Information processing device, information processing system, and information processing method

Abstract

An information processing device (100) comprises: an acquisition unit (132) that acquires, from a memory unit (120) storing episode data (D1) of a speaker, episode data (D1) related to topic information included in speech data (D30) of the speaker; and an interaction control unit (133) that controls interactions with the speaker so as to include an episode based on the episode data (D1).

Description

Technical Field

[0004] ,

[0006] , , , , , ,

[0005] , , , If the acquisition unit recognizes the speaker, it acquires the episode data that satisfies the acquisition conditions. , , ,

[0001] The present disclosure relates to an information processing apparatus, an information processing system, and an information processing method.

Background Art

[0002] In recent years, devices that perform speech recognition of user speech and conduct conversations based on the recognition results have become widespread. For example, Patent Document 1 discloses a technique for realizing conversations according to a specific field using a dictionary for each field.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the prior art, since it conducts specialized conversations in the fields recorded in the dictionaries for each field, it is difficult to realize daily conversations. For this reason, in the prior art, it has been desired to realize conversations with which users feel a sense of familiarity.

[0005] Therefore, the present disclosure proposes an information processing apparatus, an information processing system, and an information processing method capable of realizing conversations with which users feel a sense of familiarity.

Means for Solving the Problems

[0006] In order to solve the above problems, an information processing apparatus according to one aspect of the present disclosure includes an acquisition unit that acquires, from a storage unit that stores episode data of a speaker, the episode data regarding topic information included in the utterance data of the speaker, a dialogue control unit that controls a dialogue with the speaker so as to include an episode based on the episode data, and a recognition unit that recognizes the speaker, knowledge and is provided with If the acquisition unit recognizes the speaker, it acquires the episode data that satisfies the acquisition conditions. the dialogue control unit is The episode data acquired by the acquisition unit is to be spontaneously spoken, Control the dialogue with the aforementioned speaker. ru .

[0007] Furthermore, one form of information processing system relating to this disclosure is an information processing system comprising an agent device for collecting speaker utterance data and an information processing device, wherein the information processing device includes an acquisition unit for acquiring episode data relating to topic information included in the utterance data from a storage unit for storing the speaker's episode data, a dialogue control unit for controlling a dialogue with the speaker to include an episode based on the episode data, and an information processing device for recognizing the speaker. knowledge It comprises a recognition unit, If the acquisition unit recognizes the speaker, it acquires the episode data that satisfies the acquisition conditions. The aforementioned dialogue control unit, The episode data acquired by the acquisition unit is to be spontaneously spoken, Control the dialogue with the aforementioned speaker. ru .

[0008] Furthermore, one form of information processing method relating to this disclosure involves a computer acquiring episode data relating to topic information contained in the speaker's utterance data from a storage unit that stores the speaker's episode data, controlling the interaction with the speaker to include an episode based on the episode data, and recognizing the speaker. knowledge Including doing, The acquisition described above means that if the speaker is recognized, the episode data that satisfies the acquisition conditions is acquired. The aforementioned control means The episode data acquired by the acquisition unit is to be spontaneously spoken, Control the dialogue with the aforementioned speaker. ru . [Brief explanation of the drawing]

[0009] [Figure 1] This is a diagram illustrating an example of an information processing system according to an embodiment. [Figure 2] This figure shows an example of the configuration of an agent device according to the embodiment. [Figure 3] This figure shows an example of the structure of the agent device according to this embodiment. [Figure 4] This figure shows an example of the configuration of an information processing device according to the embodiment. [Figure 5] This figure shows an example of episode data according to the embodiment. [Figure 6]It is a diagram showing an example of template data related to an embodiment. [Figure 7] It is a diagram for explaining an example of generation of episode data based on utterance data according to an embodiment. [Figure 8] It is a diagram for explaining an example of generation of episode data based on utterance data according to an embodiment. [Figure 9] It is a flowchart showing an example of a processing procedure (1) of an information processing apparatus according to an embodiment. [Figure 10] It is a flowchart showing an example of a processing procedure (2) of an information processing apparatus according to an embodiment. [Figure 11] It is a diagram showing an example of question data according to an embodiment. [Figure 12] It is a flowchart showing a processing procedure of an information processing apparatus related to generation of episode data using questions. [Figure 13] It is a diagram showing an example of generation of episode data using questions of an information processing apparatus according to an embodiment. [Figure 14] It is a flowchart showing a processing procedure related to generation of episode data based on related information of an information processing apparatus. [Figure 15] It is a diagram showing an example of use of episode data based on related information of an information processing apparatus according to an embodiment. [Figure 16] It is a diagram for explaining generalization of a plurality of episode data. [Figure 17] It is a hardware configuration diagram showing an example of a computer that realizes the functions of an information processing apparatus.

Mode for Carrying Out the Invention

[0010] Hereinafter, embodiments of the present disclosure will be described in detail based on the drawings. In each of the following embodiments, the same parts are denoted by the same reference numerals, and redundant descriptions are omitted.

[0011] (Embodiment) [Outline of Information Processing System According to Embodiment] Figure 1 is a diagram illustrating an example of an information processing system according to an embodiment. The information processing system 1 shown in Figure 1 is a dialogue processing system that realizes dialogue between user U and agent device 10. User U is an example of a speaker of agent device 10 in information processing system 1. For example, user U may become bored if the dialogue with agent device 10 is not original or novel. User U may stop using agent device 10 if they become bored with it. Therefore, information processing system 1 has an episode talk function that realizes dialogue based on user U's episodes so that user U does not get bored with the dialogue with agent device 10 and feels a sense of familiarity with it.

[0012] Information processing system 1 comprises an agent device 10 and an information processing device 100. Information processing system 1 may include, for example, multiple agent devices 10 that interact with each of multiple users U. The agent devices 10 and the information processing device 100 are configured to communicate, for example, via a network or directly without a network. The information processing device 100 may, for example, be incorporated into the agent device 10. In the following description, an example will be described in which the agent devices 10 and the information processing device 100 cooperate to enable the agent device 10 to interact with user U.

[0013] The agent device 10 is a so-called IoT (Internet of Things) device and performs various information processing in cooperation with the information processing device 100. The speech recognition and voice-based dialogue processing performed by the agent device 10 are sometimes referred to as agent functions. The agent device 10 can, for example, provide various responses to the utterances of user U. For example, responses include episode-based responses, scenario-based responses, responses based on the verbalization of the situation, knowledge-based responses, and casual conversation responses.

[0014] The agent device 10 includes, for example, an autonomous mobile device, a smartphone, a tablet device, a game console, a home speaker, a household electrical appliance, an automobile, etc. Smartphones, tablet devices, etc., realize the agent function by executing a program (application) that has agent functionality. In this embodiment, the case in which the agent device 10 is an autonomous mobile device will be described.

[0015] In one example shown in Figure 1, the agent device 10 can be various devices that perform autonomous actions based on environmental perception. The agent device 10 is an elongated ellipsoidal agent-type robot device that performs autonomous movement using wheels. The agent device 10 includes, for example, two wheels and a drive mechanism for driving the two wheels. The agent device 10 controls the drive mechanism to move while maintaining an upright position. The agent device 10 enables various forms of communication with user U by performing autonomous actions according to the user U, the surroundings, and its own situation. The agent device 10 may be, for example, a small robot that is small and light enough for user U to easily lift with one hand.

