Dialogue forecasting device, dialogue forecasting system, dialogue forecasting method, and dialogue forecasting program

Abstract

The present invention provides a dialogue forecasting device, a dialogue forecasting system, a dialogue forecasting method, and a dialogue forecasting program that facilitate communication between users even when an explicit request for dialogue is not made. [Solution] The management device 7 of the dialogue forecasting system 1 comprises a collection unit 9, a prediction unit 10, and an output unit 11. The collection unit 9 collects behavioral information when user 2a desires to have a dialogue with user 2b. This behavioral information represents actions or states of user 2a that do not involve an explicit request for dialogue. Based on the behavioral information collected by the collection unit 9, the prediction unit 10 predicts the likelihood of a dialogue occurring between user 2a and user 2b. The output unit 11 notifies user 2b of the likelihood of a dialogue predicted by the prediction unit 10.

Description

Technical Field

[0001] The present disclosure relates to an interactive forecasting device, an interactive forecasting system, an interactive forecasting method, and an interactive forecasting program.

Background Art

[0002] Patent Document 1 discloses an example of a communication system. In this system, user A who desires to interact with user B, who is the interaction partner, sends a message indicating the desire for interaction to user B. User B who receives this message checks the situation of user A and contacts user A.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in a communication system such as that of Patent Document 1, a user who desires an interaction needs to explicitly request the interaction from the interaction partner, such as by sending a message with the subject line "Desire to Contact". On the other hand, for example, due to reasons such as the content or urgency of the interaction, or the relationship with the interaction partner, the request for interaction may not be explicitly made. At this time, there is a possibility that the communication between users will be insufficient without the necessary interaction occurring.

[0005] The present disclosure relates to the solution of such problems. The present disclosure provides an interactive forecasting device, an interactive forecasting system, an interactive forecasting method, and an interactive forecasting program that make the communication between users smoother even when an explicit request for interaction is not made.

Means for Solving the Problems

[0006] The dialogue prediction device according to this disclosure includes: a collection unit that collects behavioral information representing the actions or states of a first user that do not involve an explicit request for dialogue when the first user desires to engage in dialogue with a second user; a prediction unit that predicts the likelihood of a dialogue occurring between the first user and the second user based on the behavioral information of the first user collected by the collection unit; and an output unit that notifies the second user of the likelihood of a dialogue predicted by the prediction unit. The prediction unit scores the first user's manner information based on a predetermined criterion for each type of manner information, and calculates the likelihood of interaction by adding the scores multiplied by a predetermined weighting coefficient for each type of manner information. .

[0007] The dialogue forecasting system relating to this disclosure comprises: a portable device possessed by a first user; a collection unit that collects behavioral information through the portable device representing the first user's actions or states that do not involve an explicit request for dialogue when the first user desires to engage in dialogue with a second user; a prediction unit that predicts the likelihood of a dialogue occurring between the first user and the second user based on the behavioral information of the first user collected by the collection unit; and an output unit that notifies the second user of the likelihood of a dialogue predicted by the prediction unit. The prediction unit scores the first user's manner information based on a predetermined criterion for each type of manner information, and calculates the likelihood of interaction by adding the scores multiplied by a predetermined weighting coefficient for each type of manner information. .

[0008] The dialogue forecasting method relating to this disclosure is a method in which a computer collects behavioral information representing the actions or states of a first user that do not involve an explicit request for dialogue when the first user desires to interact with a second user; predicts the likelihood of a dialogue occurring between the first user and the second user based on the collected behavioral information of the first user; and notifies the second user of the predicted likelihood of a dialogue. The likelihood of interaction is calculated by scoring the first user's manner information based on a predetermined standard for each type of manner information, multiplying the score by a predetermined weighting coefficient for each type of manner information, and adding the scores together. .

[0009] The dialogue prediction program relating to this disclosure causes a computer to perform the following actions: collect behavioral information representing the actions or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user; predict the likelihood of a dialogue occurring between the first user and the second user based on the collected behavioral information of the first user; and notify the second user of the predicted likelihood of a dialogue. The likelihood of interaction is calculated by scoring the first user's aspect information based on pre-set criteria for each type of aspect information, multiplying the score by a pre-set weighting coefficient for each type of aspect information, and adding the scores together. . [Effects of the Invention]

[0010] The dialogue forecasting device, dialogue forecasting system, dialogue forecasting method, or dialogue forecasting program related to this disclosure facilitates smoother communication between users even when no explicit request for dialogue is made. [Brief explanation of the drawing]

[0011] [Figure 1] This is a diagram showing the configuration of the dialogue prediction system according to Embodiment 1. [Figure 2] This flowchart shows an example of processing in the dialogue prediction system according to Embodiment 1. [Figure 3] This flowchart shows an example of processing in the dialogue prediction system according to Embodiment 1. [Figure 4] This is a hardware configuration diagram of the main components of the dialogue prediction system according to Embodiment 1. [Modes for carrying out the invention]

[0012] The embodiments for carrying out the subject matter of this disclosure will be described with reference to the attached drawings. In each drawing, the same or corresponding parts are denoted by the same reference numerals, and redundant explanations are simplified or omitted as appropriate. However, the subject matter of this disclosure is not limited to the following embodiments, and any modification of any component of the embodiments or omission of any component of the embodiments is possible without departing from the spirit of this disclosure.

[0013] Embodiment 1. Figure 1 is a diagram showing the configuration of the dialogue forecasting system 1 according to Embodiment 1.

[0014] Dialogue forecasting system 1 is a system that notifies users of the results of its prediction of the likelihood of dialogue occurring between human users. Dialogue forecasting system 1 is used in organizations, such as companies, unions, or other groups. An organization may be a part of a larger organization, such as a branch or department of a company, or it may include multiple smaller organizations. An organization may include multiple members, each of whom is a person. In this case, dialogue forecasting system 1 treats each individual member of the organization as a user and predicts the occurrence of dialogue between users. Members of an organization may be, for example, employees of a company. Dialogue between users is explicit, two-way communication using language. Dialogue between users includes face-to-face conversations or voice calls such as telephone calls. Dialogue between users also includes written communication such as email or chat. Dialogue between users may be real-time or asynchronous. In Figure 1, users 2a, 2b, and 2c are shown. Here, when individual users such as users 2a, 2b, and 2c are not specifically distinguished, they may simply be referred to as user 2.

