fluent diary entries").

[0036] By inputting prompts like the ones described above into the large-scale language model, the diary generation unit 213 can generate diary entries that reflect events extracted from conversation data and the emotional information associated with them.

[0037] Furthermore, the diary generation unit 213 can customize the method of generating prompts to suit the user's preferences and habits. For example, it can provide additional instructions to prompts to include keywords specified by the user or to generate sentences in a specific writing style (polite, informal, dialect, etc.). The diary generation unit 213 can also learn the user's preferred expression patterns by analyzing past diary data and reflect this in the generation of prompts.

[0038] Furthermore, the management server 2 may include a conversation data generation unit 214 that generates conversation data to be sent to the user. The conversation data generation unit 214 has the function of generating conversation data (hereinafter referred to as guided conversation data) that inquires about daily events from the user and their thoughts on those events. The conversation data generation unit 214 can generate guided conversation data by creating a prompt that includes an instruction to generate conversation data inquiring about daily events from the user and their thoughts on those events, and inputting this into a large-scale language model. The prompt may also include conversation data that has been exchanged with the user in the past. The generated guided conversation data is sent to the user terminal 1 and used for conversation with the user.

[0039] The conversation data generation unit 214, when generating conversation data to be sent to the user, can not only ask standardized questions but also generate more effective conversation data according to the user's situation and characteristics. For example, the conversation data generation unit 214 can appropriately select topics and language by considering the user's attribute information such as age, gender, and occupation. By realizing a communication style tailored to the user's attributes, such as using more colloquial expressions for young people and more polite language for the elderly, it is possible to increase affinity with the user.

[0040] The conversation data generation unit 214 can also generate conversation data while considering the user's emotional state, for example. For instance, if the user is expressing negative emotions such as "sadness" or "anger," the unit can provide empathetic words or introduce topics that encourage a change of pace to help manage the user's emotions. Conversely, if the user is expressing positive emotions such as "joy" or "happiness," the unit can support the user's positive feelings by generating conversation data that shares and further enhances those emotions.

[0041] The conversation data generation unit 214 can, for example, analyze the user's behavioral history and hobbies and preferences to provide topics that match the user's interests. For instance, if the conversation data generation unit 214 infers from diary data that the user has recently traveled, it can generate conversations that are tailored to the user's interests by asking questions that encourage reflection on the travel experience or inquiring about future travel plans.

[0042] The conversation data generation unit 214 can also generate conversation data that takes the context of the conversation into account by, for example, analyzing past interactions with the user. For example, if the user mentions a certain topic, the conversation data generation unit 214 can ask questions related to that topic or quote related episodes that have come up in past conversations.

[0043] The conversation data generation unit 214 can, for example, learn the user's linguistic characteristics from the user's conversation history and reflect them in the generation of conversation data. For instance, the conversation data generation unit 214 can analyze the words, expressions, and sentence-ending habits that the user frequently uses and incorporate them into the conversation data to achieve a more user-friendly way of speaking.

[0044] The conversation data generation unit 214 may employ a rule-based method or a generative method using a large-scale language model. In the rule-based method, conversation data can be generated based on pre-prepared templates or IF-THEN rules. In the method using a large-scale language model, conversation data can be generated by providing prompts as described above to the large-scale language model. For example, the conversation data generation unit 214 can take past conversation data with the user, diary data, or user profiles as input and provide prompts to the large-scale language model instructing it to generate conversation data based on these conditions. The large-scale language model can probabilistically generate conversation data in response to the input prompts.

[0045] Furthermore, reinforcement learning techniques can be introduced when generating conversational data using language models. For example, user responses can be fed back into the learning process, and the language model can be optimized to gradually generate better conversational data. For instance, by training the model to reinforce conversational data in which users have shown positive responses and suppress conversational data in which users have shown negative responses, it becomes possible to generate more engaging conversational data for users.

[0046] The diary generation unit 213 can extract the user's thoughts from conversation data and generate diary data based on the extracted thoughts and conversation data, either in addition to or in place of emotions. The process of extracting the user's thoughts from conversation data can use natural language processing or the like. For example, the diary generation unit 213 can generate the user's thoughts by providing a large-scale language model with a prompt that includes conversation data and instructions to extract the user's thoughts from the conversation data.

[0047] <Operation> Figure 4 is a diagram illustrating the operation of the management server 2.

[0048] The management server 2 generates conversational data inquiring about the user's daily events and thoughts about those events (S301), sends and receives the conversational data with the user terminal 1 (S302), retrieves the conversational data sent and received with the user terminal 1 (S303), analyzes the conversational data to identify the user's emotions (S304), generates the user's diary data based on the conversational data and emotions (S305), and registers the generated diary data in the diary data storage unit 232 (S306).