[0016] In this embodiment, the agent device 10 performs information processing on the collected audio. For example, the agent device 10 recognizes the utterance of user U and performs information processing on that utterance. In the example shown in Figure 1, the agent device 10 is assumed to be installed in the user U's home, company, school, etc.

[0017] The agent device 10 may have various sensors for acquiring not only ambient sounds but also other types of information. For example, in addition to a microphone, the agent device 10 may have a camera for acquiring spatial information, an illuminance sensor for detecting light intensity, a gyro sensor for detecting tilt, an infrared sensor for detecting objects, and so on.

[0018] The information processing device 100 is a so-called cloud server and is a server device that performs information processing in cooperation with the agent device 10. The information processing device 100 acquires the speech data of user U collected by the agent device 10 and analyzes the speech data using natural language understanding (NLU), natural language processing (NLP), etc. When the information processing device 100 extracts topic information from the speech data, it acquires episode data related to that topic information from the episode database (episode DB). The topic information includes, for example, information indicating the classification of the episode, words such as keywords, strings, etc. In other words, the topic information is information for acquiring episode data. The episode database, for example, contains episode data related to user U and is stored in a memory device. The information processing device 100 has a function to interact with user U so as to include episodes based on the episode data.

[0019] In the example shown in Figure 1, when the agent device 10 recognizes that user U was absent, it issues an utterance C11 instructed by the information processing device 100. The utterance C11 is, for example, "Welcome back." User U responds to utterance C11 with utterance C12. The utterance C12 is, for example, "I was at the office working." The agent device 10 transmits the utterance data of utterance C12 to the information processing device 100.

[0020] The information processing device 100 analyzes the speech data of utterance C12 and extracts topic information for "company" and "I went." The information processing device 100 obtains episode data corresponding to the extracted topic information from the episode DB and instructs the agent device 10 to make an utterance C13 based on the episode data. For example, the information processing device 100 obtains episode data related to "company" and "go." In this case, the episode data is, for example, data indicating episodes about the speaker going to company yesterday and the day before yesterday. The agent device 10 makes an utterance C13 to user U. For example, utterance C13 is "Oh, really? You go every day." The method of dialogue based on episode data will be described later. User U makes an utterance C14 in response to utterance C13. For example, utterance C14 is "That's what work is." The agent device 10 transmits the speech data of utterance C14 to the information processing device 100.

[0021] The information processing device 100 analyzes the speech data of utterance C14 and controls the agent device 10 to utter a response to user U. For example, the information processing device 100 generates speech data suitable for the character of agent device 10 based on a dialogue model, dialogue scenario, etc., for responding to the speech data. In the example shown in Figure 1, agent device 10 utters utterance C15 in response to the speech data of utterance C14. Utterance C15 is, for example, "Do your best for me too."

[0022] As described above, when the information processing device 100 according to this embodiment acquires episode data relating to topic information contained in the utterance data of user U, it can cause the agent device 10 to execute a dialogue that includes user U's episode based on said episode data. As a result, the information processing device 100 can provide episodes based on user U's episode data in the dialogue with user U. Consequently, by including user U's episode in the dialogue, the information processing device 100 can realize a dialogue that user U finds familiar and engaging.

[0023] [Example configuration of an agent device according to the embodiment] Figure 2 shows an example of the configuration of the agent device 10 according to the embodiment. As shown in Figure 2, the agent device 10 includes a sensor unit 11, an input unit 12, a light source 13, an output unit 14, a drive unit 15, a control unit 16, and a communication unit 17.

[0024] The sensor unit 11 has the function of collecting various sensor information related to the user U and the surroundings. The sensor unit 11 according to this embodiment includes, for example, a camera, a ToF (Time of Flight) sensor, a microphone, an inertial sensor, etc. The sensor unit 11 may also include various sensors such as a geomagnetic sensor, a touch sensor, an infrared sensor, a temperature sensor, a humidity sensor, etc. The sensor unit 11 supplies the collected sensor information to the control unit 16. The sensor unit 11 has the function of collecting sound, etc., using a microphone. The sensor unit 11 can store the collected sound, etc., in a storage device.

[0025] The input unit 12 has the function of detecting physical input operations performed by user U. The input unit 12 includes, for example, a button such as a power switch. The input unit 12 supplies input information indicating the detected input operation to the control unit 16.

[0026] The light source 13 represents the eye movements of the autonomous mobile agent device 10. The light source 13 has, for example, two eye sections. The light source 13 represents a wide range of eye movements according to the situation and actions, based on instructions from the control unit 16.

[0027] The output unit 14 has the function of outputting various sounds, including voice. The output unit 14 includes, for example, a speaker or an amplifier. The output unit 14 outputs the sound instructed by the control unit 16.

[0028] The drive unit 15 expresses movement by driving the drive mechanism of the agent device 10, which is an autonomous mobile unit. The drive unit 15 includes, for example, two wheels or multiple motors. The drive unit 15 is driven by instructions from the control unit 16.

[0029] The control unit 16 controls the agent device 10. The control unit 16 has the function of controlling each component of the agent device 10. For example, the control unit 16 controls the starting and stopping of each component. Based on control information from the information processing device 100, the control unit 16 controls the light source 13, output unit 14, drive unit 15, etc.

[0030] When the sensor unit 11 collects the user U's speech, the control unit 16 controls the transmission of speech data indicating the speech to the information processing device 100. The control unit 16 controls the output unit 14 to output dialogue data instructed by the information processing device 100. The control unit 16 realizes dialogue with the user U by outputting dialogue data corresponding to the collected speech data.

[0031] The communication unit 17 communicates with, for example, the information processing device 100, other external devices, etc. The communication unit 17 sends and receives various data, for example, via a wired or wireless network. For example, when speech is collected, the communication unit 17 transmits speech information to the information processing device 100. The communication unit 17 may transmit not only speech data but also identification information for identifying user U, etc., to the information processing device 100. For example, the communication unit 17 supplies various data received from the information processing device 100 to the control unit 16.

[0032] The above describes an example of the functional configuration of the agent device 10 according to this embodiment. Note that the above configuration described using Figure 3 is merely an example, and the functional configuration of the agent device 10 according to this embodiment is not limited to this example. The functional configuration of the agent device 10 according to this embodiment can be flexibly modified according to specifications and operation.

[0033] [Example of the structure of an agent device according to an embodiment] Figure 3 shows an example of the structure of the agent device 10 according to this embodiment. The left figure in Figure 3 is a side view showing the posture of the agent device 10 in an upright position (including when stationary and when moving). The right figure in Figure 3 is a side view showing the posture of the agent device 10 in a seated position.

[0034] As shown in the right-hand diagram of Figure 3, in this embodiment, when the agent device 10 is stationary in a seated position, a portion of its bottom touches the floor surface. As a result, at least three points—the two wheels 570 and the bottom that is in contact with the floor—are in contact with the floor surface, and the agent device 10 is seated on three points. Therefore, the agent device 10 can be stably stationary in a seated position. Furthermore, when the agent device 10 is moving in an upright position, its bottom does not touch the floor surface.

