Information processing device and information processing method

Abstract

An information processing device according to one embodiment of the present invention includes: a first acquisition unit that acquires first recommendation information indicating the recommendation of an action to a target user based on individual action data of the target user in a first period; a second acquisition unit that acquires second recommendation information indicating the recommendation of an action to the target user based on action data of a plurality of users in a second period longer than the first period; and an output unit that outputs, to the target user, recommendation information corresponding to one option selected, on the basis of context information indicating a context of an action taken by the user, from among a plurality of options including the first recommendation information and the second recommendation information.

Description

Information processing device and information processing method

[0001] The present invention relates to a technique for an information processing device and an information processing method.

[0002] There are known techniques for outputting recommendations to a target user according to different time periods. For example, Patent Literature 1 discloses an invention that analyzes short-term interests and long-term preferences and outputs recommendation results to the user based on the recommendation level.

[0003] Patent No. 5798022

[0004] The invention described in Patent Document 1 does not output recommendation results by selectively switching between short-term interest analysis and long-term preference analysis based on the situation of the target user.

[0005] In view of the above background, the present invention provides a technology for making appropriate recommendations depending on the situation of a target user.

[0006] One aspect of the present disclosure provides an information processing device having: a first acquisition unit that acquires first recommendation information indicating behavioral recommendations for a target user based on behavioral data of the target user's individual user during a first period; a second acquisition unit that acquires second recommendation information indicating behavioral recommendations for the target user based on behavioral data of multiple users during a second period longer than the first period; and an output unit that outputs to the target user recommendation information corresponding to one option selected from multiple options including the first recommendation information and the second recommendation information based on context information indicating the context of the behavior being taken by the user.

[0007] Yet another aspect of the present disclosure provides an information processing method comprising the steps of: acquiring first recommendation information indicating behavioral recommendations for a target user based on personal behavioral data of the target user during a first period; acquiring second recommendation information indicating behavioral recommendations for the target user based on behavioral data of multiple users during a second period longer than the first period; and outputting to the target user recommendation information corresponding to one option selected from multiple options including the first recommendation information and the second recommendation information based on context information indicating the context of the behavior being taken by the user.

[0008] According to the present invention, it is possible to make appropriate recommendations depending on the situation of the target user.

[0009] A diagram illustrating an example of the system configuration of the information processing system 1. A diagram illustrating an example of the functional configuration of the information processing system 1. A diagram illustrating an example of the hardware configuration of an information processing device 10. A diagram illustrating an example of a flowchart illustrating an example of a method for collecting behavioral data and a method for acquiring recommendation information in the information processing system 1. A diagram illustrating an example of a first recommendation database 1001. A diagram illustrating an example of a second recommendation database 2001. A sequence chart illustrating an example of a method for outputting recommendation information in the information processing system 1. A diagram illustrating an example of a context database 3001. A diagram illustrating an example of a method for selecting recommendations using a machine learning model.

[0010] 1. Configuration FIG. 1 is a diagram illustrating the system configuration of an information processing system 1 according to an embodiment. In this example, the information processing system 1 (or simply referred to as the system) is a system for selecting and providing two types of recommendations based on the context of the target user's behavior. The two types of recommendations are recommendations tailored to at least one of the target user's goals and behavior, and recommendations tailored to the target user's interests and tastes. The "target user" refers to a user to whom the system outputs recommendations. Recommendations tailored to goals or behaviors are short-term recommendations based on individual data. Recommendations tailored to interests and tastes are long-term recommendations based on data from multiple users. The short period (an example of a first period) is a relatively short period (e.g., a period measured in seconds, minutes, hours, days, or weeks). The long period (an example of a second period) is a period longer than the first period (e.g., a period measured in months or years). A recommendation (same as "recommend") is information that represents a recommendation, suggestion, or recommendation made in response to the behavior of a target user. A context (same as "context") is information that represents the behavior, environment, situation, or state of a target user. Users other than the target user are simply referred to as users (or general users or mass users). The information processing system 1 has an information processing device 10, a user terminal 20, and a wearable terminal 90. In this example, the components of the system are connected via a network 9 as shown in FIG. 1. In this example, the network 9 is a computer network such as the Internet.

[0011] The inventors of the present application focused on the relationship between the time axis and user behavior patterns and classified and defined recommendations into the following two types. The first recommendation is a recommendation made to a target user who is acting with a specific (immediate) purpose, and is a recommendation (an example of first recommendation information) based on the target user's personal behavioral data over a short period of time. For example, if the target user has the purpose of "moving" soon, the target user will naturally increase their behavior related to the move, such as visiting a real estate agency or shopping at a furniture and home appliance store. Here, since "moving" is ultimately a personal matter and a short-term goal, it is reasonable to present the first recommendation information based on the target user's personal most recent behavioral data.

[0012] The second recommendation is based on the hobbies and preferences of the target user and is a recommendation based on behavioral data of multiple users over a long period of time (an example of second recommendation information). For example, if the target user is interested in "fashion," the hobbies and preferences of the target user are estimated based on the behavior (such as clothing purchase history) of multiple other users who have similar attributes to the target user. Hobbies and preferences are not temporary, like "moving," but rather are something that the target user continues to be interested in over a long period of time. Furthermore, users who share common attributes such as age, gender, and occupation tend to have similar tendencies toward hobbies and preferences. Therefore, it is reasonable to present recommendations in the second recommendation information by referring to the hobbies and preferences of other users who have attributes related to (for example, similar to) the target user.

[0013] As described above, the first recommendation information and the second recommendation information are recommendations acquired or presented based on different perspectives, and therefore, they must be used appropriately depending on the situation of the target user. For the system to estimate the situation of the target user, it is useful to use context information indicating the behavior of the target user acquired by the user terminal 20 or the wearable terminal 90, etc. Therefore, the information processing system 1 selects whether to output the first recommendation information, the second recommendation information, or no recommendation information, based on the context information acquired from various devices. This allows the information processing system 1 to reflect the optimal recommendation result depending on the situation of the target user. Details of the context information will be described later.