[0049] As described above, the information processing system of this embodiment can generate user diary data from conversations between the system and the user. The diary data will reflect the user's emotions and thoughts.

[0050] Although these embodiments have been described above, they are intended to facilitate understanding of the present invention and are not intended to limit its interpretation. The present invention can be modified and improved without departing from its spirit, and equivalents thereof are also included.

[0051] For example, the processing performed by each functional unit of the management server 2 described above may be executed by any of the functional units. Furthermore, different functional units may be added to perform some of the processing performed by each of the functional units described above. Also, the functional units of the management server 2 may be distributed across multiple computers.

[0052] Furthermore, the information stored in each memory unit of the management server 2 may be stored in any of the memory units. That is, the information stored in the multiple memory units mentioned above may be stored in a single memory unit, or a portion of the information stored in one memory unit may be stored in another memory unit.

[0053] <Example 1>

[0054] In Modification 1, a function can be added to obtain conversation data from multiple users, analyze each user's emotions, and calculate the degree of emotional similarity between users.

[0055] Specifically, the management server 2 collects conversation data from multiple user terminals 1 and analyzes the emotions of each user. It can then use emotion vectors to calculate the degree of similarity in emotions between users.

[0056] An emotion vector is a vector whose elements are the intensity of each emotion category (joy, sadness, anger, etc.). For example, one can consider an emotion vector such as "joy: 0.8, sadness: 0.1, anger: 0.2". Management server 2 analyzes emotions from each user's conversation data and generates an emotion vector.

[0057] Next, the management server 2 calculates the similarity of emotions between users by calculating the cosine similarity of the emotion vectors between users. Cosine similarity is an indicator that shows how closely two vectors point in the same direction, and it takes a value from -1 to 1. The closer the value is to 1, the more similar the directions of the two vectors are.

[0058] Management Server 2 can provide various services based on the calculated emotional similarity between users. For example, it can match users with high emotional similarity and provide a platform for them to interact online, or it can create groups of users with high similarity and provide opportunities for them to discuss common topics.

[0059] Furthermore, the management server 2 can also grasp the overall emotional trends of users by analyzing the degree of emotional similarity among them. For example, if many users' emotions are skewed towards "sadness," the management server 2 can deliver content to brighten the mood of all users or plan events that encourage each other.

[0060] Furthermore, the management server 2 can also analyze patterns of user emotional changes by tracking changes in the similarity of users' emotions. For example, if the similarity of emotions between two users gradually increases, it may indicate that empathy is developing between them. By utilizing such analysis results, it becomes possible to provide detailed support tailored to the changes in users' emotions.

[0061] <Modification 2>

[0062] In the second modification, the diary data can include keywords and summaries extracted from the conversation data. Specifically, the diary generation unit 213 can use natural language processing techniques such as morphological analysis and word importance calculation to extract keywords from the conversation data.

[0063] The diary generation unit 213 performs morphological analysis on each sentence included in the conversation data, for example, and breaks it down into parts of speech such as nouns, verbs, and adjectives. Next, it uses methods such as TF-IDF (Term Frequency-Inverse Document Frequency) to calculate the importance of each word. TF-IDF is an index that evaluates how important a particular word is in a document, and is calculated using the word's frequency of occurrence and the reciprocal of the number of documents in which that word appears. The diary generation unit 213 extracts words with high calculated TF-IDF values ​​as important keywords in the conversation data. The extracted keywords can be recorded as metadata in the diary data or inserted into the main text of the diary.

[0064] The diary generation unit 213 can use summarization techniques such as calculating sentence importance, extracting important sentences, and compressing sentences to generate summaries from conversation data. For example, the diary generation unit 213 calculates the importance of each sentence in the conversation data as the sum of the importance of the words in the sentence, and extracts sentences with high importance. Next, it compresses the extracted important sentences using a language model and rewrites them into more concise expressions. The summaries thus generated can be recorded as metadata for the diary data or inserted at the beginning or end of the diary text.

[0065] The diary generation unit 213 may generate summaries to be included in the diary data by providing a large-scale language model with prompts that include conversation data and instructions to summarize the conversation data.

[0066] <Variation 3>

[0067] Modification 3 adds a function to visualize the user's emotional changes by graphing them over time. This makes it easy to understand the progression of the user's emotions.

[0068] Specifically, the management server 2 may include an emotion history graph generation unit 215. The emotion history graph generation unit 215 has the function of graphing the type and degree of the user's emotions, as identified by the emotion identification unit 212, in chronological order.

[0069] The emotion history graph generation unit 215 retrieves emotion data for each day from the diary data stored in the diary data storage unit 232. The emotion data includes the type of emotion (joy, sadness, anger, etc.) and its degree (a real value between 0 and 1). The emotion history graph generation unit 215 arranges the retrieved emotion data in chronological order and graphs the changes in the degree of each type of emotion.