[0015] Each user 2 of the conversational forecasting system 1 uses the conversational forecasting system 1 through a terminal device. In this example, each user 2 uses one or more terminal devices corresponding to themselves. For example, user 2a uses terminal device 3a, user 2b uses terminal device 3b, and user 2c uses terminal device 3c. Here, when individual terminal devices 3 such as terminal devices 3a, terminal device 3b, and terminal device 3c are not specifically distinguished, they may simply be referred to as terminal device 3. Some or all of the multiple terminal devices 3 used in the conversational forecasting system 1 may be internal devices included in the conversational forecasting system 1, or external devices outside the system that can cooperate with the conversational forecasting system 1. Terminal devices 3 are, for example, portable information processing terminal devices such as smartphones or smartwatches owned by user 2. Portable terminal devices 3 owned by user 2 are examples of portable devices. Terminal devices 3 may also be stationary information processing terminal devices such as desktop PCs (PCs: Personal Computers). Application software corresponding to the conversational forecasting system 1 is pre-installed on each terminal device 3. Each terminal device 3 operates according to the software program and other instructions to realize the functions of the interactive forecasting system 1. Each terminal device 3 connects to a communication network 4 by wireless or wired communication so that it can communicate information with devices outside of itself. The communication network 4 includes, for example, a wide-area network such as the Internet or a telephone network, or a local network such as a LAN (Local Area Network). A terminal device 3 may also directly communicate wirelessly with external devices, such as other terminal devices 3.

[0016] Some or all of the terminal devices 3 are equipped with a function to acquire current location information by, for example, a satellite positioning system such as GPS (Global Positioning System), an autonomous navigation system, or other systems. The terminal device 3 may also acquire current location information by, for example, an indoor positioning system, based on the received strength of radio signals transmitted from multiple beacons 5 installed in facilities such as the office, company building, or business premises of the organization to which the user 2 belongs.

[0017] Some or all of the terminal devices 3 have the groupware application pre-installed. The groupware includes functions for users 2 to exchange messages with each other, such as email, chat, and instant messenger. Messages may be email or other data formats. Messages may include, for example, text information, audio information, image information, or information from other attachments. The exchange of messages in the groupware is an example of dialogue between users 2. The groupware includes functions to manage information such as user 2's presence, such as whether they are present or away, and user 2's schedule information. The groupware may also manage user 2's location information. Information such as message history, presence, and schedule in the groupware is managed, for example, by an application server 6. The application server 6 is, for example, a computer system including one or more computers. The application server 6 may be an external system that can cooperate with the dialogue forecasting system 1, or an internal system included in the dialogue forecasting system 1. The application server 6 is connected to the communication network 4.

[0018] The dialogue prediction system 1 includes a management device 7. The management device 7 is a device responsible for functions such as information processing in the dialogue prediction system 1. The management device 7 is an example of a dialogue prediction device. The management device 7 is connected to a communication network 4. The management device 7 is, for example, a computer system composed of one or more server devices. Here, a computer system composed of one or more devices may sometimes be simply referred to as a computer. The plurality of server devices constituting the management device 7 may be arranged at different locations from each other. At this time, the plurality of server devices communicate information with each other, for example, through the communication network 4. Part or all of the functions of the management device 7 may be implemented by, for example, processing, storage, or other resources on a cloud service. The management device 7 includes a position acquisition unit 8, a collection unit 9, a prediction unit 10, and an output unit 11.

[0019] The position acquisition unit 8 is a part that acquires the position information of each user 2. The position acquisition unit 8 may, for example, acquire the current position information of the terminal device 3 from the portable terminal device 3 held or worn by the user 2 through the communication network 4 and use this as the position information of the user 2. Each terminal device 3 may, for example, spontaneously transmit position information to the management device 7 at a regular pre-set timing, or may transmit position information to the management device 7 in a form that responds to a request from the management device 7. For example, when one or more cameras 12 for taking images are provided in a facility of an organization to which the user 2 belongs, etc., the position acquisition unit 8 may acquire the position information of the user 2 based on the captured images of each camera 12 acquired through the communication network 4 or the like. In the position acquisition unit 8, feature quantities for identifying the user 2 and the like are registered in advance. The position acquisition unit 8 identifies the user 2 on the captured images of each camera 12 and calculates the position information of the user 2 identified by technologies such as the positions where each camera 12 is arranged and image recognition. Each camera 12 may, for example, spontaneously transmit captured images to the management device 7 at a regular pre-set timing, or may transmit captured images to the management device 7 in a form that responds to a request from the management device 7.

[0020] The collection unit 9 is a part that collects the function of collecting mode information representing the mode when a certain user 2 desires to interact with another user 2. In the following example, it is assumed that user 2a desires to interact with user 2b. The collection unit 9 collects the mode information of user 2a regarding the interaction from user 2a to user 2b. The mode information of user 2a at this time is information representing the implicit actions of user 2a or the state of user 2a when user 2a desires to interact with user 2b. Here, the implicit action of user 2a is an action without an explicit invitation to interact with user 2b. The mode information of user 2a may be information representing the actions that user 2a takes when attempting to interact with user 2b, or may be information representing the state of user 2a that is presumed to be in a situation where user 2a desires to interact with user 2b. The collection unit 9 collects the mode information of user 2a through a communication network 4 or the like, for example, based on the usage status of groupware by user 2a, the captured images of camera 12 provided in facilities such as the organization to which user 2a belongs, the location information of user 2a, or other information. The collection unit 9 may, for example, request the provision of information from an external device at a regular pre-set timing, or may collect information spontaneously provided by the external device.