[0014] The information processing device 10 is an information processing device or server device in the information processing system 1. In this example, the information processing device 10 acquires behavioral data of a user from a user terminal 20. The behavioral data is data related to the behavior of the user who uses the user terminal 20. The behavioral data includes, for example, at least one of online purchase history, web search information, outing records based on location information, daily life records based on schedules, installation history of apps, and attribute information based on profiles. The information processing device 10 acquires first recommendation information based on the behavioral data of individual target users acquired from the user terminal 20. Furthermore, the information processing device 10 acquires second recommendation information (not shown) based on behavioral data of multiple users acquired from multiple user terminals 20.

[0015] The information processing device 10 acquires context information indicating the behavior of the target user from the user terminal 20. The context information includes biometric information, environmental information, and movement information. The biometric information is information about the target user's body. Examples of the biometric information include heart rate, number of steps, gender, age, hometown, and occupation. The environmental information is information about the environment in which the target user is located. Examples of the environmental information include time, weather, temperature, humidity, season, wind speed, publicness, noise level, and population density. The movement information is information about the movement status of the target user. The movement information is classified into binary information indicating whether the target user is stationary or moving. Therefore, the movement information includes, for example, location information (current location), means of transportation, speed, acceleration, and movement direction. As described above, the information processing device 10 can estimate the situation of the target user from the acquired context information.

[0016] The information processing device 10 selects recommendation information based on the above-described context information and outputs it to the target user. The selection by the information processing device 10 is made from a plurality of options including (A) first recommendation information and (B) second recommendation information. The options may also include (C) an option of not outputting a recommendation. Various methods for selecting options will be described later.

[0017] The user terminal 20 is a terminal operated by a user. The user terminal 20 includes, for example, a smartphone, a tablet, or a personal computer. The user terminal 20 acquires behavioral data of the user using the terminal and transmits the data to the information processing device 10. The user terminal 20 acquires the user behavioral data using, for example, an app, an API (Application Programming Interface), website tracking, or cookies. This allows the information processing device 10 to obtain the behavioral data necessary for acquiring recommendation information. The user terminal 20 acquires context information of the target user and transmits the information processing device 10. Based on the context information acquired from the user terminal 20, the information processing device 10 acquires an option that corresponds to the user's situation from recommendation information consisting of multiple options. The information processing device 10 outputs the acquired recommendation information to the user terminal 20. The user terminal 20 acquires various types of recommendation information from the information processing device 10 and notifies the user of the information.

[0018] The wearable device 90 is a device that a user wears on at least a part of their body and operates. Examples of the wearable device 90 include a smart watch, smart glasses, a smart ring, or a virtual reality (VR) head-mounted display. The wearable device 90 acquires behavioral data or context information of the user wearing the device and transmits it to the information processing device 10. The behavioral data acquired by the wearable device 90 is used by the information processing device 10 to acquire recommendation information. The user terminal 20 or the wearable device 90 acquires recommendation information selected based on the context information of the target user from the information processing device 10. This allows the user to receive recommendations tailored to their own situation. Note that in the information processing system 1, a user may use only either the user terminal 20 or the wearable device 90.

[0019] 2 is a diagram illustrating an example of the functional configuration of the information processing system 1. In this embodiment, the information processing device 10 has functional blocks (components) including a first acquisition unit 11, a second acquisition unit 12, an output unit 13, an estimation unit 14, a selection unit 15, a storage unit 18, and a control unit 19. In this example, the storage unit 18 stores various types of data including, for example, a database. In this example, the control unit 19 performs various types of control.

[0020] The first acquisition unit 11 acquires first recommendation information indicating behavioral recommendations for the target user based on behavioral data of the target user in a first period. The first acquisition unit 11 acquires the behavioral data of the target user from the user terminal 20 or the wearable device 90.

[0021] The second acquisition unit 12 acquires second recommendation information indicating behavioral recommendations for the target user based on behavioral data of the plurality of users for a second period longer than the first period. The second acquisition unit 12 acquires behavioral data from the plurality of users (general public) who are users of various devices from the plurality of user terminals 20 or the plurality of wearable devices 90, and aggregates the data.

[0022] The output unit 13 outputs to the target user recommendation information corresponding to one option selected from a plurality of options including the first recommendation information and the second recommendation information based on context information indicating the context of the behavior of the target user. Alternatively, the output unit 13 does not output a recommendation. This allows the output unit 13 to perform processing according to the content of the context information.

[0023] If the context information includes biometric information of the target user and the target user's tolerance to recommendations based on the biometric information does not meet a standard, the output unit 13 does not output recommendation information. Tolerance is the degree to which a user's physical or psychological condition allows them to accept recommendations. If tolerance is high, the user is likely to be mentally stable and is more likely to accept (read) recommendations sent to the terminal. Conversely, if tolerance is low, the user is likely to be mentally unstable (busy, panicked, etc.), and there is a high probability that recommendations sent to the terminal will not be used (will not be read). In this case, the output unit 13 does not output recommendation information.

[0024] If the context information includes environmental information indicating the environment in which the target user is located, and the environment indicated by the environmental information satisfies a specific condition, the output unit 13 does not output recommendation information. The condition is an arbitrary setting related to the environmental information that is predetermined by the system. For example, even if outdoor parks, leisure activities such as camping, or restaurants are specifically recommended on a day when the weather (an example of environmental information) is rainy (an example of a condition), the recommendation is likely to be inappropriate for the user (not useful to the user). In this case, the output unit 13 does not output recommendation information.

