Information providing device, information providing method, and program

Abstract

To provide a method for providing information on a person who has performed an action when the person who has performed the action is not specified.SOLUTION: An information providing device includes: a prediction unit for using a prediction model generated by learning a plurality of pieces of behavior information, each of which includes a person and information of a plurality of items, to predict a person who performs an action from behavior data including information corresponding to at least a part of the information of the plurality of items of behavior information; and an information display unit for displaying the prediction result of the prediction unit on a display device.SELECTED DRAWING: Figure 1

Description

Technical Field

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

Background Art

[0002] As a related art, Patent Document 1 discloses a criminal investigation support system. The criminal investigation support system described in Patent Document 1 learns an estimation model using information on the action history and human relationships of the persons involved in the case, to which the case type has been assigned as a label, for one or more cases that have already been solved. The criminal investigation support system estimates the type of the case under investigation from the action history and human relationships of the persons involved in the case under investigation using the learned estimation model.

[0003] In Patent Document 1, the criminal investigation support system generates a graph representing the action history information and human relationship information of the persons involved in each of one or more cases that have already been solved. The graph includes nodes representing the persons involved, and edges representing the types of human relationships between the persons involved and the occurrence status of problems between the persons involved. In generating the estimation model, the criminal investigation support system extracts the characteristics of the time-series changes in the actions and human relationships of the persons involved in the case using a predetermined algorithm from the graph to which the case type determined after the solution of the case has been assigned as a label. The criminal investigation support system determines explanatory variables related to the type of the case based on the extraction result, and generates an estimation model including the explanatory variables.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] In Patent Document 1, using an estimation model, the type of an incident under investigation is estimated from the action history and human relations of the persons concerned. However, the crime investigation support system described in Patent Document 1 cannot predict a person who has performed a specific action and provide information about that person.

[0006] One of the objects of the present disclosure is to provide an information providing apparatus, method, and program that can provide information about a person related to an action for which the actor is not specified.

Means for Solving the Problems

[0007] The program according to the first aspect of the present disclosure causes a computer to execute a process including predicting a person related to action data including information corresponding to at least a part of the information of the plurality of items using a prediction model generated by learning a plurality of action information each including a person and information of a plurality of items, and displaying the result of the prediction on a display device.

[0008] The information providing method according to the second aspect of the present disclosure uses a prediction model generated by learning a plurality of action information each including a person and information of a plurality of items, and predicts a person related to the action data from the action data including at least a part of the information of the plurality of items. The method includes displaying the result of the prediction on a display device.

[0009] The information providing apparatus according to the third aspect of the present disclosure includes a prediction unit that predicts a person related to action data including at least a part of the information of the plurality of items using a prediction model generated by learning a plurality of action information each including a person and information of a plurality of items, and an information display unit that displays the prediction result of the prediction unit on a display device.

Effects of the Invention

[0010] The information providing apparatus, method, and program according to the present disclosure can provide information about a person related to an action for an action in which the actor is not specified.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Embodiments for Carrying Out the Invention

[0012] Prior to the description of the embodiments of the present disclosure, the outline of the present disclosure will be described. FIG. 1 shows a schematic configuration example of a relationship visualization apparatus according to the present disclosure. The information providing apparatus 10 shown in FIG. 1 includes a prediction unit 11 and an information display unit 12.

[0013] The prediction unit 11 predicts a person related to the action of the action data from the action data using a prediction model generated by learning a plurality of pieces of action information. Each of the plurality of pieces of action information includes a person and information on a plurality of items. The action data includes information corresponding to at least a part of the information on the plurality of items of the action information. The information display unit 12 displays the prediction result of the prediction unit 11 on a display device.

[0014] In the present disclosure, the prediction unit 11 can predict the person who performed the action from the action data not associated with the person by using the prediction model generated by learning the action information. In other words, the prediction unit 11 can predict the person associated with the action data from the action data for which the actor is not specified. The information display unit 12 displays the prediction result of the prediction unit 11 on the display device. By doing so, the information providing apparatus 10 according to the present disclosure can provide the user with information about the person associated with the action for the action for which the actor is not specified.