[0021] The collection unit 9 collects, for example, the check frequency by user 2a of information related to user 2b as the mode information of user 2a regarding the interaction from user 2a to user 2b. Information related to user 2b is, for example, the schedule of user 2b, the location information of user 2b, the message exchange history between user 2b and user 2a, and the like. The collection unit 9 collects, for example, through groupware, the browsing by user 2a of information related to user 2b, or the number of times per unit time of browsing or the time interval between multiple browsings, etc., as the check frequency from an application server 6 or the like. The collection unit 9 collects, for example, through groupware, the number of times of browsing per unit time by user 2a of information related to user 2b, or the time interval during which user 2a browses information related to user 2b, etc., as the check frequency from an application server 6 or the like.

[0022] The collection unit 9 collects information on communication gaps in a communication tool used for communication between user 2a and user 2b, for example, as information on user 2a's behavior in the dialogue from user 2a to user 2b. The communication tool between users 2a and user 2b includes, for example, asynchronous message exchange tools such as chat on groupware. The collection unit 9 collects information on user 2b's reply delay as a communication gap, for example, the time it takes for user 2b to reply to a message sent by user 2a. The collection unit 9 may also collect information on differences, such as the difference or ratio between user 2a's reply delay and user 2b's reply delay, as a communication gap. The collection unit 9 may also detect the degree of discrepancy between user 2a and user 2b by analyzing the content of messages based on the message exchange history, and collect the degree of such discrepancy as a communication gap. The collection unit 9 collects information on communication gaps from an application server 6 or the like, for example, based on the message exchange history managed in groupware.

[0023] The data collection unit 9 collects, for example, the positional relationship between the location information of user 2a and the location information of user 2b as information about user 2a's behavior regarding the interaction from user 2a to user 2b. The location information of user 2a and user 2b is acquired by the location acquisition unit 8. The positional relationship between user 2a and user 2b includes, for example, the distance between user 2a and user 2b. The positional relationship between user 2a and user 2b may also include, for example, the estimated time until one of user 2a or user 2b arrives at the location where the other is located.

[0024] The collection unit 9 collects, for example, user 2a's biometric information as behavioral information of user 2a regarding the dialogue between user 2a and user 2b. The collection unit 9 also collects, for example, body temperature, pulse rate, or other biometric information measured by a terminal device 3 worn by user 2 as user 2a's biometric information from the terminal device 3, etc. For example, when user 2a explicitly or implicitly desires a dialogue, it may affect user 2a's stress level. Also, for example, if user 2b is in a position such as a superior who can resolve the source of stress, user 2a may desire a dialogue with user 2b when experiencing stress. Therefore, biometric information reflecting user 2a's stress level can become behavioral information reflecting user 2a's desire for a dialogue.

[0025] The data collection unit 9 collects, for example, user 2a's behavioral patterns as user 2a behavioral information regarding the dialogue between user 2a and user 2b. User 2a's behavioral patterns may include, for example, user 2a's movement history based on user 2a's location information. The data collection unit 9 may also detect behavioral patterns of user 2a that differ from normal behavior and collect the detected behavioral patterns as behavioral information. For example, if user 2a wishes to converse with user 2b, user 2a may move to a place where user 2b might be. A place where user 2b might be might be, for example, user 2b's room, a room such as a conference room or workroom that user 2b frequently uses, or a smoking area if user 2b is a smoker. In particular, if the place where user 2b might be is a place that user 2a does not normally enter, user 2a's movement to that place may be a behavioral pattern resulting from user 2a's desire to converse with user 2b. Therefore, user 2a's behavioral patterns may be behavioral information that reflects user 2a's desire for conversation.

[0026] The collection unit 9 may collect other information besides the information exemplified herein as user 2a's characteristic information. The collection unit 9 may also collect some or all of the information exemplified herein or other information as characteristic information.

[0027] The prediction unit 10 is the part that predicts the likelihood of a dialogue occurring between two users 2, based on the interaction information collected by the collection unit 9 from one user 2 to another user 2. The likelihood of a dialogue may be continuous numerical information, such as the probability of a dialogue occurring. The likelihood of a dialogue may also be a dialogue occurrence level, which represents the likelihood of a dialogue occurring at multiple levels. The likelihood of a dialogue may be represented by a numerical value or indicator, for example, that allows for comparison of the degree of likelihood of a dialogue occurring. The prediction unit 10 may predict the likelihood of a dialogue for several future points in time. The prediction unit 10 may predict the likelihood of a dialogue for each point in time after a predetermined amount of time has elapsed from the present. That is, the prediction unit 10 may predict the time-series change in the likelihood of a dialogue at each future point in time, for example, every 15 minutes or every hour. In this example, the prediction unit 10 predicts the likelihood of a dialogue using a dialogue occurrence score, which numerically represents the degree of likelihood of a dialogue occurring.

[0028] The prediction unit 10 calculates a dialogue occurrence score by, for example, scoring the manner information collected by the collection unit 9 based on pre-set criteria for each type of manner information, multiplying these scores by pre-set weight coefficients according to the type of manner information, and adding them together. The prediction unit 10 may gradually decrease the score for each type of manner information according to the elapsed time since the manner information was collected. The strength of the decrease according to the elapsed time may be set for each type of manner information. When new manner information is collected, the prediction unit 10 may update the predicted value of the dialogue occurrence score by adding the score based on that manner information to the dialogue occurrence score calculated so far. Furthermore, the prediction unit 10 may update the predicted value of the dialogue occurrence score even when no new manner information has been collected, in response to the decrease in the score for each type of manner information.

[0029] The prediction unit 10 scores the frequency with which user 2a checks information related to user 2b, for example, as follows: For each time user 2a views information related to user 2b, the prediction unit 10 accumulates and adds a pre-set score to the current check frequency score. The prediction unit 10 then decays the check frequency score at a predetermined decay rate according to the elapsed time. As a result, when user 2a views the information frequently, the check frequency score is added before it has decayed sufficiently, resulting in a high check frequency score. The prediction unit 10 may also calculate a check frequency score for each type of information about user 2b that user 2a views. For example, the prediction unit 10 may rate the check frequency score for information that more directly represents user 2b's behavior, such as user 2b's schedule and user 2b's location information, higher than the check frequency score for asynchronous communication tools such as chat.