[0035] The center of gravity CoG of the agent device 10 is located on the perpendicular line V1 above the axle of the wheel 570 when the agent device 10 is in a forward-leaning position (standing position) (see the left diagram in Figure 3). As a result, the agent device 10 maintains balance and remains in an upright position.

[0036] On the other hand, in a seated position, as shown in the right diagram of Figure 3, the agent device 10 is tilted backward, causing at least three points—the two wheels 570 and the bottom (protrusion 701)—to contact the floor. At this time, the center of gravity CoG of the agent device 10 is located between the perpendicular V1 passing through the axle of the wheel 570 and the perpendicular V2 passing through the contact point between the bottom (protrusion 701) and the floor. By positioning the center of gravity CoG of the agent device 10 between the perpendicular V1 passing through the axle of the wheel 570 and the perpendicular V2 passing through the contact point between the bottom (protrusion 701) and the floor, the agent device 10 can be kept stably stationary in a seated position.

[0037] [Example configuration of an information processing device according to the embodiment] Figure 4 shows an example of the configuration of an information processing device 100 according to an embodiment. As shown in Figure 4, the information processing device 100 comprises a communication unit 110, a storage unit 120, and a control unit 130. The control unit 130 is electrically connected to the communication unit 110 and the storage unit 120.

[0038] The communication unit 110 communicates with, for example, the agent device 10, other external devices, etc. The communication unit 110 sends and receives various types of data, for example, via a wired or wireless network. The communication unit 110 supplies data received from, for example, the agent device 10 to the control unit 130. The communication unit 110 transmits data instructed by the control unit 130 to the instructed destination.

[0039] The memory unit 120 stores various data and programs. For example, the memory unit 120 may be a semiconductor memory element such as RAM (Random Access Memory) or flash memory, a hard disk, or an optical disk. The memory unit 120 stores information received via the communication unit 110. The memory unit 120 stores various data such as episode data D1, template data D2, user data D3, question data D4, dialogue data D5, knowledge data D10, management data D20, and utterance data D30.

[0040] Episode data D1 is data that shows episodes of user U. Episodes include, for example, information about user U's past events, user U's future plans, and user U's hobbies. Template data D2 is data that shows the algorithm for generating dialogue data D5 from episode data D1. User data D3 is data that shows information about user U, which includes, for example, user U's personal information, authentication information, and settings information. Question data D4 is data that shows questions, etc., for eliciting episodes from user U. Question data D4 is data used when generating episode data D1, etc. Dialogue data D5 is data used for dialogue with user U. Knowledge data D10 is data for realizing dialogue based on knowledge, scenarios, etc. Knowledge data D10 includes, for example, information such as intent identification, common sense knowledge, specialized knowledge, and corpus. Management data D20 is data that shows the correspondence between user U and the agent device 10 used by user U. Speech data D30 is speech data from agent device 10, and the analysis results are associated with it.

[0041] In this embodiment, the information processing device 100 realizes an episode database by storing and managing the episode data D1 in the storage unit 120. The information processing device 100 may also be configured to store the episode database in an external storage device or the like.

[0042] The control unit 130 is, for example, a dedicated or general-purpose computer. The control unit 130 is, for example, an integrated control unit that controls the information processing device 100. The control unit 130 works in conjunction with the agent device 20 to provide various functional units that enable the agent device 10 to interact with the user U.

[0043] The control unit 130 comprises the following functional units: a recognition unit 131, an acquisition unit 132, an interaction control unit 133, a generation unit 134, a data collection unit 135, and an operation control unit 140. Each functional unit of the control unit 130 is implemented, for example, by a CPU (Central Processing Unit) or MPU (Micro Processing Unit) executing a program stored inside the information processing device 100 using RAM or the like as a working area. Alternatively, each functional unit may be implemented by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field-Programmable Gate Array).

[0044] The recognition unit 131 has the function of authenticating user U based on sensor information from the agent device 10. The recognition unit 131 performs personal authentication, for example, by facial recognition, voice recognition, fingerprint recognition, etc., and recognizes user U with whom the agent device 10 is interacting. If the recognition unit 131 authenticates a user U available to the agent device 10 based on the sensor information from the agent device 10, user data D3, management data D20, etc., it enables interaction between the user U and the agent device 10.

[0045] The recognition unit 131 converts the speech data received from the agent device 10 into a string. For example, the recognition unit 131 can use automatic speech recognition (ASR) processing. The recognition unit 131 analyzes the recognized string using NLU, NLP processing, etc. In this embodiment, the recognition unit 131 analyzes the dependency structure of the string and stores the analysis information indicating the analysis result in the storage unit 120 in association with the speech data D30. The analysis information includes, for example, information indicating the dependency structure such as phrases and cases.

[0046] The acquisition unit 132 acquires episode data D1 related to topic information contained in user U's speech data D30 from the storage unit 120, which stores user U's episode data D1. The acquisition unit 132 extracts topic information from the speech data D30, for example. The topic information is information for acquiring user U's episodes and includes pre-set keywords, words, strings, and other information.

[0047] The acquisition unit 132 acquires episode data D1 whose topic information satisfies the dialogue conditions. The dialogue conditions are set in, for example, user data D3, management data D20, etc. The dialogue conditions include, for example, conditions for extracting episode data D1 based on topic information. For example, if the condition is that the words or strings contained in episode data D1 match the topic information in a set number of cases or more, the acquisition unit 132 acquires episode data D1 that satisfies the dialogue conditions from the storage unit 120.

[0048] If the acquisition unit 132 recognizes user U, it acquires episode data D1 that satisfies the acquisition conditions, regardless of user U's utterances. The acquisition conditions include, for example, conditions for acquiring episode data D1 that indicates user U's most recent episode, time-of-day episodes, date or season episodes, etc. The most recent episode is, for example, user U's latest episode. Time-of-day episodes include, for example, episodes related to today's schedule, yesterday's events, etc., if the time-of-day episode is morning. Time-of-day episodes include, for example, episodes related to afternoon's schedule, evening's schedule, etc., if the time-of-day episode is nighttime. Time-of-day episodes include, for example, episodes related to today's events, tomorrow's schedule, etc. Date or season episodes include, for example, episodes within a set period. The acquisition unit 132 supplies the acquired episode data D1 to the dialogue control unit 133. If the acquisition unit 132 is unable to acquire episode data D1, it notifies the dialogue control unit 133 that it could not acquire it.

[0049] The dialogue control unit 133 controls the dialogue with user U (speaker) so that it includes an episode based on the episode data D1 acquired by the acquisition unit 132. Controlling the dialogue includes, for example, controlling it to make utterances to user U, and controlling it to talk face-to-face with user U. For example, the dialogue control unit 133 generates dialogue data D5 including an episode and controls the agent device 10 to output audio based on the dialogue data D5. The dialogue control unit 133 instructs the agent device 10 to output audio of the dialogue data D5 via the communication unit 110. As a result, the agent device 10 realizes dialogue with user U by outputting audio based on the dialogue data D5 from the output unit 14.

[0050] The dialogue control unit 133 generates dialogue data D5, which includes an episode, based, for example, on episode data D1 and an algorithm for generating dialogue text. The algorithm can be implemented, for example, by a template, application, program, etc. In this embodiment, the case in which dialogue data D5 is generated using template data D2 will be described.