[0025] When the context information includes movement information indicating the movement state of the target user and the movement information indicates that the target user is moving at a specific speed in a specific location, the output unit 13 outputs first recommendation information. For example, in the above-mentioned example of moving, if the user is moving through a shopping mall (an example of a specific location) with a store selling home appliances and furniture, it can be estimated that the action is related to the user's immediate purpose, i.e., moving. Therefore, if the output unit 13 outputs first recommendation information that is in line with the user's purpose, the recommendation may be easily accepted by the user.

[0026] When the context information includes movement information indicating the movement state of the target user and the movement information indicates that the target user is stationary, the output unit 13 outputs second recommendation information. If the user is stationary in a specific location, it can be inferred that the user simply has nothing to do (is bored), is killing time by looking at their smartphone, or is engaged in a hobby. Therefore, if the output unit 13 outputs second recommendation information related to the user's hobbies and preferences, the recommendation may be more acceptable to the user.

[0027] The estimation unit 14 estimates the situation of the target user based on the context information of the target user acquired from the user terminal 20. The situation of the user includes a physical or psychological state, the environment surrounding the user, or the user's behavior. The estimation unit 14 performs estimation based on a machine learning model including, for example, logistic regression, a support vector machine (SVM), or a neural network.

[0028] The selection unit 15 selects one option from a plurality of options including first recommendation information and second recommendation information based on the user's situation estimated from the context information of the target user. These options include an option of not outputting a recommendation. More specifically, the selection unit 15 determines which recommendation method is effective among the options of (A) the first recommendation information, (B) the second recommendation information, and (C) no recommendation, and selects the more effective option. The selection unit 15 makes the selection based on, for example, a machine learning model including 3-way classification using a graph neural network (GNN), an SVM, or a random forest, or a predetermined rule base.

[0029] FIG. 3 is a diagram illustrating an example of the hardware configuration of the information processing device 10. Physically, the information processing device 10 is configured as a computer including a processor 101, a memory 102, a storage 103, a communication device 104, an input device (not shown), an output device (not shown), and a bus connecting these. Each of these devices operates using power supplied from a battery (not shown). In the following description, the term "device" can be interpreted as a circuit, device, unit, or the like. The hardware configuration of the information processing device 10 may be configured to include one or more of the devices shown in FIG. 3, or may be configured without including some of the devices. Furthermore, the information processing device 10 may be configured by communicating with multiple devices each having a different housing.

[0030] Each function of the information processing device 10 is realized by loading specified software (programs) onto hardware such as the processor 101, memory 102, etc., so that the processor 101 performs calculations, controls communication via the communication device 104, and controls at least one of reading and writing data in the memory 102 and storage 103.

[0031] The processor 101 controls the entire computer by running, for example, an operating system. The processor 101 may be configured as a central processing unit (CPU) including an interface with peripheral devices, a control device, an arithmetic unit, a register, etc. Furthermore, for example, a baseband signal processing unit, a call processing unit, etc. may be realized by the processor 101.

[0032] The processor 101 reads programs (program codes), software modules, data, etc. from at least one of the storage 103 and the communication device 104 into the memory 102 and executes various processes in accordance with these. The programs used are those that cause a computer to execute at least some of the operations described below. The functional blocks of the information processing device 10 may be implemented by a control program stored in the memory 102 and running on the processor 101. Various processes may be executed by one processor 101, or may be executed simultaneously or sequentially by two or more processors 101. The processor 101 may be implemented by one or more chips. The programs may be transmitted to the information processing device 10 via a telecommunications line.

[0033] The memory 102 is a computer-readable recording medium and may be configured by at least one of, for example, a read-only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a random access memory (RAM), etc. The memory 102 may also be called a register, a cache, a main memory (primary storage device), etc. The memory 102 can store executable programs (program codes), software modules, etc. for implementing the method according to this embodiment.

[0034] Storage 103 is a computer-readable recording medium, and may be composed of at least one of, for example, an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disk, a digital versatile disk, a Blu-ray (registered trademark) disk), a smart card, a flash memory (e.g., a card, a stick, a key drive), a floppy (registered trademark) disk, a magnetic strip, etc. Storage 103 may also be called an auxiliary storage device.

[0035] The communication device 104 is hardware (transmission / reception device) for communicating between computers via at least one of a wired network and a wireless network, and is also called, for example, a network device, a network controller, a network card, or a communication module.

[0036] Each device, such as the processor 101 and the memory 102, is connected by a bus for communicating information. The bus may be configured using a single bus, or different buses may be used between each device.

[0037] The information processing device 10 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA), and some or all of the functional blocks may be realized by the hardware. For example, the processor 101 may be implemented using at least one of these pieces of hardware.

[0038] When the processor 101 is executing the above-mentioned program, the processor 101, memory 102, storage 103, and communication device 104 are examples of functional blocks for operating the information processing device 10. The processor 101 is an example of a first acquisition unit 11, a second acquisition unit 12, an estimation unit 14, a selection unit 15, and a control unit 19. At least one of the memory 102 and the storage 103 is an example of a storage unit 18. The communication device 104 is an example of an output unit 13. The configuration of the information processing system 1 has been described above. Next, the operation of the information processing system 1 will be described.

[0039] 4 is a flowchart illustrating a method for collecting behavioral data and a method for acquiring recommendation information in the information processing system 1. In step S1, the information processing device 10 acquires behavioral data of individual target users from the user terminal 20 (wearable terminal 90). The information processing device 10 acquires the behavioral data periodically (periodically). Alternatively, the behavioral data may be acquired at any timing, continuously, or at any frequency.

[0040] In step S2, the information processing device 10 acquires first recommendation information for the target user. The first recommendation information is acquired, for example, by the information processing device 10 generating a recommendation by itself. Alternatively, the information processing device 10 acquires a recommendation generated by an external device different from the information processing device 10 from the external device.