[0030] The prediction unit 10 scores the communication gap between user 2a and user 2b, for example, as follows: The prediction unit 10 rates the communication gap score higher the greater the delay in user 2b's response. The prediction unit 10 rates the communication gap score higher the greater the difference in response delays between user 2b and user 2a. The prediction unit 10 rates the communication gap score higher the greater the degree of discrepancy between user 2a and user 2b. The prediction unit 10 accumulates and adds a predetermined score to the current communication gap score each time a communication gap occurs. The prediction unit 10 attenuates the communication gap score at a predetermined constant decay rate according to the elapsed time. As a result, communication gaps that occurred more recently are evaluated with a higher score. Also, if communication gaps occur frequently, the communication gap score will be high because it is added before sufficient decay occurs.

[0031] The prediction unit 10 scores the positional relationship between user 2a and user 2b, for example, as follows: The prediction unit 10 evaluates the position score according to the distance between user 2a and user 2b. The prediction unit 10 calculates a position score based on a range of distances between user 2a and user 2b, with a score predetermined for that range. As a more specific example, the prediction unit 10 assigns a position score of 10 points if the distance between user 2a and user 2b is less than 10m, a position score of 5 points if it is 10m or more but less than 50m, and a position score of 1 point if it is 50m or more. When updating the positional relationship based on the current position, the prediction unit 10 does not need to allow the position score to decay over time. The prediction unit 10 may predict the future position score by predicting user 2's future position information. The prediction of user 2's future position information may be based on the history of user 2's past position information, or it may be based on Brownian motion or other probabilistic models.

[0032] The prediction unit 10 scores the user 2a's biometric information, for example, as follows: The prediction unit 10 evaluates the biometric information score more highly the higher the stress level of user 2a calculated based on the biometric information.

[0033] The prediction unit 10 scores the user 2a's behavior patterns, for example, as follows: For example, each time the prediction unit 10 detects a behavior pattern that differs from user 2a's normal behavior, it accumulates and adds a pre-set score to the current behavior pattern score. For example, the prediction unit 10 attenuates the behavior pattern score at a predetermined constant decay rate according to the elapsed time. As a result, when user 2a frequently behaves in a pattern different from normal, the behavior pattern score is added before sufficient decay occurs, resulting in a high value for the behavior pattern.

[0034] The prediction unit 10 calculates a dialogue occurrence score as a measure of the likelihood of a dialogue, for example, as follows. In this example, the prediction unit 10 evaluates the location score as 8 points, the check frequency score as 6 points, and the communication gap score as 7 points. The prediction unit 10 also sets the weight coefficients for the location score, check frequency score, and communication gap score to 0.4, 0.3, and 0.3, respectively. At this time, the prediction unit 10 calculates the predicted value of the dialogue occurrence score as (8 × 0.4) + (6 × 0.3) + (7 × 0.3) = 7.1 by multiplying these scores by the weight coefficients and adding them together.

[0035] The output unit 11 is the part that outputs the likelihood of a conversation predicted by the prediction unit 10. The output unit 11 notifies user 2b of the likelihood of a conversation predicted for a conversation between user 2a and user 2b. The output unit 11 notifies user 2b, for example, via the communication network 4 and an application on the terminal device 3b that corresponds to the conversation forecasting system 1. The output unit 11 notifies user 2b whenever the prediction unit 10 predicts or updates the likelihood of a conversation with user 2b. The output unit 11 may also notify user 2b when the likelihood of a conversation with user 2b exceeds a preset threshold. When the prediction unit 10 predicts the time-series change in the likelihood of a conversation at each future point in time, the output unit 11 notifies user 2b of the likelihood of a conversation at each point in time. The application on the terminal device 3b that corresponds to the conversation forecasting system 1 notifies user 2b of the predicted likelihood of a conversation, for example, by numerical values, icons, or graphs of time-series data on a dedicated dashboard screen. The output unit 11 may notify user 2a of the predicted likelihood of a conversation between user 2a and user 2b.

[0036] User 2b attempts to engage in dialogue with User 2a by referring to the likelihood of dialogue notified by the output unit 11. This facilitates the occurrence of necessary dialogue between User 2a and User 2b, even if User 2a has not made an explicit request for dialogue. This makes communication between Users 2 and User 2b smoother.

[0037] In addition, the output unit 11 may have information disclosure policies and forecast notification policies set for each user 2. The information disclosure policy is a guideline for disclosing information such as the type of information collected and the predicted likelihood of interaction. The information disclosure policy includes, for example, whether or not to disclose the likelihood of interaction and type of information, the scope of disclosure of the likelihood of interaction and type of information, and the selection of information to disclose. If the disclosure requirement for the likelihood of interaction for user 2a is set to not disclose, the output unit 11 will not disclose the predicted likelihood of interaction for the interaction from user 2a to user 2b to anyone other than user 2a and user 2b. On the other hand, if the disclosure requirement for the likelihood of interaction for user 2a is set to disclose, the output unit 11 will make the predicted likelihood of interaction for the interaction from user 2a to user 2b viewable by anyone other than user 2a and user 2b. In this case, the output unit 11 will disclose the likelihood of interaction to the scope of disclosure set by user 2a. The scope of disclosure may be, for example, within the organization to which user 2a or user 2b belongs, or to user 2a's superior. User 2, who is included in the scope of public access, can view the likelihood of interaction with User 2a through User 2's terminal device 3, etc. The forecast notification policy is a guideline for notifying users of the predicted likelihood of interaction. The forecast notification policy may include, for example, a threshold for the likelihood of interaction regarding whether or not to send a notification. The information disclosure policy and the forecast notification policy are set, for example, by User 2 through User 2's terminal device 3. The information disclosure policy and the forecast notification policy are set, for example, by User 2 through User 2's terminal device 3. The information disclosure policy and the forecast notification policy may be set collectively for multiple Users 2 by an administrator, for example. The administrator may be User 2c, etc. User 2c, who is the administrator, may set the information disclosure policy and the forecast notification policy collectively through their own terminal device 3c, or they may set the information disclosure policy and the forecast notification policy collectively through a dedicated management terminal device. The management terminal device is, for example, an information processing terminal device connected to the communication network 4, such as a smartphone or a desktop PC.