[0015] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. Note that, for clarity of explanation, the following description and drawings are appropriately omitted and simplified. Also, in each drawing, the same elements and similar elements are denoted by the same reference numerals, and redundant descriptions are omitted as necessary.

[0016] An embodiment will be described. FIG. 2 shows a configuration example of the information providing apparatus according to the present disclosure. The information providing apparatus 100 includes an action information acquisition unit 101, a model generation unit 102, a data input unit 103, an actor prediction unit 104, and an information display unit 105. Physically, the information providing apparatus 100 can be configured as, for example, an information processing apparatus having one or more processors and one or more memories, or a server apparatus. At least a part of the functions of each unit in the information providing apparatus 100 can be realized by one or more processors operating according to a program read from one or more memories. The information providing apparatus 100 corresponds to the information providing apparatus 10 shown in FIG. 1.

[0017] The action information database 130 stores action information. The action information includes, for example, regarding actions performed in the past, the person who performed the action, that is, the actor, and information indicating the characteristics of the actor's action. The action information includes, for example, an ID (identifier) for identifying the action, the actor (its identification information), and information related to the action characteristics of the actor in the action. The information related to the action characteristics may include, for example, for each type of action, information on a plurality of items according to the type. The action information includes, for example, information on the action location, the means of transportation, the tools used in the action, and the time zone of the action as information related to the action characteristics.

[0018] The action information may be generated, for example, from information published on articles or social media related to actions by natural language processing (NLP) using artificial intelligence (AI) using a server (not shown). Alternatively, the action information may be created manually. The action information may include detailed actor information. The detailed actor information may include, for example, information such as the actor's age, the actor's place of residence, the area where the actor appeared, the time zone when the actor appeared, and the purpose of the action.

[0019] The action information acquisition unit 101 acquires action information from the action information database 130. The action information acquisition unit 101 may acquire, for example, action information related to actions performed within a predetermined period, such as the past 10 years, from the action information database 130.

[0020] The model generation unit 102 generates a prediction model for predicting an actor from the acquired action information, that is, data indicating action features, i.e., action data. For example, the model generation unit 102 uses the actor included in the acquired action information as a label and learns information related to the action features, thereby learning a prediction model for predicting an actor from the information related to the action features. For example, the model generation unit 102 generates a prediction model for estimating hidden relationships in the data. The prediction model may be a model that outputs the basis of the prediction result and a score indicating the accuracy or reliability of the prediction result. The prediction model is also called an AI model.

[0021] The data input unit 103 inputs the action data given to the prediction model to the actor prediction unit 104. The action data includes, for example, information on a plurality of items related to the action features of a certain action. For example, the user can input the action data of an action with an unknown actor that has been newly performed or has been performed in the past to the actor prediction unit 104 via the data input unit 103. Alternatively, the user can also input the action data of an action that is not currently being performed but may be performed in the future to the actor prediction unit 104 via the data input unit 103. The number of information items included in the input action data may be less than the number of information items related to the action features included in the action information stored in the action information DB 130.

[0022] The data input unit 103 may acquire action data from communication records stored in a portable communication device such as a smartphone and input the acquired action data to the action prediction unit 104. For example, a messaging application is installed on the portable communication device. The user of the portable communication device uses the messaging application to conduct a two-person conversation or a group chat. The data input unit 103 acquires conversation history data including the conversation data between the users participating in the conversation.

[0023] The data input unit 103 analyzes the conversation data, and from the conversation data, for a certain action, obtains the actor and information on a plurality of items related to the action characteristics. The data input unit 103 may generate action data including the information extracted from the mobile communication device, and input the generated action data to the action side unit 104. Further, the data input unit 103 may register the action data and the actor in the action information DB 103 as action information. The registered action information can be used as learning data in the generation or update of the prediction model.