[0070] For example, the emotion history graph generation unit 215 can generate a line graph with the horizontal axis representing time and the vertical axis representing the degree of emotion. On the graph, lines corresponding to each type of emotion are displayed in different colors, and the height of the line indicates the degree of emotion. This allows users to grasp the changes in their emotions for each type at a glance.

[0071] Furthermore, the emotion history graph generation unit 215 can also use a radar chart to visualize the correlation between different types of emotions. A radar chart is a graph that represents multivariate data on a two-dimensional plane. By plotting the values ​​of each variable on axes radiating from a central point and connecting the plotted points with lines, the relationships between the variables can be represented.

[0072] The emotion history graph generation unit 215 generates a radar chart with each emotion type as an axis and can plot the emotion data for each day on the radar.

[0073] The emotion history graph generation unit 215 can also generate an animation showing the progression of emotions in order to summarize the changes in the user's emotions. In the animation, points indicating the degree of emotion on the graph move over time, or areas showing combinations of emotions on the radar chart change, thereby representing the dynamic changes in the user's emotions.

[0074] The generated emotion history graph and animation are sent to user terminal 1 and presented to the user. The user can visually understand the changes in their emotions.

[0075] <Modification 4>

[0076] In variation 4, a feature can be added that allows users to share their diary data with other users based on their permission. This promotes communication among users and allows them to receive empathy and advice.

[0077] Specifically, the management server 2 may include a diary sharing unit 216. The diary sharing unit 216 has the function of sharing all or part of the diary data with other users in accordance with the user's instructions.

[0078] The diary sharing unit 216 receives a request to share diary data from the user terminal 1. The sharing request includes information to identify the user to whom the data will be shared (user ID, email address, SNS account, etc.) and the scope of the diary data to be shared (specific dates, period, entries, etc.).

[0079] The diary sharing unit 216 sends the specified diary data to the recipient user based on the sharing request. At that time, the diary sharing unit 216 adds the source user's information (user ID, nickname, etc.) to the diary data so that the recipient user can identify the creator of the diary.

[0080] The recipient user can view the diary data received on their own user terminal 1 and add comments and reactions (likes, emojis, etc.). The comments and reactions are then fed back to the original user via the diary sharing unit 216.

[0081] Furthermore, the diary sharing unit 216 can also calculate the degree of emotional empathy among users by analyzing comments and reactions to shared diary data. For example, if multiple users leave similar comments or reactions to a particular diary entry, it can be determined that there is a high degree of emotional empathy among those users.

[0082] The diary sharing section 216 can also match users based on the calculated emotional empathy level. For example, it can connect users with high emotional empathy levels to promote deeper communication. Matched users can understand each other's emotions through their diary data and exchange empathy and advice.

[0083] The diary sharing unit 216 can also discover emotional trends common to many users by analyzing shared diary data. For example, if it is found that many users feel similar emotions about a particular event, it suggests that the event has social significance.

[0084] The diary sharing section 216 can be equipped with functions such as allowing users to finely configure the sharing scope and restricting the users to whom it can be shared according to their level of trust. Furthermore, technical measures such as prohibiting copying and forwarding can be implemented to prevent the secondary dissemination of shared diary data.

[0085] <Modification 5>

[0086] In Modification 5, a function can be added to estimate the user's interests, concerns, and problems from diary data and provide information and advice tailored to them. The management server 2 in Modification 5 may include a user understanding unit 217 and a content recommendation unit 218.

[0087] The user understanding unit 217 has the function of estimating the user's interests, concerns, and worries by analyzing diary data.

[0088] The user understanding unit 217 estimates the user's interests and concerns by analyzing keywords and descriptions included in the diary data. For example, if keywords such as "cooking" and "recipes" appear frequently, it can be estimated that the user is interested in cooking. Also, if negative expressions such as "stress" and "fatigue" are frequently observed, it can be estimated that the user is experiencing some kind of problem.

[0089] The user understanding unit 217 categorizes the estimated interests and concerns by applying them to a pre-defined classification system. For example, interests such as "cooking" and "travel" can be classified into the "hobbies" category, while concerns such as "work stress" and "relationship problems" can be classified into the "mental health" category.

[0090] The content recommendation unit 218 has the function of selecting and providing information and advice suitable for the user based on the estimation results by the user understanding unit 217. The content recommendation unit 218 selects appropriate information and advice according to the interests and concerns estimated and classified by the user understanding unit 217. For example, it can provide new recipes and gourmet information to a user interested in "cooking," or provide advice on stress relief methods and time management techniques to a user troubled by "work stress."