[0038] In this example, the collection unit 9 collects feedback evaluations regarding the accuracy of the predicted likelihood of a dialogue. The collection unit 9 collects feedback evaluations from, for example, one or both of the user 2a who wanted to have a dialogue and / or the user 2b who was the other party to the dialogue, via the terminal device 3.

[0039] User 2b, for example, after engaging in dialogue with User 2a based on a notification from Output Unit 11, evaluates the accuracy of the notified dialogue likelihood level according to the content of the dialogue with User 2a. Based on the accuracy of the dialogue likelihood level that User 2b has evaluated, User 2b sends a feedback evaluation to Collection Unit 9, for example, through an application on Terminal Device 3b corresponding to Dialogue Forecasting System 1. The feedback evaluation for the dialogue likelihood level may be a binary value, for example, whether the notified dialogue likelihood level was accurate or not. In this case, the input for the feedback evaluation may be the selection of an evaluation button, for example. The feedback evaluation for the dialogue likelihood level may be a selection from multiple options, for example, whether the notified dialogue likelihood level was appropriate, whether the dialogue likelihood level should be evaluated higher in the future, or whether the dialogue likelihood level should be evaluated lower in the future. The feedback evaluation for the dialogue likelihood level may be the dialogue likelihood level value itself evaluated by User 2b. The feedback evaluation for the dialogue likelihood level may include additional information such as supplementary information or comments.

[0040] User 2a evaluates the accuracy of the dialogue likelihood notified to User 2b, for example, depending on whether or not there was a dialogue with User 2b, or the content of the dialogue with User 2b. Here, the dialogue likelihood notified to User 2b is also notified to User 2a by, for example, the output unit 11. User 2a, like User 2b, sends a feedback evaluation to the collection unit 9 based on the accuracy of the dialogue likelihood they evaluated, for example, through an application on the terminal device 3a corresponding to the dialogue forecasting system 1.

[0041] The prediction unit 10 updates the prediction parameters used to predict the likelihood of a dialogue based on the feedback evaluations collected by the collection unit 9. The prediction parameters include, for example, threshold or range boundary values ​​in a predetermined standard for each type of personality information, a conversion coefficient for the score to be added for each type of personality information, a decay rate of the score for each type of personality information, or a weight coefficient for the score for each type of personality information. When the prediction unit 10 predicts the likelihood of a dialogue using personality information as input by a pre-trained prediction model, the prediction parameters may be the learning parameters of the prediction model. The prediction unit 10 accumulates and aggregates the feedback evaluations collected by the collection unit 9 and updates the prediction parameters based on the aggregation results. The prediction parameters may be common to each user 2, set for each user 2 who wants to have a dialogue, set for each user 2 who is the dialogue partner, or set for each combination of user 2 who wants to have a dialogue and user 2 who is the dialogue partner. The prediction unit 10 makes a new prediction of the likelihood of a dialogue using the updated prediction parameters. At this point, the newly predicted likelihood of dialogue is expected to be more accurate because it reflects the collected feedback evaluations.

[0042] Next, we will explain an example of processing in the dialogue forecasting system 1 using Figures 2 and 3. Figures 2 and 3 are flowcharts showing an example of processing in the dialogue forecasting system 1 according to Embodiment 1.

[0043] Figure 2 shows an example of the processing performed by the management device 7 for collecting behavioral information and notifying the likelihood of interaction. The series of processes shown in Figure 2 may be started, for example, at a regularly scheduled, pre-set timing, or may be triggered by other events. In this example, the management device 7 collects information on location, check frequency, and communication gaps as behavioral information.

[0044] The collection unit 9 collects behavioral information of user 2 who wishes to engage in dialogue (step S21). Then, the prediction unit 10 predicts the likelihood of dialogue based on the collected behavioral information. In predicting the likelihood of dialogue, the prediction unit 10 sequentially calculates a location score (step S22), a check frequency score (step S23), and a communication gap score (step S24). The prediction unit 10 may calculate each score in an order different from the one exemplified here, or it may calculate each score in parallel. Then, the prediction unit 10 calculates a dialogue occurrence score by multiplying each score by a weight coefficient and adding them together (step S25). Then, the output unit 11 notifies user 2, who will be the dialogue partner, of the calculated dialogue occurrence score as the likelihood of dialogue (step S26).

[0045] Figure 3 shows an example of the processing performed by the management device 7 in relation to updating prediction parameters.

[0046] The management device 7 performs the processing of collecting behavioral information and notifying the likelihood of a dialogue, for example, as shown in Figure 2 (step S31). Subsequently, the collection unit 9 collects feedback evaluations from the user 2 regarding the accuracy of the likelihood of a dialogue (step S32). Then, the prediction unit 10 stores and aggregates the collected feedback evaluations (step S33).

[0047] Subsequently, the management device 7 determines whether or not it is time to update the prediction parameters (step S34). The timing for updating the prediction parameters may be, for example, a regularly scheduled time, a time specified by an administrator, or a time when an administrator performs an update operation. The timing for updating the prediction parameters may also be, for example, when the number of collected feedback evaluations exceeds a predetermined criterion. If it is not time to update the prediction parameters, the management device 7 proceeds to step S31. On the other hand, if it is time to update the prediction parameters, the prediction unit 10 performs the prediction parameter update process based on the aggregated results of the feedback evaluations (step S35). Subsequently, the management device 7 proceeds to step S31.

[0048] The management device 7 uses the updated prediction parameters to collect behavioral information and notify the likelihood of interaction (step S31). Through this feedback loop type learning mechanism, the management device 7 continuously and gradually improves the accuracy of its prediction of the likelihood of interaction.

[0049] As described above, the management device 7 of the dialogue prediction system 1 according to Embodiment 1 comprises a collection unit 9, a prediction unit 10, and an output unit 11. The collection unit 9 collects behavioral information when user 2a desires to have a dialogue with user 2b. This behavioral information represents actions or states of user 2a that do not involve an explicit request for dialogue. Based on the behavioral information collected by the collection unit 9, the prediction unit 10 predicts the likelihood of a dialogue occurring between user 2a and user 2b. The output unit 11 notifies user 2b of the likelihood of a dialogue predicted by the prediction unit 10.