[0024] The data input unit 103 may obtain the WiFi connection history from the mobile communication device, and obtain information on the location of the action from the information of the access point of the connection destination. Alternatively, the data input unit 103 may obtain the phone numbers of the phone book registrants and the phone numbers included in the outgoing call history / incoming call history from the mobile communication device, and obtain information on the location of the action from the area codes of the obtained phone numbers. The data input unit 103 may obtain photo data including location information from the mobile communication device, and obtain information on the location of the action from the photo data. The data input unit 103 may obtain the schedule data from the mobile communication device, and obtain information on the location of the action from the location information included in the schedule data.

[0025] The actor prediction unit 104 predicts the actor from the action data input from the data input unit 103 using the prediction model generated by the model generation unit 102. The actor prediction unit 104 may predict a plurality of actors from among the actors included in the action information as actor candidates. Further, the actor prediction unit 104 may output, for each predicted actor, the basis for being predicted as the actor and a score indicating the accuracy or reliability that the predicted actor is the actor of the action in the action data. The actor prediction unit 104 corresponds to the prediction unit 11 shown in FIG. 1.

[0026] The information display unit 105 displays the prediction result of the actor prediction unit 104, that is, the prediction result of the prediction model, on the screen of the display device 150. When a plurality of actors are predicted and scores are assigned to each actor, the information display unit 105 may display the plurality of actors in an order according to the scores. For example, the information display unit 105 displays the plurality of actors in descending order of scores.

[0027] The user can select any actor from the displayed actors. When the user selects an actor, the information display unit 105 may display the actor detailed information of the selected actor. The information display unit 105 acquires the actor detailed information from, for example, the action information DB 130, and displays information such as the age, place of residence, and action area of the actor on the screen. The information display unit 105 may acquire the actor detailed information from an external database that manages the actor detailed information and display the acquired actor detailed information on the screen. The information display unit 105 corresponds to the information display unit 12 shown in FIG. 1.

[0028] In the present embodiment, the prediction model may be a model that predicts an actor having a link with the input action data when the action information is represented as data in a graph format, that is, graph data. Here, the graph data is data in a data format representing the connection relationship of things with nodes and links. The model generation unit 102 generates a prediction model by learning, for example, data obtained by converting a plurality of action information into graph data. The data representing the action information in graph format includes, for example, each of the information of a plurality of items related to the action ID, actor, and action characteristics as nodes. In the graph data, links are formed between the action ID and each of the information of a plurality of items related to the actor and action characteristics. The prediction model can be generated, for example, by comprehensively learning latent feature amounts, relationships, and numerical features from the graph data, that is, the knowledge graph.

[0029] Figure 3 shows an example of graph data of action information. In the example shown in Figure 3, for example, the action information with action ID = 1 includes information that the person is "Actor A", the means of movement is "walking", the time of action is "daytime", and the location is "residential area". In that case, as shown in Figure 3, this action information is represented as graph data in which a link is formed between the node indicating action ID = 1 and the nodes indicating "Actor A", "walking", "daytime", and "residential area" respectively.

[0030] Also, the action information with action ID = 2 includes information that the person is "Actor B", the means of movement is "bicycle", and the location is "city hall". In that case, as shown in Figure 3, this action information is represented as graph data in which a link is formed between the node indicating action ID = 2 and the nodes indicating "Actor B", "city hall", and "bicycle" respectively. The action information DB130 stores action information for a large number of actions.

[0031] Figure 4 shows an example of action data input to the prediction model. In this example, the action data includes information that the means of movement is "bicycle" and the location is "city hall". The action data does not include information about the actor, that is, the person. The prediction model is used to predict an actor who has a hidden link with the node representing the action data among the actors included in the action information.

[0032] Figure 5 shows an example of the prediction result of the prediction model. The model generation unit 102 (see Figure 2) generates a prediction model that discovers hidden connections between nodes, for example, by learning the graph data shown in Figure 3. The prediction model analyzes the similarity between past action information and the input action data in a graph, and predicts an actor linked to the action information similar to the input action data as the person who performed the action, that is, the actor of the action data.