[0041] More specifically, the information processing device 10 reads the target user's immediate purpose from short-term personal behavioral data. For example, if behavioral data indicating daily visits to an electronics retail store or home improvement center is acquired, it can be inferred that the user is planning to "move" soon. Furthermore, if the target user's behavior related to "moving" includes spending more than 30 minutes at different government offices within a week, or increasing the number of times they search for routes starting from the nearest station to their new home (a station that they do not normally use frequently), the possibility of "moving" becomes increasingly likely. It is thought that these behaviors will increase in frequency as the moving date approaches (i.e., people are more likely to take purpose-specific actions closer to the moving date).

[0042] Therefore, the information processing device 10 uses a machine learning model to acquire first recommendation information from short-term individual behavioral data. The machine learning model MR1 (not shown) is configured with a function for accepting input of the target user's behavioral data as an explanatory variable and a function for outputting first recommendation information as a target variable. The machine learning model MR1 can perform the above-mentioned processing, for example, by learning features contained in the behavioral data of multiple past users (or the target user alone) in advance as training data. In the example of "moving," the machine learning model MR1 learns the user's behavioral pattern (style) for the user's purpose by extracting features from the user's behavioral data up to the moving date for multiple past users. As an example, the machine learning model MR1 includes a learning model capable of processing the user's past behavioral data in a time series or for future prediction, such as a recurrent neural network (RNN) or a convolutional long short-term memory (LSTM). The machine learning model MR1 has a function not only to generate first recommendation information but also to calculate an evaluation value corresponding to each recommendation (i.e., to evaluate the first recommendation information). The evaluation value is, for example, a numerical value representing the degree of relationship (relevance) between the content of the recommendation and the purpose of the target user, that is, the degree of certainty. If the degree of certainty is high, the recommendation is likely to have content that is more suited to the purpose of the target user.

[0043] In step S3, the information processing device 10 records the first recommendation information in the database. Here, the first recommendation information will be described.

[0044] FIG. 5 is a diagram illustrating an example of the first recommendation database 1001. In this example, the first recommendation database 1001 includes multiple records. Each record corresponds to one piece of recommendation information. Each record includes a timestamp, a recommendation ID, recommendation content, a category, and an evaluation value. The timestamp is information indicating the date and time when the recommendation was generated. When a predetermined time has elapsed based on each timestamp, the information processing device 10 deletes the recommendation corresponding to that timestamp from the first recommendation database 1001. This makes it possible to delete old recommendations that no longer suit the target user's purpose over time. The recommendation ID is ID information of the first recommendation information that can be uniquely identified for each recommendation. The recommendation content is recommendation data to be output to the target user's terminal. The recommendation content includes data for displaying or notifying the recommendation according to the terminal's UI (User Interface). The types of data include, for example, location information, product information, news information, or a uniform resource locator (URL). The category is a feature that represents a classification of the recommendation content. For example, a category is set based on the user's purpose or behavior over a short period of time. The evaluation value is a feature that is used when the information processing device 10 selects a recommendation for a target user. For example, the information processing device 10 can select one piece of first recommendation information depending on the magnitude of the evaluation value.

[0045] Returning to Fig. 4, in step S4, the information processing device 10 acquires behavioral data of multiple users from multiple terminals (such as user terminals 20 or wearable terminals 90). Note that the multiple users here refer to general users (the general public) who use terminals connected to the information processing device 10 via the network 9. The information processing device 10 periodically acquires behavioral data of the general public.

[0046] In step S5, the information processing device 10 acquires second recommendation information for the target user based on the behavioral data of multiple users. The acquisition of the second recommendation information is performed, for example, by collaborative filtering. Collaborative filtering is a method (a type of machine learning) for estimating the preferences of a target user based on the hobbies and tastes of multiple general users who have attributes related to the target user. Related attributes refer to, for example, a commonality of gender, age, and place of residence, or a combination of these.

[0047] More specifically, the information processing device 10 uses a machine learning model MR2 (not shown) that implements collaborative filtering to identify and analyze multiple users similar to the target user from a large dataset. For example, it is effective to learn patterns from the past choices or behaviors (e.g., product purchase histories or reviews) of multiple users. To this end, the information processing device 10 acquires, for example, customer data from affiliated e-commerce businesses or department stores, or purchasing behavior data from businesses that conduct marketing research. The information processing device 10 can utilize this data to generate second recommendation information. Furthermore, when generating an evaluation value for each recommendation, a method is used that calculates the evaluation value based on the degree of match (match rate) between the attributes of the target user and the attributes of a typical user group that has the hobbies and preferences indicated by the recommendation. Note that the calculation of the evaluation value is not limited to this method and any method may be used.

[0048] In step S6, the information processing device 10 records the second recommendation information in the database. Here, the second recommendation information will be described.

[0049] FIG. 6 is a diagram illustrating an example of the second recommendation database 2001. In this example, the second recommendation database 2001 includes multiple records. Each record corresponds to one piece of recommendation information. Each record includes a recommendation ID, recommendation content, category, and rating value. The recommendation ID, recommendation content, category, and rating value are data common to the data types included in the first recommendation database 1001. The recommendation content records recommendations that are in line with the hobbies and preferences of the target user and are acquired based on behavioral data of the general public over a long period of time. The recommendation content includes, for example, recommendations based on data acquired over a certain period (long period) without identifying individuals. When selecting recommendations for the target user, the information processing device 10 selects one piece of second recommendation information, for example, depending on the category or the magnitude of the rating value. Each piece of second recommendation information may also be assigned a timestamp indicating the date and time when the recommendation was created, and the oldest information may be deleted first.

[0050] As a result, the information processing device 10 can generate first recommendation information and second recommendation information for the target user by using the above-described machine learning. Each of the first recommendation information and the second recommendation information is counted as one option per recommendation, and is used when the information processing device 10 selects a recommendation.