[0050] With this configuration, user 2b, the conversation partner, is motivated to engage in conversation with user 2a by being notified of the likelihood of conversation, even if user 2a, who desires the conversation, has not explicitly requested it. This promotes the occurrence of necessary conversations between users 2a and 2b, leading to smoother communication. User 2a may be unable to explicitly request a conversation, hesitate to do so, or be unsure whether to initiate a conversation, for reasons such as the content or urgency of the conversation, or their relationship with user 2b. Even in such cases, the notification of the likelihood of conversation promotes the occurrence of conversations between users 2a and 2b. Furthermore, even in situations where direct observation of each other's appearance or behavior is not possible, such as in remote work or a free-address system, user 2b can understand that other users 2a are interested in conversation through the notification of the likelihood of conversation. This leads to smoother communication between users and mitigates isolation due to lack of conversation.

[0051] Furthermore, the prediction unit 10 updates the predicted likelihood of dialogue after predicting it. The prediction unit 10 updates the prediction based on, for example, behavioral information collected after predicting the likelihood of dialogue. The output unit 11 notifies user 2b of the updated likelihood of dialogue each time the prediction unit 10 updates the prediction of the likelihood of dialogue from user 2a to user 2b. As a result, user 2b is notified of the latest likelihood of dialogue. This enables user 2b to communicate with user 2a more smoothly.

[0052] Furthermore, the collection unit 9 collects the frequency with which user 2a checks information related to user 2b as behavioral information. The prediction unit 10 uses the collected check frequency to predict the likelihood of a conversation between user 2a and user 2b. When user 2a wishes to have a conversation with user 2b, they may frequently check user 2b's location, schedule, or other information. Thus, the frequency with which user 2a checks can reflect user 2a's behavior when they wish to have a conversation. For this reason, the prediction unit 10 can detect the possibility of a conversation between user 2a and user 2b occurring without relying on an explicit request for a conversation from user 2a.

[0053] Furthermore, the collection unit 9 collects communication gaps in the communication tool used for communication between user 2a and user 2b as behavioral information. The prediction unit 10 uses the collected communication gaps to predict the likelihood of a dialogue between user 2a and user 2b. User 2a may want to have a dialogue with user 2b if there is a communication gap between them. Therefore, based on the communication gaps, the prediction unit 10 can detect the possibility of a dialogue between user 2a and user 2b occurring, without relying on an explicit request for dialogue from user 2a.

[0054] The management device 7 also includes a location acquisition unit 8. The location acquisition unit 8 acquires location information of users 2a and 2b. The collection unit 9 collects the positional relationship of the location information between users 2a and 2b acquired by the location acquisition unit 8 as behavioral information. The prediction unit 10 uses the collected positional relationship to predict the likelihood of a conversation between user 2a and user 2b. User 2a may approach user 2b if they wish to have a conversation with user 2b. Also, when users 2a and 2b are close to each other, users 2a and 2b can start a conversation with less resistance or obstacle. Thus, the positional relationship between users 2a and 2b can reflect user 2a's behavior when they wish to have a conversation, and the likelihood of a conversation occurring. For this reason, the prediction unit 10 can detect the possibility of a conversation between users 2a and 2b occurring without relying on an explicit request for a conversation from user 2a.

[0055] Furthermore, the collection unit 9 collects user 2a's biometric information as behavioral information. The prediction unit 10 uses the collected biometric information to predict the likelihood of a dialogue between user 2a and user 2b. For example, if user 2b is user 2a's superior, user 2a may want to talk to user 2b when they are in a stressful situation. Thus, biometric information that reflects user 2a's stress level can reflect user 2a's desire for dialogue. Therefore, the prediction unit 10 can detect the possibility of a dialogue between user 2a and user 2b without relying on an explicit request for dialogue from user 2a.

[0056] Furthermore, the collection unit 9 collects user 2a's behavioral patterns as behavioral information. The prediction unit 10 uses the collected behavioral patterns to predict the likelihood of a dialogue between user 2a and user 2b. When user 2a desires to interact with user 2b, they may adopt behavioral patterns tailored to user 2b. Thus, user 2a's behavioral patterns can reflect user 2a's desire for dialogue. Therefore, the prediction unit 10 can detect the possibility of a dialogue between user 2a and user 2b without relying on an explicit request for dialogue from user 2a.

[0057] Furthermore, the prediction unit 10 predicts the likelihood of a conversation at several future points in time. The output unit 11 notifies user 2b of the predicted likelihood of a conversation at each point in time. User 2b can understand the time-series changes in the likelihood of a conversation at each future point in time through the notifications from the output unit 11. This allows user 2b to consider the optimal timing to start a conversation with user 2a. As a result, smoother communication between user 2a and user 2b is further promoted.

[0058] The output unit 11 publishes the likelihood of dialogue to the scope set by user 2a when user 2a has set the disclosure requirement to "disclose likelihood of dialogue". Because the likelihood of dialogue is disclosed, other users 2 can also understand that user 2a is interested in dialogue with user 2b. This promotes the occurrence of dialogue between user 2a and user 2b, for example, through the intervention of other users 2. In addition, since the likelihood of dialogue is disclosed to the set disclosure scope, user 2a's information is not unnecessarily disseminated.

[0059] Furthermore, the collection unit 9 collects feedback evaluations from the user 2 regarding the accuracy of the likelihood of a conversation. The prediction unit 10 updates the prediction parameters used to predict the likelihood of a conversation based on the collected feedback evaluations. This allows the management device 7 to further improve the accuracy of its predictions of the likelihood of a conversation.

[0060] Furthermore, the collection unit 9 collects user 2's behavioral information through the terminal device 3 that user 2 possesses and carries with them. Because user 2's behavioral information is collected through the terminal device 3 that user 2 carries with them, more accurate behavioral information is collected.

[0061] Furthermore, the Dialogue Prediction System 1 is not limited to use in organizations such as companies. For example, the Dialogue Prediction System 1 can be applied to facilitating dialogue between customers and staff in commercial stores, or between teachers and students in educational institutions. The Dialogue Prediction System 1 can also be applied to other situations where dialogue may occur, such as in public services.