[0051] The dialogue control unit 133 generates dialogue data D5 by setting the data of episode data D1 as a template. The dialogue control unit 133 controls the dialogue with user U so that the agent device 10 spontaneously speaks the episode. For example, when the dialogue control unit 133 recognizes user U, it controls the agent device 10 to emit audio of the episode based on episode data D1 that satisfies the acquisition conditions.

[0052] If the acquisition unit 132 has not acquired episode data D1, the dialogue control unit 133 controls the dialogue with user U based on dialogue data D5 which is different from the episode. That is, the dialogue control unit 133 generates dialogue data D5 that does not include an episode based on knowledge data D10 and controls the agent device 10 to output speech based on the dialogue data D5. The dialogue data D5 that does not include an episode is data for making responses based on knowledge data D10, such as scenario dialogue, situation verbalization, knowledge-based dialogue, and casual conversation responses. Scenario dialogue includes, for example, responses according to a pre-designed scenario. Situation verbalization includes, for example, responses based on a corpus for spontaneous speaking, situation definitions, etc., according to the current situation. Knowledge-based dialogue includes, for example, personalized responses based on knowledge data D10. The dialogue control unit 133 selects scenario dialogue, situation verbalization, knowledge-based dialogue, and casual conversation responses according to the dialogue with user U and generates dialogue data D5 that shows the selected knowledge-based response.

[0053] If the dialogue control unit 133 has authenticated user U of the agent device 10, it controls the dialogue with user U based on the question data D4 in order to elicit an episode from user U. The dialogue control unit 133 can, for example, select question data D4 according to the time of day. The dialogue control unit 133 controls the agent device 10 to output audio based on the question data D4.

[0054] The generation unit 134 generates episode data D1 based on the analysis results of the speech data D30. The generation unit 134 generates episode data D1 based, for example, on the analysis results of the speech data D30 and the state and status of the agent device 10. By focusing on the state and status of the agent device 10, the generation unit 134 generates episode data D1 based on information that cannot be recognized from the analysis results of the speech data D30.

[0055] The generation unit 134 generates episode data D1 based on the question data D4 that the dialogue control unit 133 asks the user U, and the analysis result of the utterance data D30 corresponding to that question. For example, the generation unit 134 generates episode data D1 by combining the content of the question data D4 with the user U's response. For example, if the question data D4 is "Where are you going today?", the generation unit 134 can generate episode data D1 that shows the user's episode for the day based on the user U's response to the question. In other words, the generation unit 134 can generate episode data D1 based on the response elicited from the user U by the question. The generation unit 134 stores the generated episode data D1 in the storage unit 120, associating it with the user U.

[0056] The collection unit 135 collects relevant information about the recognized user U. This relevant information includes, for example, information indicating user U's past situations. For example, the collection unit 135 collects information indicating the situation of user U recognized based on sensor information from the agent device 10. For example, the collection unit 135 collects relevant information indicating, for example, "Suzuki came to the house yesterday." For example, the collection unit 135 collects relevant information indicating, for example, "Mom was watching TV over the weekend." The collection unit 135 can also collect relevant information such as user U's schedule and user U's information on social networks. The collection unit 135 supplies the collected relevant information to the generation unit 134.

[0057] The generation unit 134 generates episode data D1 for user U based on the collected relevant information and stores it in the storage unit 120. For example, the generation unit 134 analyzes the relevant information and generates episode data D1 based on information about people, animals, etc., related to user U.

[0058] The motion control unit 140 controls the operation of the agent device 10. The motion control unit 140 controls the operation of the agent device 10 based on an action plan based on the user U's status, learned knowledge, etc. The motion control unit 140 controls the agent device 10 to move while maintaining its upright position, for example, the agent device 10 having an elongated ellipsoid shape. The movement includes, for example, forward and backward movement, turning movement, rotational movement, etc. The motion control unit 140 controls the agent device 10 to actively perform inducement actions that induce communication between the user U and the agent device 10. Inducement actions include, for example, actions to induce dialogue between the user U and the agent device 10.

[0059] The above describes an example of the functional configuration of the information processing device 100 according to the embodiment. Note that the above configuration described using Figure 4 is merely an example, and the functional configuration of the information processing device 100 according to the embodiment is not limited to this example. The functional configuration of the information processing device 100 according to the embodiment can be flexibly modified according to specifications and operation.

[0060] [Example of episode data structure] Figure 5 shows an example of episode data D1 according to the embodiment. As shown in Figure 5, the information processing device 100 has multiple episode data D1 related to user U. In the example shown in Figure 5, the case where there are six episode data D1 is described, but the number of episode data D1 will differ depending on the number created and registered. Episode data D1 has items such as Ep_id, When, Who, Where, Action, State, Target, Why, How, With, Cause, etc.

[0061] The Ep_id field contains information that identifies user U's episode data D1. The When field contains information about the date, time, season, etc., of episode data D1. The Who field contains information about the name, proper noun, etc., of episode data D1. The Where field contains information about the location of episode data D1. The Action field contains information about the predicate of episode data D1. The State field contains information about the state of episode data D1. The Target field and Why field contain information about the cause of episode data D1. The How field contains information about the means, methods, procedures, etc., of episode data D1. The With field contains information about the person, animal, object, etc., involved in episode data D1. The Cause field contains information about the cause of episode data D1. In other words, the Cause field contains information that links episode data D1 together. Episode data D1 has multiple fields, and information is set in the fields that correspond to the episode, while fields that do not correspond are left blank.

[0062] In the example shown in Figure 5, the information processing device 100 manages six episode data D1s, from Ep1 to Ep6, as episode data D1 related to user U. Episode data D1 with the Ep_id field set to Ep1 represents the episode, "Yesterday, Tanaka petted aibo at home." Episode data D1 for Ep1 has "Ep2" set in the Cause field, indicating that it is the cause of episode data D1 for Ep2. Episode data D1 with the Ep_id field set to Ep2 represents the episode, "Yesterday, Aibo was happy at home." In other words, episode data D1 for Ep1 and Ep2 represents the episode where Aibo was happy because Tanaka petted it at home yesterday. Episode data D1 with the Ep_id field set to Ep3 represents the episode, "This morning, Tanaka was hungry, so he hurriedly ate curry with Yamada." Episode data D1 for Ep3 has a blank Cause field, indicating that it is not related to the other episode data D1s.

[0063] Episode data D1 with the Ep_id field set to Ep4 represents the episode "Yamada runs to the toilet at full speed." Episode data D1 for Ep4 is caused by "Ep5" in the Cause field, indicating that it is the cause of episode data D1 for Ep5. Episode data D1 with the Ep_id field set to Ep5 represents the episode "Yamada falls down." In other words, episode data D1 for Ep4 and Ep5 represents the episode where Yamada falls down because he ran to the toilet at full speed. Episode data D1 with the Ep_id field set to Ep6 represents a future episode "Tomorrow, everyone goes to a hot spring in Izu." Episode data D1 for Ep6 is not related to any other episode data D1 because the Cause field is blank.

[0064] In this embodiment, for the sake of simplicity, the "When" item in episode data D1 will be described as being set to information such as "yesterday" or "this morning," but it may also be set to date and time information. In this case, when the information processing device 100 creates dialogue data D5 based on episode data D1, it should replace the date and time with "yesterday," "the day before yesterday," "tomorrow," etc.