[0051] 7 is a sequence chart illustrating a method for outputting recommendation information in the information processing system 1. In step S101, the user terminal 20 (wearable terminal 90) acquires context information from the target user. The user terminal 20 acquires in advance from the target user consent to acquire the context of the behavior of the target user as data.

[0052] In step S102, the information processing device 10 acquires the context data of the target user acquired by the user terminal 20. The information processing device 10 records the context data acquired from the user terminal 20 in a database. Here, the database for managing the context will be described.

[0053] FIG. 8 is a diagram illustrating an example of the context database 3001. In this example, the context database 3001 includes multiple records. Each record corresponds to a time when the user terminal 20 acquired context data from the target user. Each record includes a timestamp, biometric information, environmental information, and movement information. The timestamp is time information, particularly date information (year, month, day) and time information (hour, minute, second), when the user terminal 20 acquired the context data. The biometric information, environmental information, and movement information are all examples of context information of the target user. The information processing device 10 identifies the type of context data according to the nature of the data acquired from the user terminal 20.

[0054] Returning to Fig. 7, in step S103, the information processing device 10 selects a recommendation based on the context information. The recommendation selection process by the information processing device 10 is performed using, for example, a machine learning model. Here, the machine learning model will be described.

[0055] 9 is a diagram illustrating a recommendation selection method using a machine learning model. This diagram shows an example of a machine learning model that estimates the situation of a target user based on the target user's context information acquired from the user terminal 20 and selects recommendations according to the situation. The information processing device 10 is pre-implemented with machine learning models M1 and M2 that can select recommendations according to the context information. The roles (functions) of each model will be described.

[0056] The machine learning model M1 is configured with a function to accept input of context information, such as biometric information, environmental information, or movement information, as an explanatory variable, and a function to select and output recommendation information options, such as (A) first recommendation information, (B) second recommendation information, or (C) no recommendation, as a target variable. The options A, B, and C do not specify the content of the recommendation. Therefore, the machine learning model M2 is used to narrow down personalized (customized) recommendations for the target user.

[0057] The machine learning model M2 has a function of accepting input of the above-mentioned context information and the option (A, B, or C) selected by the machine learning model M1 as explanatory variables. The machine learning model M2 also has a function of selecting from a database, as a target variable, specific recommendation content corresponding to the option, and outputting it. For example, when option A is output by the machine learning model M1, the machine learning model M2 selects and outputs one piece of recommendation information from each record in the first recommendation database 1001 according to the context information. Alternatively, when option B is output by the machine learning model M1, the machine learning model M2 selects and outputs one piece of recommendation information from each record in the second recommendation database 2001 according to the context information. Note that the machine learning models M1 and M2 may be functionally integrated, or the machine learning model M2 may be omitted.

[0058] The machine learning models M1 and M2 can classify (find patterns in) unknown input data by learning in advance, for example, from a dataset containing correct answer data (correct answer rate) corresponding to past input data as training data. The machine learning model M1 learns using a dataset that answers one of the options A, B, or C based on past context information. The machine learning model M2 learns using a dataset that identifies specific recommendation content from past context information and an option (A, B, or C) linked to that context information.

[0059] A specific example of training data is data in which labels classified as (A), (B), and (C) are assigned (labeled) to a dataset (input data) of past context information acquired from multiple users. Furthermore, for example, when labeling (B), an annotation of the category or genre of the recommendation (e.g., food and drink, travel) may be added. Such labeling and annotation of training data is performed based not only on the correct answer labels (correct answer data) of the multiple-choice questions, but also on the addition of features or related information of the input data. This makes it easier for the machine learning model M1 and the machine learning model M2 to understand the relationship between the datasets as multiple-choice questions, enabling more accurate predictions. In other words, it is possible to improve the performance and accuracy of the machine learning model M1 and the machine learning model M2.

[0060] For example, consider a case in which the above-described machine learning models M1 and M2 acquire context information of a target user as of May 4th in the context database 3001. At 12:00:00, the target user's state is a heart rate of 60 or less (resting state), the step count has not changed for 30 minutes (constant at 4,000 steps), and the current location is a cafe, so it is estimated that the target user is taking a break (having lunch). Generally, a target user taking a break has a high tolerance for recommendations, so the information processing device 10 can determine that there is room for recommendation selection. Here, the tolerance is calculated using various machine learning models. A machine learning model M3 (not shown) that calculates the tolerance has a function of accepting input of the target user's context information as an explanatory variable. Unlike the machine learning models M1 and M2 that select recommendation information, this machine learning model M3 may be a machine learning model specialized for calculating tolerance.

[0061] Returning to the selection of recommendation information, in this case, the target user is likely to be in a relaxed mood and enjoying his or her hobbies, so it is possible to determine that it is preferable to output second recommendation information (fashion, clothing, etc.) that is in line with the target user's hobbies and preferences. The machine learning models M1 and M2 select recommendations from the database in which the second recommendation information is stored that share common characteristics, such as recommendation content or category, or that have high evaluation values. This series of processes is performed based on the recommendation selection function of the machine learning models M1 and M2. Note that, among the functions of the machine learning models M1 and M2, the estimation of the target user's situation based on context information and the recommendation selection based on the estimation of the target user's situation may be implemented using the same or different machine learning models. As described above, the information processing device 10 can select options corresponding to recommendation information based on the target user's context information. Note that the selection of recommendation information is performed based on various data in the first recommendation database 1001 or the second recommendation database 2001.

[0062] Returning to FIG. 7 , in step S104, the information processing device 10 outputs recommendation information corresponding to the option to the user terminal 20. For example, in the above example, the information processing device 10 can output a recommendation related to fashion to the target user as the second recommendation information. Note that if the machine learning model M1 selects option (C) "No recommendation," the information processing device 10 does not output a recommendation to the user terminal 20 and ends the processing at that point.