[0033] In the example of FIG. 5, the prediction model discovers the input action data and the action information of action ID = 2 where the location and means of movement match. The prediction model predicts the person "Actor B" linked to the discovered action information of action ID = 2 as the actor of the input action data. As the reason for predicting Actor B, the prediction model outputs that the means of movement is the same bicycle and the location is the same civic hall. Also, the prediction model outputs the score of Actor B.

[0034] The actor prediction unit 104 outputs, as a prediction result, a person who may be the actor of the action of the input action data from among the persons who have taken actions in the past. The information display unit 105 displays the actor B predicted by the actor prediction unit 104 on the display device 150. When a plurality of actors are predicted in the actor prediction unit 104, the information display unit 105 displays the actors in a ranking format according to the scores. Also, the information display unit 105 displays the reason for predicting as the actor on the display device 150.

[0035] Next, the operation procedure will be described. FIG. 6 shows the operation procedure of the information providing apparatus 100. The operation procedure of the information providing apparatus 100 corresponds to the information providing method. As a preliminary preparation, the information providing apparatus 100 generates a prediction model. Alternatively, the information providing apparatus 100 performs re-learning of the prediction model. In the generation or re-learning of the prediction model, the action information acquisition unit 101 collects action information from the action information DB 130. The model generation unit 102 generates a prediction model for predicting an actor from the collected action information from the action data. The re-learning of the prediction model may be performed, for example, when a predetermined number of new action information is accumulated in the action information DB 130. Alternatively, the re-learning of the prediction model may be performed, for example, every time a predetermined period such as six months or one month has elapsed. Note that the prediction model does not necessarily have to be generated within the information providing apparatus 100. The information providing apparatus 100 may acquire a prediction model generated in an external server or the like from the external server.

[0036] The data input unit 103 inputs the behavior data to the actor prediction unit 104 (step S1). The actor prediction unit 104 inputs the behavior data input in step S1 to the prediction model and predicts the actor from the behavior data (step S2). The information display unit 105 displays the prediction result of step S2 on the display device 150 (step S3). In step S3, the information display unit 105 displays a list of persons predicted as the actor on the display device 150. When the user selects an actor from the list, the information display unit 105 may display detailed information of the selected actor on the display device 150.

[0037] In this embodiment, the model generation unit 102 learns a plurality of behavior information and generates a prediction model for predicting an actor from behavior data in which the person who performed the behavior is not specified. The actor prediction unit 104 predicts an actor related to the input behavior data from the input behavior data. In this embodiment, the actor prediction unit 104 can predict, as an actor, an actor having behavior characteristics similar to the input behavior data among the actors who have acted in the past. Therefore, this embodiment can predict an actor for a behavior in which the person is not specified and provide the user with information on the predicted actor.

[0038] The above information providing device 100 can be used, for example, in a security company. In a security company, a user can input behavior data of a behavior that may occur in the future and is related to a nuisance behavior to the information providing device 100. The actor prediction unit 104 inputs the behavior data to the prediction model and predicts the actor, that is, the person who performs the behavior of the behavior data. The information display unit 105 displays a list of the predicted persons.

[0039] When the user selects a specific person from the list, the information display unit 105 displays detailed information on the behavioral characteristics of that person on the display device 150. The information display unit 105 displays, for example, information such as the appearance area, appearance time zone, and purpose of the action of the selected person on the display device 150. By referring to the detailed information, the user can predict in what areas, time zones, and for what purposes the actions are carried out by the selected person. In this case, the security company can focus on marking the person predicted as the actor in the security operations in order to prevent nuisance acts. Also, since the security company can refer to the information of people who have taken similar actions in the past, it can efficiently formulate countermeasures and the like.

[0040] Next, the hardware configuration of the information providing device 100 will be described. FIG. 7 shows a configuration example of a computer device that can be used as the information providing device 100. The computer device 500 includes a processor 510 such as a CPU (Central Processing Unit), a storage unit 520, a ROM (Read Only Memory) 530, a RAM (Random Access Memory) 540, a communication interface (IF: Interface) 550, and a user interface 560.