[0065] The configuration described above using Figure 5 is merely an example, and the data structure of episode data D1 according to this embodiment is not limited to this example. The data structure of episode data D1 according to this embodiment can be flexibly modified according to specifications and operation. For example, episode data D1 may have items such as the purpose and destination related to the episode.

[0066] [Example data structure for template data] Figure 6 shows an example of template data D2 according to the embodiment. As shown in Figure 6, template data D2 has items such as classification and template. Classification includes, for example, episode classifications such as past, future, and conversation (currently recognized). Template has data for performing dialogue based on the items of episode data D1. Template is an example of an algorithm. In this embodiment, the case in which the information processing device 100 generates dialogue data D5 using a template is described, but dialogue data D5 may also be generated using an application, module, etc. for executing an algorithm.

[0067] In one example shown in Figure 6, template data D2 classified as past (1) shows the algorithm "(when)(who)(where+de)(why+kara)(with) and(target)(how)(action+ta) yo ne". Template data D2 classified as past (2) shows the algorithm "(when)(who)(where+de)(why+kara)(with) and(target)(how)(action+te)(state+ta) yo ne". Template data D2 classified as future shows the algorithm "(when)(who)(where+de)(why+kara)(with) and(target)(how)(action terminal form) n desho". Template data D2 classified as conversational shows the algorithm "(when)(who)(where+de)(why+kara)(with) and(target)(how)(action terminal form) ne".

[0068] The information processing device 100 generates dialogue data D5 representing an episode by setting the information of the items set in episode data D1 to the corresponding items in the algorithm of template data D2. For example, in the case of episode data D1 of Ep1 shown in Figure 5, the information processing device 100 uses past template data D2 to generate dialogue data D5 saying, "Yesterday, I petted Aibo at the Tanaka house." The information processing device 100 enables dialogue of episodes appropriate to the character by storing template data D2 corresponding to the character of the agent device 10 in the storage unit 120. The information processing device 100 can enable dialogue of episodes that feel familiar by storing template data D2 corresponding to the language of the region where user U lives in the storage unit 120.

[0069] [Example of generating episode data] Figures 7 and 8 illustrate an example of generating episode data D1 based on speech data D30 according to this embodiment.

[0070] In the example shown in the left diagram of Figure 7, the utterance data 30 has its dependency structure analyzed by syntactic analysis to show that it was "yesterday," "Taro," "at home," "because it was cold," "heating," "immediately," and "turned on." "Yesterday" is a noun phrase and is in the temporal case. "Taro" is a noun phrase and is in the nominative case. "At home" is a noun phrase and is in the causative case. "Because it was cold" is an adjective phrase and is in the adverbial clause case. "Heating" is a noun phrase and is in the object case. "Immediately" is an adverbial phrase and is in the adverbial clause case. "Turned on" is a verb phrase and is in the predicate clause case. In this case, the information processing device 100 generates the episode data D1 shown in the right diagram of Figure 7 based on the analysis results of the utterance data D30 and stores it in the storage unit 120 as the user U's episode data D1. Episode data D1 has "Yesterday" set in the When field, "Taro" in the Who field, "Turn on" in the Action field, "Heating" in the Target field, "Because it's cold" in the Why field, and "Immediately" in the How field, with the other fields left blank.

[0071] In the example shown in the left diagram of Figure 8, the utterance data D30 has its dependency structure analyzed by syntactic analysis to determine that it consists of "last weekend," "with family," "zoo," and "I went." "Last weekend" is a noun phrase and is in the temporal case. "With family" is a noun phrase and is in the causal case. "Zoo" is a noun phrase and is in the accusative case. "I went" is a verb phrase and is in the predicate clause case. In this case, the information processing device 100 generates the episode data D1 shown in the right diagram of Figure 8 based on the analysis results of the utterance data D30 and stores it in the storage unit 120 as the user U's episode data D1. In the episode data D1, "Last weekend" is set in the When field, "Speaker" in the Who field, "Go" in the Action field, and "Zoo" in the Target field, while the other fields are left blank. If the speech data D30 does not contain a nominative case, the Who field in the episode data D1 will be set to the speaker of agent device 10.

[0072] [Example of processing procedure (1) for the information processing apparatus according to the embodiment] Figure 9 is a flowchart of an example of a processing procedure (1) of the information processing device 100 according to the embodiment. The processing procedure shown in Figure 9 is realized by the control unit 130 of the information processing device 100 executing a program. The processing procedure shown in Figure 9 is repeatedly executed when the information processing device 100 controls the interaction of the agent device 10. That is, the processing procedure shown in Figure 9 is repeatedly executed by the information processing device 100, for example, during interaction with user U.

[0073] As shown in Figure 9, the control unit 130 of the information processing device 100 extracts topic information from the analysis results of the speech data D30 (step S101). For example, the control unit 130 extracts words, strings, etc. that satisfy extraction conditions such as when, where, who, what happened, what state it was in, what, with whom, why, how, purpose, and what the result was as topic information. Once the control unit 130 stores the extraction results in the storage unit 120, it proceeds to step S102.

[0074] The control unit 130 acquires episode data D1 related to topic information from the storage unit 120 (step S102). For example, the control unit 130 searches for episode data D1 that has topic information and acquires episode data D1 that satisfies the dialogue conditions from the storage unit 120. The dialogue conditions include, for example, that the number, percentage, etc., of words in the episode data D1 that match the topic information is above a preset threshold. If the control unit 130 acquires multiple episode data D1, it uses the episode data D1 with the highest degree of match or uses episode data D1 that has not been used in past dialogues. Once the control unit 130 stores the acquisition result indicating whether or not episode data D1 has been acquired in the storage unit 120, it proceeds to step S103.

[0075] The control unit 130 determines whether or not episode data D1 has been acquired based on the acquisition results from the storage unit 120 (step S103). If the control unit 130 determines that episode data D1 has been acquired (Yes in step S103), it proceeds to step S104. The control unit 130 generates dialogue data D5 based on episode data D1 and template data D2 (step S104). For example, the control unit 130 acquires template data D2 from the storage unit 120 for a classification corresponding to the information in the "when" item of episode data D1. If the "when" item indicates the past, the control unit 130 acquires template data D2 for a past classification. The control unit 130 generates dialogue data D5 by setting the information set in the items of episode data D1 into the template of the acquired template data D2. Once the control unit 130 stores the generated dialogue data D5 in the storage unit 120, it proceeds to step S106, which will be described later.

[0076] Furthermore, if the control unit 130 determines that it has not acquired episode data D1 (No in step S103), it proceeds to step S105. The control unit 130 generates dialogue data D5 that is different from the episode (step S105). For example, based on the history of past conversations with user U, the control unit 130 selects one of scenario dialogue, situation verbalization, knowledge-based dialogue, and small talk to generate dialogue data D5. If the control unit 130 selects scenario dialogue, it generates dialogue data D5 based on a scenario corresponding to the content of user U's utterance. If the control unit 130 selects situation verbalization, it generates dialogue data D5 corresponding to the current situation recognized via the agent device 10. If the control unit 130 selects knowledge-based dialogue, it generates dialogue data D5 based on the content of user U's utterance and knowledge data D10. If the control unit 130 selects small talk, it generates dialogue data D5 that responds to the content of user U's utterance. Once the control unit 130 stores the generated dialogue data D5 in the storage unit 120, it proceeds to step S106.