[0062] For example, in a commercial store, the collection unit 9 collects customer behavior information through cameras 12 installed in the store. The collection unit 9 identifies the customer's location by image recognition from the camera 12 and collects the customer's behavioral patterns as behavioral information. The prediction unit 10 adds a preset value to the customer's dialogue occurrence score when behavioral patterns such as staying in a specific location in the store, looking around, or raising a hand are collected as behavioral information. The prediction unit 10 may add a higher dialogue occurrence score for longer dwell times depending on the time spent in a specific location in the store. The output unit 11 notifies the store staff of the likelihood of a customer's dialogue occurring, as predicted by the prediction unit 10, along with the customer's location. The commercial store may be a store where customers move around to select products, such as a clothing store, or a store where customers remain seated, such as a restaurant. If a customer is using a store's dedicated application via their mobile device, the collection unit 9 may collect the usage status of that application as behavioral information. The collection unit 9 may acquire the customer's location information through the customer's mobile device and the application. The prediction unit 10 may add a pre-set value to the customer's conversation occurrence score if the customer is viewing specific information, such as the store's "Frequently Asked Questions."

[0063] For example, in an educational institution, the collection unit 9 collects student behavior information through online learning tools or learning management systems used by students. The collection unit 9 collects behavior information from the student to their teacher or the teacher in charge of the material, such as the frequency of students logging into learning tools or the frequency of accessing specific learning materials. The prediction unit 10 uses behavior information such as login frequency or access frequency to predict the likelihood of dialogue in a similar manner to, for example, the frequency of checking the information of the corresponding teacher. The collection unit 9 may also collect communication gaps between students and teachers as behavior information, for example, based on the message exchange history in a communication tool for instruction. The collection unit 9 may also collect the student's level of understanding, such as the correct answer rate on assignments about specific learning content, as a communication gap between the student and the teacher in charge of that learning content. Furthermore, the dialogue prediction system 1 may calculate the likelihood of dialogue from teachers to students, not just from students to teachers.

[0064] Next, we will explain an example of the hardware configuration of the dialogue forecasting system 1 using Figure 4. Figure 4 is a hardware configuration diagram of the main components of the dialogue prediction system 1 according to Embodiment 1.

[0065] Some or all of the functions of the interactive forecasting system 1 can be implemented by a processing circuit. The processing circuit comprises at least one processor 100a and at least one memory 100b. The processing circuit may also include at least one dedicated hardware component along with the processor 100a and memory 100b, or as a substitute for them.

[0066] When the processing circuit includes a processor 100a and a memory 100b, each function of the interactive forecasting system 1 is implemented by software, firmware, or a combination of software and firmware. At least one of the software and firmware is written as a program. This program is stored in the memory 100b. The processor 100a implements each function of the interactive forecasting system 1 by reading and executing the program stored in the memory 100b. The program may also be a program package that includes multiple subprograms, modules, or libraries. The program is sometimes called a program product.

[0067] The processor 100a is also called a CPU (Central Processing Unit), processing unit, arithmetic unit, microprocessor, microcomputer, or DSP. The memory 100b is composed of non-volatile or volatile semiconductor memory such as RAM, ROM, flash memory, EPROM, or EEPROM.

[0068] When a processing circuit has dedicated hardware, it can be implemented as, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof.

[0069] Each function of the interactive forecasting system 1 can be implemented by a separate processing circuit. Alternatively, each function of the interactive forecasting system 1 can be implemented collectively by a single processing circuit. For each function of the interactive forecasting system 1, some may be implemented by dedicated hardware, while others are implemented by software or firmware. Thus, the processing circuit implements each function of the interactive forecasting system 1 using dedicated hardware, software, firmware, or a combination thereof.

[0070] To summarize the above explanation, the possible configurations of the technology relating to this disclosure include the configurations listed below as appendices. (Note 1) A collection unit that collects behavioral information representing the actions or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user, Based on the behavioral information of the first user collected by the collection unit, a prediction unit predicts the likelihood of a dialogue occurring between the first user and the second user, and An output unit that notifies the second user of the likelihood of the dialogue predicted by the prediction unit, An interactive forecasting device equipped with the following features. (Note 2) The prediction unit updates the prediction of the likelihood of dialogue after predicting the likelihood of dialogue, The output unit notifies the second user of the updated likelihood of dialogue each time the prediction unit updates the likelihood of dialogue. The interactive prediction device described in Appendix 1. (Note 3) The collection unit collects the frequency with which the first user checks information related to the second user as the aspect information, The prediction unit predicts the likelihood of the dialogue using the collected check frequency. The interactive forecasting device described in Appendix 1 or Appendix 2. (Note 4) The collection unit collects the communication gap in the communication tool used for communication between the first user and the second user as the manner information. The prediction unit uses the collected communication gap to predict the likelihood of the conversation. The interactive forecasting device described in Appendix 1 or Appendix 2. (Note 5) Location acquisition unit that acquires the location information of the first user and the location information of the second user. Equipped with, The collection unit collects the positional relationship between the location information of the first user and the location information of the second user acquired by the location acquisition unit as the configuration information. The prediction unit predicts the likelihood of the dialogue using the collected positional relationships. The interactive forecasting device described in Appendix 1 or Appendix 2. (Note 6) The collection unit collects the first user's biometric information as the characteristic information, The prediction unit uses the collected biometric information to predict the likelihood of the interaction. The interactive forecasting device described in Appendix 1 or Appendix 2. (Note 7) The collection unit collects the behavioral patterns of the first user as behavioral information, The prediction unit predicts the likelihood of the dialogue using the collected behavior patterns. The interactive forecasting device described in Appendix 1 or Appendix 2. (Note 8) The prediction unit predicts the likelihood of the dialogue for each of several future points in time, The output unit notifies the second user of the predicted likelihood of the dialogue for each of the multiple time points. An interactive forecasting device as described in any one of the appendices 1 through 7. (Note 9) The output unit, when the requirement for disclosure of the likelihood of dialogue is set to "disclosure required" by the first user, discloses the likelihood of dialogue to the disclosure range predetermined by the first user. An interactive forecasting device as described in any one of the appendices 1 through 8. (Note 10) The collection unit collects feedback evaluations from either or both of the first user or the second user regarding the accuracy of the likelihood of the dialogue, The prediction unit updates the prediction parameters used to predict the likelihood of the dialogue based on the collected feedback evaluation. An interactive forecasting device as described in any one of the appendices 1 through 9. (Note 11) The collection unit collects information about the first user's characteristics through a portable device possessed by the first user. An interactive forecasting device as described in any one of the appendices 1 through 10. (Note 12) The first user's mobile device and A collection unit that collects information through the portable device that represents behaviors or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user, Based on the behavioral information of the first user collected by the collection unit, a prediction unit predicts the likelihood of a dialogue occurring between the first user and the second user, and An output unit that notifies the second user of the likelihood of the dialogue predicted by the prediction unit, An interactive forecasting system equipped with the following features. (Note 13) Computers To collect behavioral information that represents the actions or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user, Based on the collected behavioral information of the first user, predict the likelihood of a dialogue occurring between the first user and the second user, To notify the second user of the predicted likelihood of the aforementioned interaction, An interactive forecasting method that performs this operation. (Note 14) On the computer, To collect behavioral information that represents the actions or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user, Based on the collected behavioral information of the first user, predict the likelihood of a dialogue occurring between the first user and the second user, To notify the second user of the predicted likelihood of the aforementioned interaction, An interactive forecasting program that executes [the following]. [Explanation of Symbols]