[0041] The communication interface 550 is an interface for connecting the computer device 500 to a communication network via wired communication means, wireless communication means, or the like. The user interface 560 includes a display unit such as a display, for example. Also, the user interface 560 includes an input unit such as a keyboard, a mouse, and a touch panel.

[0042] The storage unit 520 is an auxiliary storage device that can hold various data. The storage unit 520 can be used as the product information DB 110. The storage unit 520 does not necessarily have to be a part of the computer device 500, and it may be an external storage device or a cloud storage connected to the computer device 500 via a network.

[0043] The ROM 530 is a non-volatile memory device. A semiconductor memory device such as a flash memory with relatively small capacity, for example, is used for the ROM 530. The program executed by the CPU 510 may be stored in the storage unit 520 or the ROM 530. The storage unit 520 or the ROM 530 stores a program for realizing the functions of each part of the information providing apparatus 100.

[0044] When the above program is read into a computer, it includes a group of instructions (or software code) for causing the computer to perform one or more functions described in the embodiment. The program may be stored in a non-transitory computer-readable medium or a tangible storage medium. By way of example and not limitation, the computer-readable medium or tangible storage medium includes random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drive (SSD) or other memory technologies, Compact Disc (CD), digital versatile disc (DVD), Blu-ray (registered trademark) disc or other optical disc storage, magnetic cassette, magnetic tape, magnetic disc storage or other magnetic storage devices. The program may be transmitted on a transitory computer-readable medium or communication medium. By way of example and not limitation, the transitory computer-readable medium or communication medium includes electrical, optical, acoustic, or other forms of propagated signals.

[0045] The RAM 540 is a volatile memory device. Various semiconductor memory devices such as DRAM (Dynamic Random Access Memory) or SRAM (Static Random Access Memory) are used for the RAM 540. The RAM 540 can be used as an internal buffer for temporarily storing data and the like. The CPU 510 expands and executes the program stored in the storage unit 520 or the ROM 530 in the RAM 540. By the CPU 510 executing the program, the functions of each part within the information providing apparatus 100 can be realized. The CPU 510 may have an internal buffer capable of temporarily storing data and the like.

[0046] Note that in the present disclosure, the information providing apparatus 100 does not necessarily have to be physically configured as one device. The information providing apparatus 100 may be configured using a plurality of physically separated devices. For example, the information providing apparatus 100 may have a configuration including a first device having the behavior information acquisition unit 101 and the model generation unit 102, and a second device having the actor prediction unit 104 and the information display unit 105.

[0047] As described above, the present disclosure has been described with reference to the embodiments, but the present disclosure is not limited to the above-described embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Each embodiment can be combined with other embodiments as appropriate.

[0048] The figures are merely illustrative for explaining one or more embodiments. Each figure may be associated with not only one specific embodiment but also one or more other embodiments. As can be understood by those skilled in the art, various features or steps described with reference to any one figure can be combined with features or steps shown in one or more other figures to create, for example, embodiments not explicitly illustrated or described. Not all of the features or steps shown in any one figure for explaining exemplary embodiments are necessarily essential, and some features or steps may be omitted. The order of the steps described in any figure may be changed as appropriate.

[0049] Some or all of the above embodiments may be described as follows, but are not limited thereto.

[0050] [Appendix 1] Using a prediction model generated by learning a plurality of action information each including a person and information of a plurality of items, predicting a person related to the action data from the action data including information corresponding to at least a part of the information of the plurality of items, A program for causing a computer to execute a process including displaying the result of the prediction on a display device.

[0051] [Appendix 2] The predicting in the program according to Appendix 1 includes outputting, as the result of the prediction, a reason why the predicted person is predicted with respect to the action data.

[0052] [Appendix 3] The predicting in the program according to Appendix 1 or 2 includes outputting, as the result of the prediction, a score indicating the probability that the predicted person is the actor of the action data.