[0077] The control unit 130 controls the dialogue with the speaker based on the generated dialogue data D5 (step S106). For example, the control unit 130 instructs the agent device 10 to perform a dialogue based on the dialogue data D5 via the communication unit 110. As a result, the agent device 10 performs an utterance based on the instructed dialogue data D5, thereby realizing a dialogue with the user U. When the processing in step S106 is completed, the control unit 130 terminates the processing procedure shown in Figure 9.

[0078] In the processing procedure shown in Figure 9, the control unit 130 functions as the acquisition unit 132 by executing step S102. The control unit 130 functions as the dialogue control unit 133 by executing the processes from step S103 to step S106.

[0079] Thus, when the information processing device 100 acquires episode data D1 related to topic information contained in the user U's utterance data D30, it can realize a dialogue that includes the speaker's episode based on the episode data D1. The information processing device 100 can provide the episode based on the user U's episode data D1 during the dialogue with the user U. As a result, by including the user U's episode in the dialogue, the information processing device 100 can realize a dialogue that the user U feels a sense of familiarity with.

[0080] [Example of processing procedure (2) for the information processing device according to the embodiment] Figure 10 is a flowchart of an example of a processing procedure (2) of the information processing device 100 according to the embodiment. The processing procedure shown in Figure 10 is realized by the control unit 130 of the information processing device 100 executing a program. The processing procedure shown in Figure 10 is executed by the information processing device 100 to spontaneously engage in dialogue with user U (speaker) of the agent device 10 when it recognizes user U and has not yet engaged in dialogue with user U.

[0081] As shown in Figure 10, the control unit 130 of the information processing device 100 determines whether or not it has recognized the speaker of the agent device 10 (step S201). For example, the control unit 130 acquires sensor information from the agent device 10 via the communication unit 110 and performs recognition processing of the speaker's face, voiceprint, etc., based on the sensor information and user data D3 associated with the agent device 10. If the control unit 130 determines, based on the results of the recognition processing, that it has not recognized the speaker (No in step S201), it terminates the processing procedure shown in Figure 10.

[0082] If the control unit 130 determines that it has recognized the speaker based on the results of the recognition process (Yes in step S201), it proceeds to step S202. The control unit 130 obtains episode data D1 that satisfies the acquisition conditions from the storage unit 120 (step S202). The acquisition conditions include, for example, conditions for acquiring the speaker's most recent episode or an episode corresponding to the time of day. For example, if the acquisition condition is to acquire the most recent episode, the control unit 130 obtains the latest episode data D1 from the storage unit 120. For example, if the acquisition condition is to acquire an episode corresponding to the time of day, the control unit 130 refers to the When item of the episode data D1 and obtains episode data D1 appropriate for the time of day from the storage unit 120. For example, if the time of day is morning, the control unit 130 obtains episode data D1 from the storage unit 120 that shows today's schedule, yesterday's events, etc. For example, if the time of day is daytime, the control unit 130 retrieves episode data D1 from the storage unit 120 that indicates afternoon schedules, evening schedules, etc. For example, if the time of day is nighttime, the control unit 130 retrieves episode data D1 from the storage unit 120 that indicates today's events, tomorrow's schedule, etc. When the processing in step S202 is completed, the control unit 130 proceeds to step S203.

[0083] The control unit 130 generates dialogue data D5 based on the episode data D1 and template data D2 (step S203). For example, the control unit 130 obtains template data D2 of the classification corresponding to the information in the "when" item of the episode data D1 from the storage unit 120. The control unit 130 generates dialogue data D5 for the agent device 10 to spontaneously speak by setting the information set in the items of the episode data D1 into the template of the obtained template data D2. Once the control unit 130 stores the generated dialogue data D5 in the storage unit 120, it proceeds to step S204.

[0084] The control unit 130 controls the dialogue with the speaker based on the generated dialogue data D5 (step S204). For example, the control unit 130 instructs the agent device 10 to perform a dialogue based on the dialogue data D5 via the communication unit 110. As a result, the agent device 10 spontaneously engages in dialogue with the user U by making utterances based on the instructed dialogue data D5. When the processing in step S204 is completed, the control unit 130 terminates the processing procedure shown in Figure 10.

[0085] In the processing procedure shown in Figure 10, the control unit 130 functions as the acquisition unit 132 by executing step S202. The control unit 130 functions as the dialogue control unit 133 by executing the processing from step S203 to step S204.

[0086] Thus, the information processing device 100 can acquire episode data D1 that satisfies acquisition conditions, regardless of the speaker's utterance data D30, and spontaneously provide episodes based on said episode data D1. Furthermore, since the information processing device 100 can acquire diverse episode data D1 depending on the acquisition conditions, it can spontaneously provide diverse episodes. As a result, the information processing device 100 can provide episodes of user U even when not interacting with user U, thereby fostering a sense of familiarity with user U.

[0087] [Example of data structure for question results] Figure 11 shows an example of question data D4 according to the embodiment. As shown in Figure 11, the question data D4 includes items such as time of day and question content. The time of day is, for example, the time of day when the question is asked, such as morning, noon, or evening. The question content contains data for asking questions appropriate to the time of day.

[0088] In the example shown in Figure 11, the question data D4 for the morning time slot includes questions such as, "Are you going anywhere today?", "What are your plans for today?", "What are you doing today?", "What did you do yesterday?", and "Where did you go yesterday?". The question data D4 for the daytime time slot includes questions such as, "What are you eating for dinner?", and "What are you doing this afternoon?". The question data D4 for the evening time slot includes questions such as, "What are you doing tomorrow?", and "What did you do today?".

[0089] [Processing procedure for information processing equipment that generates episode data using questions] Figure 12 is a flowchart showing the processing procedure of the information processing device 100 for generating episode data using questions. The processing procedure shown in Figure 12 is implemented by the control unit 130 of the information processing device 100 executing a program. The processing procedure shown in Figure 12 is executed by the information processing device 100 when the user U (speaker) of the agent device 10 is recognized.

[0090] As shown in Figure 12, the control unit 130 of the information processing device 100 determines whether or not it has recognized the speaker of the agent device 10 (step S301). If the control unit 130 determines that it has not recognized the speaker (No in step S301), it terminates the processing procedure shown in Figure 12. If the control unit 130 determines that it has recognized the speaker (Yes in step S301), it proceeds to step S302.

[0091] The control unit 130 controls the dialogue with the speaker based on the question data D4 (step S302). For example, the control unit 130 obtains question data D4 corresponding to the current time period from the storage unit 120 and instructs the agent device 10 via the communication unit 110 to perform a dialogue based on the question data. As a result, the agent device 10 issues a question to the user U by making an utterance based on the instructed dialogue data D5. When the processing in step S302 is completed, the control unit 130 proceeds to step S303.

[0092] The control unit 130 obtains the analysis result of the speech data D30 corresponding to the question data D4 (step S303). For example, the control unit 130 obtains the analysis result of the speech data D30 after instructing a dialogue based on the question data D4. Once the control unit 130 obtains the analysis result of the speech data D30, it proceeds to step S304.