[0091] The Content Recommendation Department 218 can use a variety of methods, such as the following, to provide information and advice. (1) Adding relevant information to diary data: Recommended information and advice will be inserted as annotations in the relevant parts of the diary data. Users will be able to refer to the relevant information when reviewing their diary. (2) Push notifications: Information and advice tailored to the user's interests and concerns are periodically sent to the user's device 1. Users can receive information at the appropriate time in their daily lives. (3) Distribution of email newsletters: Email newsletters will be distributed on themes that match the user's interests, concerns, and problems. Users will be able to receive compiled information and advice on a regular basis. (4) Dialogue with chatbot: Dialogue scenarios with the chatbot are generated according to the user's interests and concerns. Users can interactively obtain information and advice through conversations with the chatbot.

[0092] Furthermore, the content recommendation unit 218 can improve the information and advice it recommends based on user feedback. For example, if a user "likes" or comments that the recommended information was "helpful," the recommendation level of that information will be increased. Conversely, if a user ignores the recommended information or comments that they are "not interested," the recommendation level of that information will be decreased. By analyzing this feedback using technologies such as machine learning, it becomes possible to provide information and advice that is more tailored to the user's preferences.

[0093] The content recommendation unit 218 can also analyze the user's estimated interests and concerns, as determined by the user understanding unit 217, in relation to the user's life events. For example, a user's interests and concerns may change with life events such as marriage, childbirth, finding a job, or changing jobs. By capturing such changes, it becomes possible to provide detailed information and advice tailored to each life stage.

[0094] <Disclosure Items> Furthermore, this disclosure also includes the following configurations. [Item 1] A conversation data acquisition unit that acquires conversation data with the user, An emotion identification unit analyzes the aforementioned conversation data to identify the user's emotions, A diary generation unit generates diary data of the user based on the aforementioned conversation data and emotions, An information processing system characterized by comprising the following features. [Item 2] The information processing system described in item 1, The aforementioned conversation data includes at least two of the following: text data, audio data, or video data. An information processing system characterized by the following. [Item 3] The information processing system described in item 1, The conversation data acquisition unit acquires the conversation data over time in association with time information, The emotion identification unit identifies the emotion and the degree of the emotion, The diary generation unit generates diary data by providing a large-scale language model with a prompt that includes conversation data corresponding to a predetermined emotion, in which the degree of the emotion is above a predetermined threshold, information indicating the emotion, and a statement indicating that the conversation data should be used to identify an event and output in diary format along with the emotion. An information processing system characterized by the following. [Item 4] A conversation data generation unit that generates the conversation data to be sent to the user, comprising the conversation data generation unit that generates guidance conversation data to inquire about the user's daily events and their thoughts on those events, An information processing system characterized by the following. [Item 5] The information processing system described in item 4, The conversation data generation unit generates the guided conversation data by providing a large-scale language model with a prompt that includes an instruction to generate conversation data in which the user is asked about daily events and their thoughts about those events. An information processing system characterized by the following. [Item 6] Steps to obtain conversation data with the user, The steps include analyzing the aforementioned conversation data to identify the user's emotions, A step of generating the user's diary data based on the conversation data and the emotions, An information processing method characterized by a computer executing the following. [Explanation of symbols]

[0095] 1 User terminal 2 Management Server

Claims

1. A conversation data acquisition unit that acquires conversation data with the user, An emotion identification unit analyzes the aforementioned conversation data to identify the user's emotions, A diary generation unit generates diary data of the user based on the aforementioned conversation data and emotions, An information processing system characterized by comprising the following features.

2. The information processing system according to claim 1, The aforementioned conversation data includes at least two of the following: text data, audio data, or video data. An information processing system characterized by the following.

3. The information processing system according to claim 1, The conversation data acquisition unit acquires the conversation data over time in association with time information, The emotion identification unit identifies the emotion and the degree of the emotion, The diary generation unit generates diary data by providing a large-scale language model with a prompt that includes conversation data corresponding to a predetermined emotion, in which the degree of the emotion is above a predetermined threshold, information indicating the emotion, and a statement indicating that the conversation data should be used to identify an event and output in diary format along with the emotion. An information processing system characterized by the following.

4. A conversation data generation unit that generates the conversation data to be sent to the user, comprising the conversation data generation unit that generates guidance conversation data to inquire about the user's daily events and their thoughts on those events, An information processing system characterized by the following.

5. The information processing system according to claim 4, The conversation data generation unit generates the guided conversation data by providing a large-scale language model with a prompt that includes an instruction to generate conversation data in which the user is asked about daily events and their thoughts about those events. An information processing system characterized by the following.

6. Steps to obtain conversation data with the user, The steps include analyzing the aforementioned conversation data to identify the user's emotions, A step of generating the user's diary data based on the conversation data and the emotions, An information processing method characterized by a computer executing the following.

Images