[0053] [Appendix 4] The program according to Supplementary Note 3, wherein displaying the result of the prediction includes displaying a plurality of persons included in the result of the prediction in an order according to the score.

[0054] [Supplementary Note 5] The program according to any one of Supplementary Notes 1 to 4, wherein displaying the result of the prediction includes displaying detailed information of the predicted person on the display device.

[0055] [Supplementary Note 6] The program according to any one of Supplementary Notes 1 to 5, further causing the computer to execute a process of learning the plurality of pieces of behavior information and generating the prediction model.

[0056] [Supplementary Note 7] The program according to any one of Supplementary Notes 1 to 6, wherein the information of the plurality of items includes information related to characteristics of behavior.

[0057] [Supplementary Note 8] The program according to Supplementary Note 7, wherein the information related to the characteristics of the behavior includes at least one of a location, a means of transportation, a tool, and a time zone of the behavior.

[0058] [Supplementary Note 9] The program according to any one of Supplementary Notes 1 to 8, wherein the behavior information is generated from information on behaviors performed in the past.

[0059] [Supplementary Note 10] Predicting a person related to the behavior data from behavior data including information on at least a part of the plurality of items, using a prediction model generated by learning a plurality of pieces of behavior information each including a person and information on a plurality of items, An information providing method including displaying the result of the prediction on a display device.

[0060] [Supplementary Note 11] Using a prediction model generated by learning a plurality of behavior information each including a person and information of a plurality of items, a prediction unit that predicts a person associated with the behavior data from behavior data including information of at least a part of the plurality of items, An information providing apparatus comprising an information display unit that displays the prediction result of the prediction unit on a display device.

[0061] Some or all of the elements (e.g., configurations and functions) described in Appendices 2 to 9 that are subordinate to Appendix 1 may be subordinate to Appendices 10 and 11 in the same subordinate relationship as Appendices 2 to 8. Some or all of the elements described in any appendix may be applied to various hardware, software, recording means for recording software, systems, and methods.

Explanation of Reference Numerals

[0062] 10: Information providing apparatus 11: Prediction unit 12: Information display unit 100: Information providing apparatus 101: Behavior information acquisition unit 102: Model generation unit 103: Data input unit 104: Actor prediction unit 105: Information display unit 130: Behavior information DB 150: Display device

Claims

1. Using a prediction model generated by learning a plurality of behavior information each including a person and information on a plurality of items, predicting a person associated with the behavior data from the behavior data including information corresponding to at least a part of the information on the plurality of items, A program for causing a computer to execute a process including displaying the result of the prediction on a display device.

2. The program according to claim 1, wherein the predicting includes outputting, as the result of the prediction, a reason why the predicted person is predicted with respect to the behavior data.

3. The program according to claim 1 or 2, wherein the predicting includes outputting, as the result of the prediction, a score indicating the probability that the predicted person is the actor of the behavior data.

4. The program according to claim 3, wherein the displaying the result of the prediction includes displaying a plurality of persons included in the result of the prediction in an order according to the score.

5. The program according to claim 1 or 2, wherein the displaying the result of the prediction includes displaying detailed information on the predicted person on the display device.

6. The program according to claim 1 or 2, further causing the computer to execute a process of learning the plurality of behavior information and generating the prediction model.

7. The program according to claim 1 or 2, wherein the information on the plurality of items includes information related to characteristics of the behavior.

8. The program according to claim 7, wherein the information related to the characteristics of the behavior includes at least one of a location, a means of movement, a tool, and a time zone of the behavior.

9. Using a prediction model generated by learning a plurality of behavior information each including a person and information on a plurality of items, predicting a person associated with the behavior data from the behavior data including at least a part of the information on the plurality of items, An information providing method including displaying the result of the prediction on a display device.

10. A prediction unit that predicts a person associated with behavior data from the behavior data including at least a part of the information on a plurality of items, using a prediction model generated by learning a plurality of behavior information each including a person and information on the plurality of items, An information providing apparatus including an information display unit that displays the prediction result of the prediction unit on a display device.

Images