[0071] 1. Interactive forecasting system; 2, 2a, 2b, 2c. Users; 3, 3a, 3b, 3c. Terminal devices; 4. Communication network; 5. Beacon; 6. Application server; 7. Management device; 8. Location acquisition unit; 9. Data collection unit; 10. Prediction unit; 11. Output unit; 12. Camera; 100a. Processor; 100b. Memory; 200. Dedicated hardware

Claims

1. A collection unit that collects behavioral information representing actions or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user, A prediction unit predicts the likelihood of a dialogue occurring between the first user and the second user, based on the behavioral information of the first user collected by the collection unit, An output unit that notifies the second user of the likelihood of the dialogue predicted by the prediction unit, Equipped with, The prediction unit scores the first user's manner information based on a predetermined criterion for each type of manner information, and calculates the likelihood of dialogue by adding the scores multiplied by a predetermined weighting coefficient for each type of manner information. Dialogue forecasting device.

2. The prediction unit updates the prediction of the likelihood of dialogue after predicting the likelihood of dialogue, The output unit notifies the second user of the updated likelihood of dialogue each time the prediction unit updates the prediction of the likelihood of dialogue. The dialogue prediction device according to claim 1.

3. The collection unit collects the frequency with which the first user checks information related to the second user as the aspect information. The prediction unit predicts the likelihood of the dialogue using the collected check frequency. The dialogue prediction device according to claim 1.

4. The collection unit collects the communication gap in the communication tool used for communication between the first user and the second user as the manner information. The prediction unit uses the collected communication gap to predict the likelihood of the conversation. The dialogue prediction device according to claim 1.

5. Location acquisition unit that acquires the location information of the first user and the location information of the second user. Equipped with, The collection unit collects the positional relationship between the location information of the first user and the location information of the second user acquired by the location acquisition unit as the configuration information. The prediction unit predicts the likelihood of the dialogue using the collected positional relationships. The dialogue prediction device according to claim 1.

6. The collection unit collects the first user's biometric information as the characteristic information, The prediction unit uses the collected biometric information to predict the likelihood of the interaction. The dialogue prediction device according to claim 1.

7. The collection unit collects the behavioral patterns of the first user as behavioral information, The prediction unit predicts the likelihood of the dialogue using the collected behavior patterns. The dialogue prediction device according to claim 1.

8. The prediction unit predicts the likelihood of the dialogue for each of several future points in time, The output unit notifies the second user of the predicted likelihood of the dialogue for each of the multiple time points. The dialogue prediction device according to any one of claims 1 to 7.

9. The output unit, when the requirement for disclosure of the likelihood of dialogue is set to "disclosure required" by the first user, discloses the likelihood of dialogue to the disclosure range predetermined by the first user. The dialogue prediction device according to any one of claims 1 to 7.

10. The collection unit collects feedback evaluations from either or both of the first user or the second user regarding the accuracy of the likelihood of the conversation, The prediction unit updates the prediction parameters used to predict the likelihood of the dialogue based on the collected feedback evaluation. The dialogue prediction device according to any one of claims 1 to 7.

11. The collection unit collects information on the characteristics of the first user through a portable device possessed by the first user. The dialogue prediction device according to any one of claims 1 to 7.

12. The first user's mobile device and A collection unit that collects information through the portable device that represents behaviors or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user, A prediction unit predicts the likelihood of a dialogue occurring between the first user and the second user, based on the behavioral information of the first user collected by the collection unit, An output unit that notifies the second user of the likelihood of the dialogue predicted by the prediction unit, Equipped with, The prediction unit scores the first user's manner information based on a predetermined criterion for each type of manner information, and calculates the likelihood of dialogue by adding the scores multiplied by a predetermined weighting coefficient for each type of manner information. Dialogue forecasting system.

13. Computers To collect behavioral information that represents the actions or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user, Based on the collected behavioral information of the first user, predict the likelihood of a dialogue occurring between the first user and the second user, To notify the second user of the predicted likelihood of the aforementioned interaction, This is a method for performing the task, The likelihood of interaction is calculated by scoring the first user's aspect information based on a predetermined criterion for each type of aspect information, multiplying the score by a predetermined weighting coefficient for each type of aspect information, and adding the scores together. A method of predicting through dialogue.

14. On the computer, To collect behavioral information that represents the actions or states of the first user that do not involve an explicit request for dialogue when the first user desires to interact with the second user, Based on the collected behavioral information of the first user, predict the likelihood of a dialogue occurring between the first user and the second user, To notify the second user of the predicted likelihood of the aforementioned interaction, It is a program that executes, The likelihood of interaction is calculated by scoring the first user's aspect information based on a predetermined criterion for each type of aspect information, multiplying the score by a predetermined weighting coefficient for each type of aspect information, and adding the scores together. A dialogue-based forecasting program.

Images