[0093] The control unit 130 generates episode data D1 based on the analysis results of the speech data D30 (step S304). For example, the control unit 130 generates episode data D1 based on the analysis results of the speech data D30 as shown in Figures 7 and 8. For example, if the analysis results of the speech data D30 include information about an episode, the control unit 130 generates episode data D1. For example, if the analysis results of the speech data D30 do not include information for generating an episode, the control unit 130 does not generate episode data D1. The information for generating an episode includes, for example, information corresponding to items such as Action and When. Returning to Figure 12, the control unit 130 functions as a generation unit 134 by executing step S304. When the processing of step S304 is completed, the control unit 130 proceeds to step S305.

[0094] The control unit 130 determines in step S304 whether or not episode data D1 was generated (step S305). If the control unit 130 determines that episode data D1 was not generated (No in step S305), it terminates the processing procedure shown in Figure 12. If the control unit 130 determines that episode data D1 was generated (Yes in step S305), it proceeds to step S306. The control unit 130 stores the generated episode data D1 in the storage unit 120 in association with the speaker (step S306). For example, the control unit 130 stores the episode data D1 in the storage unit 120 in association with the user data D3 of the speaker who recognizes the episode data D1. After the processing in step S306 is completed, the control unit 130 terminates the processing procedure shown in Figure 12.

[0095] [Example of generating episode data using questions] Figure 13 shows an example of generating episode data using questions in the information processing device 100 according to this embodiment.

[0096] In one example shown in Figure 13, when the agent device 10 recognizes user U, it is instructed by the information processing device 100 to utter an utterance C21 based on question data D4. Utterance C21 is, for example, "Where are you going today?". User U responds to utterance C21 with utterance C22. Utterance C22 is, for example, "I'm going on a picnic with my friends." The agent device 10 transmits the utterance data of utterance C22 to the information processing device 100.

[0097] The information processing device 100 analyzes the speech data D30 of utterance C22 and extracts the topic information "friends" and "picnic". The question sentence of utterance C21 also contains the topic information "today" and "go". In this case, the information processing device 100 generates the episode data D1 shown in the lower right of Figure 13 based on the analysis result of the speech data D30 and the question data D4, and stores it in the storage unit 120 as the episode data D1 of user U. In the episode data D1, "Today" is set in the When field, "Speaker" in the Who field, "Go" in the Action field, "Picnic" in the Target field, and "Friends" in the With field, while the other fields are left blank.

[0098] In this way, the information processing device 100 can ask questions to user U based on the question data D4, and generate episode data D1 from the analysis results of the utterance data D30 corresponding to the questions and the question data D4, thereby constructing diverse episode data D1 for user U. As a result, the information processing device 100 can maintain the freshness of the episodes provided to user U by enriching the episodes that can be provided to user U.

[0099] [Processing procedure for information processing device that generates episode data using related information] Figure 14 is a flowchart showing the processing procedure for generating episode data based on relevant information from the information processing device 100. The processing procedure shown in Figure 14 is implemented by the control unit 130 of the information processing device 100 executing a program. The processing procedure shown in Figure 14 is executed by the information processing device 100 when the user U (speaker) of the agent device 10 is recognized.

[0100] As shown in Figure 14, the control unit 130 of the information processing device 100 collects relevant information about user U via the communication unit 110 (step S401). The relevant information includes, for example, sensor information from the agent device 10 that indicates the status of user U. The relevant information is information for telling stories about user U's past situations. The relevant information includes, for example, information indicating that user U's friends came to visit, or information indicating that parents or relatives were watching television on the weekend. The relevant information may also include, for example, user U's schedule, SNS (Social Networking Service) information, etc. Once the control unit 130 stores the collected relevant information in the storage unit 120, it proceeds to step S402.

[0101] The control unit 130 generates episode data D1 based on the collected relevant information (step S402). For example, the control unit 130 analyzes the relevant information and generates episode data D1 that shows past events based on the analysis results. For example, if the relevant information is that "a friend came over yesterday," the control unit 130 generates episode data D1 in which the When field is set to "yesterday," the Who field to "friend," the Action field to "come," and the Where field to "home," with the other fields left blank. Once the processing in step S402 is complete, the control unit 130 proceeds to step S403.

[0102] The control unit 130 determines whether or not episode data D1 was generated in step S402 (step S403). If the control unit 130 determines that episode data D1 was not generated (No in step S403), it terminates the processing procedure shown in Figure 14. If the control unit 130 determines that episode data D1 was generated (Yes in step S403), it proceeds to step S404. The control unit 130 stores the generated episode data D1 in the storage unit 120, associating it with the speaker (step S404). Once the processing in step S404 is completed, the control unit 130 terminates the processing procedure shown in Figure 14.

[0103] [Example of using episode data based on related information] Figure 15 shows an example of how the information processing device 100 according to this embodiment uses episode data based on related information.

[0104] In one example shown in Figure 15, the information processing device 100 generates episode data D1, shown in the lower right of Figure 15, based on the collected relevant information, and stores it in the storage unit 120 as user U's episode data D1. Episode data D1 has "Yesterday" set in the When field, "Friend" in the Who field, "Home" in the Where field, and "Come" in the Action field, with the other fields left blank.

[0105] When the information processing device 100 recognizes user U via the agent device 10, it generates dialogue data D5 based on the episode data D1 and template data D2 shown in Figure 15, which represent user U's past situation. The information processing device 100 instructs the agent device 10 to make an utterance based on the generated dialogue data D5. The agent device 10 makes an utterance S31 based on the dialogue data D5. For example, utterance S31 is, "My friend came over yesterday." User U responds to utterance C31 with utterance C32. For example, utterance C32 is, "That's right."

[0106] In this way, the information processing device 100 can collect relevant information about user U and generate episode data D1 based on that relevant information, so it can construct episode data D1 that corresponds to user U's past situation. As a result, the information processing device 100 can further enrich the episodes that can be provided to user U, thereby maintaining the freshness of the episodes provided to user U.

[0107] [Generalization function for episode data in information processing equipment] The information processing device 100 further includes a function to generate generalized episode data D1 based on common data from multiple episode data D1. Figure 16 is a diagram illustrating the generalization of multiple episode data D1. As shown in Figure 16, the information processing device 100 assumes that as episode data D1 for user U, it has episode data D1 for Ep11, Ep12, and Ep13, where the Ep_id item is Ep11. Episode data D1 for Ep11 is data that represents the episode "On September 1st, Yamada goes to the company by taxi." Episode data D1 for Ep12 is data that represents the episode "On September 2nd, Yamada goes to the company with a subordinate because they met by chance." Episode data D1 for Ep13 is data that represents the episode "On September 3rd, Yamada goes to the company late."

[0108] In this case, the information in the Who, Where, and Action fields of the episode data D1 for Ep11, Ep12, and Ep13 matches. When the control unit 130 of the information processing device 100 detects a group of episode data D1 from among multiple episode data D1s in which the number of matching fields exceeds a preset threshold, it generalizes the episode data D1. In the example shown in Figure 16, the information processing device 100 generates episode data D1 with the three fields Who, Where, and Action generalized, with the Ep_id field set as Ep20. That is, in the episode data D1 of Ep20, "Yamada" is set in the Who field, "Company" in the Where field, and "Go" in the Action field, while the other fields are blank.