[0063] In step S105, the user terminal 20 notifies or proposes a recommendation to the target user. The user terminal 20 notifies the target user of the recommendation by, for example, using a notification function of a specific app. Alternatively, the layout may be changed so that the recommendation catches the target user's eye when the target user launches the app. The user terminal 20 may notify the target user of the recommendation in any manner.

[0064] As a result, the information processing system 1 can make recommendations at an appropriate timing and with appropriate accuracy (precision) based on the situation of the target user.

[0065] 3. Modifications The present invention is not limited to the above-described embodiment, and various modifications are possible. Some modifications will be described below. Two or more of the following features may be combined and applied.

[0066] (1) Information Processing System 1 The hardware configuration and network configuration of the information processing system 1 are not limited to those exemplified in the embodiment. The information processing system 1 may have any hardware configuration and network configuration as long as the required functions can be realized. For example, multiple physical devices may cooperate to function as the information processing system 1. For example, at least a portion of the user terminal 20 may be implemented in a wearable terminal 90 and configured to be able to obtain recommendations from the information processing device 10. Furthermore, the entities, configuration, or system structure shown in FIG. 1 are merely an example and merely represent an overview of the system. Therefore, the information processing system 1 may be a network system in which, for example, the information processing device 10 and different devices, such as an external device for outputting recommendations based on context information, are interconnected.

[0067] (2) Information Processing Device 10 Some of the functions of the information processing device 10 may be implemented in another server. This server may be, for example, a physical server or a virtual server (including a so-called cloud). Furthermore, the correspondence between functional elements and hardware is not limited to that illustrated in the embodiment. For example, in the embodiment, at least some of the functions described as being implemented in the information processing device 10 may be implemented in another device or system. Conversely, at least some of the functions described as being implemented in another device or system may be implemented in the information processing device 10. In this example, at least some of the functions of the information processing device 10 may be implemented in, for example, a user terminal 20 (wearable terminal 90). In this case, a database or machine learning model managed by the information processing device 10 may be implemented in the user terminal 20. This enables the user terminal 20 to acquire recommendation information or select recommendations based on context information.

[0068] (3) User Terminal 20 The user terminal 20 is not limited to the one exemplified in the embodiment. The user terminal 20 may perform the above-described processing using any display screen, input device, or various UIs. The user terminal 20 may be equipped with a function for notifying the target user of recommendations output from the information processing device 10 via an application (or a web browser) within the terminal. The user terminal 20 may be equipped with various sensors and may use these sensors to acquire (output to the information processing device 10) behavioral data or context information of the target user. The sensors include, for example, an acceleration or angular velocity sensor, a Global Navigation Satellite System (GNSS), a Light Detection and Ranging (LiDAR), or an Inertial Measurement Unit (IMU).

[0069] (4) Wearable Terminal 90 The wearable terminal 90 is not limited to the one exemplified in the embodiment. The wearable terminal 90 may have any hardware configuration as long as it can realize the required functions or operations. The wearable terminal 90 may have at least some of the functions of the user terminal 20. The wearable terminal 90 may be used only to acquire context information of the target user. The wearable terminal 90 may have the various sensors described above.

[0070] (5) Method for Collecting Behavioral Data and Method for Acquiring Recommendation Information The flowchart shown in FIG. 4 merely illustrates an example of the operation, and the operation of the information processing system 1 is not limited thereto. Some of the illustrated operations may be changed or omitted, the order may be changed, or new operations may be added. In step S2 (or step S5), the first (or second) recommendation information may be acquired by any method, and various machine learning models may be used. In this case, the machine learning model is configured with a function for accepting input of behavioral data of target users or general users as explanatory variables, and a function for generating and outputting various types of recommendation information as objective variables. Examples of machine learning models include collaborative filtering, content-based filtering, deep learning-based recommendation systems using neural networks, and matrix factorization. As another variation, the entity (device or system) that generates the first recommendation information and the second recommendation information is not limited to the information processing device 10, and may be generated by another device within the information processing system 1 or another system separate from the information processing system 1. The entities that generate the first recommendation information and the second recommendation information may be different from each other. The timing and frequency at which the information processing device 10 acquires the first recommendation information and the second recommendation information may be determined individually.

[0071] (6) Method for Outputting Recommendation Information The sequence chart shown in FIG. 7 merely illustrates an example of the operation, and the operation of the information processing system 1 is not limited thereto. Some of the illustrated operations may be changed or omitted, the order may be changed, or new operations may be added. In step S103, for example, processing by the machine learning model M2 may be omitted when selecting recommendation information. In this case, the information processing device 10 may select the recommendation information with the highest evaluation value from among multiple pieces of recommendation information corresponding to the option (A, B, or C) selected by the machine learning model M1, or may randomly select recommendation information. More specifically, when option A is output by the machine learning model M1, the information processing device 10 selects the recommendation information with the highest evaluation value from the first recommendation database 1001. Alternatively, when option B is output by the machine learning model M1, the information processing device 10 randomly selects one piece of recommendation information from the second recommendation database 2001.

[0072] Furthermore, the tolerance calculated by the machine learning model M3 need not be a numerical value, but may be, for example, a rank (a five-point scale). The explanatory variable (input) of the machine learning model M3 may be only one of biometric information, environmental information, or movement information. As an example of context information, the operation frequency (usage frequency) of the user terminal 20 (wearable terminal 90) may be accepted as an explanatory variable of the machine learning model M3. Note that the information processing device 10 does not need to use tolerance in the first place to estimate the state of the target user or select recommended information.

[0073] Furthermore, in step S103, various methods may be used to select recommendations based on context information. For example, a rule-based method may be used instead of (or in addition to) the above-described machine learning model. Specifically, the information processing device 10 may predetermine rules based on specific conditions. The specific conditions may be, for example, exclusion conditions set for the evaluation values ​​of the recommendation information. The exclusion conditions are conditions for excluding a specific recommendation from the selection of the information processing device 10 when the evaluation values ​​of the first recommendation information and the second recommendation information are equal to or lower than a threshold, or when the recommendation itself is determined to be inappropriate based on context information (such as late-night time, weather (rain, storm, etc.), season, women-only area, etc.) (i.e., the option "Do not recommend" is automatically selected). The threshold value of the evaluation value may be set based on the context information, for example. Various patterns will now be described.

[0074] The first pattern is when both the first recommendation information and the second recommendation information meet the exclusion condition. In this case, the information processing device 10 does not output a recommendation. The second pattern is when either the first recommendation information or the second recommendation information meets the exclusion condition. In this case, the information processing device 10 selects one of the options corresponding to the remaining recommendation information (and recommends it to the target user). The third pattern is when neither the first recommendation information nor the second recommendation information meets the exclusion condition. In this case, the information processing device 10 selects, from the two remaining recommendations, the recommendation that is more relevant to the context of the target user or the recommendation that has an option with a higher evaluation value.

[0075] (7) Biometric Information Biometric information is not limited to the examples given in the embodiments. For example, the biometric information may be information acquired from the wearable device 90, body temperature, blood oxygen concentration, pupil size, or health data. At least a portion of the operation of the information processing system 1 may be a process that uses biometric information. For example, the information processing device 10 may estimate or calculate the target user's tolerance for recommendations based on the biometric information. In this case, the information processing device 10 may, for example, set a reference value (threshold value) for the tolerance and select recommendations based on the reference value.

[0076] (8) Environmental Information The environmental information is not limited to the examples shown in the embodiments. The environmental information may be, for example, information about companions, acquaintances, friends, family, or pets surrounding the target user based on information acquired from multiple nearby user terminals 20 or multiple wearable terminals 90. At least a portion of the operation of the information processing system 1 may utilize environmental information. For example, the information processing device 10 may estimate the environment in which the target user is located based on the environmental information. In this case, whether or not to output a recommendation may be determined based on whether the environment satisfies certain conditions. For example, for the time period from 9:00 to 18:00 on weekdays (day shift), since this is during working hours, the information processing device 10 may estimate that the target user has a low tolerance for recommendations and refrain from outputting recommendations. On the other hand, for the nighttime time period from 18:00 to 23:00 on weekdays, since it is unlikely that the target user will start activities at this time, the information processing device 10 may determine that the second recommendation information is more effective than the first recommendation information.

[0077] (9) Travel Information Travel information is not limited to the examples shown in the embodiments. Travel information may be, for example, the target user's travel history, travel route, transfer of transportation means, or stay time. At least a portion of the operation of the information processing system 1 may be a process that uses travel information. For example, the information processing device 10 determines that the recommendation tolerance is low when the target user is walking or driving, and determines that the recommendation tolerance is high when the target user is stationary. Alternatively, if the target user is stationary, such as at home or on a train, the information processing device 10 can determine that the second recommendation information is effective because the target user is likely to have time on their hands. The information processing device 10 selects a recommendation from options that include the second recommendation information. Note that the recommendation tolerance is set for any context information (biometric information, environmental information, travel information) and is used when selecting a recommendation for the target user.

[0078] (10) Database The databases (or the data itself) of the information processing system 1 shown in FIGS. 5, 6, and 8 are not limited to those illustrated in the embodiments. In this example, any data may be registered in the database. For example, the data recorded in the first recommendation database 1001 may be a recommendation priority (feature) other than the evaluation value, such as the order of shortest recommendation period. The data recorded in the second recommendation database 2001 may be a feature based on attribute information of the target user. The data recorded in the context database 3001 may be data of users other than the target user (companions close to the target user). Furthermore, the layout of the database is not limited to that shown in the drawings, and data may be managed in any layout. Note that the data output by the information processing device 10 to each device may be any data registered in the database.

[0079] (11) Machine Learning The machine learning model or various machine learning (AI) functions of the information processing system 1 shown in FIG. 9 are not limited to those exemplified in the embodiment. Any machine learning model may be adopted as long as it can realize the required function or operation. The machine learning model may be configured with any algorithm. For example, the target user's situation may be estimated using machine learning models corresponding to biometric information, environmental information, and movement information, respectively, among the context information. Alternatively, all context information may be input into a single integrated machine learning model.

[0080] (12) Recommendations Recommendations are not limited to those exemplified in the embodiments. Recommendations may include options that include perspectives other than options based on different time periods, such as short-term or long-term. For example, recommendations may be based on a medium period between short-term and long-term. The configurations or operations described in the embodiments may be modified as appropriate depending on the type of recommendation.

[0081] (13) Context The context is not limited to the example illustrated in the embodiment. The context may be in any form as long as it is related to the behavior of the target user. The context may include, for example, information about sleep, travel, work, etc. The configuration or operation described in the embodiment may be appropriately changed depending on the type of context.

[0082] (14) Others The various programs executed by the processor 101 may be provided by downloading via a network such as the Internet, or may be provided in a state recorded on a computer-readable non-transitory recording medium such as a DVD-ROM. Each processor may be, for example, a CPU, an MPU (Micro Processing Unit), or a GPU (Graphics Processing Unit).

[0083] The block diagrams used to explain the above embodiments show functional blocks. These functional blocks (components) are realized by any combination of hardware and / or software. Furthermore, the method for realizing each functional block is not particularly limited. That is, each functional block may be realized using a single device that is physically or logically coupled, or may be realized using two or more physically or logically separated devices that are directly or indirectly connected (e.g., wired, wireless, etc.) and these multiple devices. The functional block may also be realized by combining software with the single device or multiple devices.

[0084] Functions include, but are not limited to, judgment, determination, judgment, calculation, computation, processing, derivation, investigation, search, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, consideration, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating, mapping, and assignment. For example, a functional block (component) that performs transmission is called a transmitting unit or transmitter. As mentioned above, there are no particular limitations on how these functions are implemented.

[0085] For example, the information processing device 10 according to an embodiment of the present disclosure may function as a computer that performs the processing of the present disclosure.

[0086] Each aspect or embodiment described in the present disclosure may be applied to at least one of systems using LTE (Long Term Evolution), LTE-Advanced (LTE-A), SUPER 3G, IMT-Advanced, 4G (4th generation mobile communication system), 5G (5th generation mobile communication system), FRA (Future Radio Access), NR (New Radio), W-CDMA (registered trademark), GSM (registered trademark), CDMA2000, UMB (Ultra Mobile Broadband), IEEE 802.11 (Wi-Fi (registered trademark)), IEEE 802.16 (WiMAX (registered trademark)), IEEE 802.20, UWB (Ultra-Wide Band), Bluetooth (registered trademark), or other suitable systems, and next-generation systems enhanced based on these. Furthermore, a combination of multiple systems (e.g., a combination of at least one of LTE and LTE-A with 5G) may also be applied.

[0087] The order of the procedures, sequences, flowcharts, etc. of each aspect or embodiment described in this disclosure may be changed unless it is inconsistent. For example, the methods described in this disclosure present elements of various steps using an example order, and are not limited to the particular order presented.

[0088] Input and output information may be stored in a specific location (for example, memory) or may be managed using a management table. Input and output information may be overwritten, updated, or added to. Output information may be deleted. Input information may be sent to another device.

[0089] The determination may be made based on a value represented by one bit (0 or 1), a Boolean value (true or false), or a numerical comparison (e.g., comparison with a predetermined value).

[0090] Although the present disclosure has been described in detail above, it is clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the spirit and scope of the present disclosure as defined by the claims. Therefore, the description of the present disclosure is intended to be illustrative and does not have any limiting meaning on the present disclosure.

[0091] Software, whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise, should be construed broadly to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, threads of execution, procedures, functions, etc. Additionally, software, instructions, information, etc. may be transmitted or received over a transmission medium. For example, if software is transmitted from a website, server, or other remote source using wired technologies (such as coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL)), and / or wireless technologies (such as infrared, microwave), then such wired and / or wireless technologies are included within the definition of a transmission medium.

[0092] The information, signals, etc. described in this disclosure may be represented using any of a variety of different technologies. For example, data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof. Note that terms described in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings.

[0093] Furthermore, the information, parameters, etc. described in this disclosure may be expressed using absolute values, may be expressed using relative values ​​from a predetermined value, or may be expressed using other corresponding information.

[0094] As used in this disclosure, the phrase "based on" does not mean "based only on," unless expressly stated otherwise. In other words, the phrase "based on" means both "based only on" and "based at least on."

[0095] As used in this disclosure, any reference to an element using a designation such as "first," "second," etc. does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient method of distinguishing between two or more elements. Thus, a reference to a first and a second element does not imply that only two elements may be employed or that the first element must in some way precede the second element.

[0096] The "unit" in the configuration of each of the above devices may be replaced with "means," "circuit," "device," or the like.

[0097] When the terms "include," "including," and variations thereof are used in this disclosure, these terms are intended to be inclusive, similar to the term "comprising." Furthermore, when the term "or" is used in this disclosure, it is not intended to be an exclusive or.

[0098] In this disclosure, where articles are added by translation, such as a, an, and the in English, the disclosure may include that the nouns following these articles are in the plural form.

[0099] In the present disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "coupled" may also be interpreted in the same way as "different."

[0100] 1...information processing system, 10...information processing device, 11...first acquisition unit, 12...second acquisition unit, 13...output unit, 14...estimation unit, 15...selection unit, 18...storage unit, 19...control unit, 101...processor, 102...memory, 103...storage, 104...communication device, 20...user terminal, 90...wearable terminal, 9...network, 1001...first recommendation database, 2001...second recommendation database, 3001...context database, MR1, MR2, M1, M2, M3...machine learning model

Claims

1. An information processing device having: a first acquisition unit that acquires first recommendation information indicating behavioral recommendations for a target user based on the target user's individual behavioral data during a first period; a second acquisition unit that acquires second recommendation information indicating behavioral recommendations for the target user based on behavioral data of multiple users during a second period longer than the first period; and an output unit that outputs to the target user recommendation information corresponding to one option selected from multiple options including the first recommendation information and the second recommendation information based on context information indicating the context of the behavior being taken by the user.

2. The information processing device according to claim 1, wherein the plurality of options includes an option of not outputting a recommendation.

3. The information processing device according to claim 1, wherein the context information includes biometric information of the target user.

4. The information processing device according to claim 3, wherein the output unit does not output the recommendation information if the target user's tolerance for the recommendation based on the biometric information does not satisfy a standard.

5. The information processing device according to claim 1, wherein the context information includes environmental information indicating an environment in which the target user is placed.

6. The information processing device according to claim 5, wherein the output unit does not output the recommendation information when the environment indicated by the environmental information satisfies a specific condition.

7. The information processing device according to claim 1, wherein the context information includes movement information indicating a movement state of the target user.

8. The information processing device according to claim 7, wherein the output unit outputs the first recommendation information when the movement information indicates that the target user is moving in a specific location at a specific speed.

9. The information processing device according to claim 7, wherein the output unit outputs the second recommendation information when the movement information indicates that the target user is stationary.

10. An information processing method comprising the steps of: acquiring first recommendation information indicating behavioral recommendations for a target user based on the target user's personal behavioral data during a first period; acquiring second recommendation information indicating behavioral recommendations for the target user based on behavioral data of multiple users during a second period longer than the first period; and outputting recommendation information to the target user corresponding to one option selected from multiple options including the first recommendation information and the second recommendation information based on context information indicating the context of the behavior being taken by the user.