[0109] Thus, when multiple similar episode data D1 exist, the information processing device 100 can generate episode data D1 that generalizes the user U's behavioral tendencies. As a result, the information processing device 100 can use the generalized episode data D1 to realize a dialogue with the user U based on the generalized episode. For example, when a generalized episode data D1 of Ep20 is generated, the information processing device 100 can realize a dialogue based on dialogue data D5 such as, "Are you going to work today?", "Aren't you going to work today?", and "How was work today?".

[0110] [Hardware configuration] The information equipment of the information processing system 1 according to the above embodiment is realized by a computer 1000 having a configuration such as that shown in Figure 17. Hereinafter, the information processing device 100 according to the embodiment will be used as an example. Figure 17 is a hardware configuration diagram showing an example of a computer 1000 that realizes the functions of the information processing device 100. The computer 1000 has a CPU 1100, RAM 1200, ROM (Read Only Memory) 1300, HDD (Hard Disk Drive) 1400, a communication interface 1500, and an input / output interface 1600. The various parts of the computer 1000 are connected by a bus 1050.

[0111] The CPU 1100 operates based on programs stored in the ROM 1300 or HDD 1400, and controls various parts. For example, the CPU 1100 loads the programs stored in the ROM 1300 or HDD 1400 into the RAM 1200 and executes processing corresponding to various programs.

[0112] ROM1300 stores boot programs such as the BIOS (Basic Input Output System) executed by CPU1100 when computer 1000 starts up, as well as programs that depend on the computer 1000's hardware.

[0113] HDD1400 is a computer-readable recording medium that non-temporarily records programs executed by CPU1100 and data used by such programs. Specifically, HDD1400 is a recording medium that records an information processing program related to this disclosure, which is an example of program data 1450.

[0114] The communication interface 1500 is an interface for the computer 1000 to connect to an external network 1550 (e.g., the Internet). For example, the CPU 1100 can receive data from other devices or transmit data it generates to other devices via the communication interface 1500.

[0115] The input / output interface 1600 is an interface for connecting the input / output device 1650 and the computer 1000. For example, the CPU 1100 receives data from input devices such as a keyboard or mouse via the input / output interface 1600. The CPU 1100 also transmits data to output devices such as a display, speaker, or printer via the input / output interface 1600. The input / output interface 1600 may also function as a media interface for reading programs recorded on a predetermined recording medium (media). Examples of media include optical recording media such as DVDs (Digital Versatile Discs), magneto-optical recording media such as MOs (Magneto-Optical disks), tape media, magnetic recording media, or semiconductor memory.

[0116] For example, when computer 1000 functions as an information processing device 100 according to the embodiment, the CPU 1100 of computer 1000 realizes functions such as the recognition unit 131, acquisition unit 132, dialogue control unit 133, generation unit 134, collection unit 135, and operation control unit 140 by executing an information processing program loaded on RAM 1200. The HDD 1400 stores the information processing program according to this disclosure and data in the storage unit 120. The CPU 1100 reads and executes program data 1450 from HDD 1400, but as another example, these programs may be obtained from other devices via an external network 1550.

[0117] In the above embodiment, the information processing system 1 was described in a case where the agent device 10 and the information processing device 100 work together to perform interactive processing, but it is not limited to this. For example, the information processing system 1 may have the agent device 10 perform interactive processing independently. In this case, the agent device 10 can implement the acquisition unit 132, dialogue control unit 133, generation unit 134, collection unit 135, operation control unit 140, etc. of the information processing device 100 using the control unit 16.

[0118] While preferred embodiments of the present disclosure have been described in detail above with reference to the attached drawings, the technical scope of the present disclosure is not limited to such examples. It is clear to any person with ordinary skill in the art of the present disclosure that various modifications or alterations may be conceived within the scope of the technical ideas described in the claims, and these will naturally also fall within the technical scope of the present disclosure.

[0119] Furthermore, the effects described herein are merely descriptive or illustrative and not limiting. In other words, the technology relating to this disclosure may produce other effects that are obvious to those skilled in the art from the description herein, in addition to or in lieu of the effects described herein.

[0120] Furthermore, it is possible to create a program that enables the CPU, ROM, RAM, and other hardware built into the computer to perform functions equivalent to those of the information processing device 100, and a computer-readable recording medium on which such a program is stored can also be provided.

[0121] Furthermore, each step of the processing of the information processing device 100 described herein does not necessarily have to be processed chronologically in the order shown in the flowchart. For example, each step of the processing of the information processing device 100 may be processed in an order different from the order shown in the flowchart, or may be processed in parallel.

[0122] Although this specification describes an information processing device 100 that provides episode data D1 through interaction (voice) with user U via agent device 10, it is not limited to this. For example, the information processing device 100 may be configured to provide episode data D1 via display device, or to provide episode data D1 through a combination of display and voice output.

[0123] (effect) The information processing device 100 includes an acquisition unit 132 that acquires episode data D1 relating to topic information contained in the speaker's utterance data D30 from a storage unit 120 that stores the speaker's episode data D1, and a dialogue control unit 133 that controls the dialogue with the speaker so as to include an episode based on the episode data D1.

[0124] As a result, when the information processing device 100 acquires episode data D1 related to topic information contained in the speaker's utterance data D30, it can realize a dialogue that includes the speaker's episode based on the episode data D1. The information processing device 100 can provide the episode based on the speaker's episode data D1 within the dialogue with the speaker. As a result, by including the speaker's episode in the dialogue, the information processing device 100 can realize a dialogue that makes the speaker feel a sense of familiarity.

[0125] In the information processing device 100, the dialogue control unit 133 generates dialogue data D5 containing an episode based on the episode data D1 and an algorithm for generating dialogue sentences, and controls the dialogue based on the dialogue data D5.

[0126] As a result, the information processing device 100 can generate dialogue data D5, which includes the speaker's episodes, based on the episode data D1 and the algorithm, thus simplifying the structure of the episode data D1. Consequently, the information processing device 100 can construct diverse episode data D1 for the speaker, making it possible to include a variety of episodes in the dialogue with the speaker and enabling a dialogue that fosters a greater sense of familiarity with the speaker.

[0127] In the information processing device 100, the algorithm has templates corresponding to the classification of episodes, and the dialogue control unit 133 generates dialogue data D5 by setting the data of episode data D1 into the template.

[0128] As a result, the information processing device 100 can generate dialogue data D5 by setting the data of episode data D1 as a template, thereby simplifying the process of generating episode data D1. Furthermore, by classifying the episode data D1, the information processing device 100 can obtain episode data D1 that is appropriate for the content of the speaker's utterances. Consequently, the information processing device 100 can include episodes appropriate for the content of the speaker's utterances in the dialogue, thereby enabling a dialogue that the speaker will find more relatable.

[0129] In the information processing device 100, the acquisition unit 132 acquires episode data D1 whose topic information satisfies the dialogue conditions.

[0130] As a result, the information processing device 100 can acquire episode data D1 suitable for topic information based on dialogue conditions and generate dialogue data D5 based on the episode data D1. Consequently, the information processing device 100 can include episodes that satisfy the dialogue conditions in the dialogue, thereby enabling a dialogue that fosters a greater sense of familiarity among the speakers.

[0131] In the information processing device 100, the acquisition unit 132 acquires episode data D1 that satisfies the acquisition conditions when it recognizes a speaker, and the dialogue control unit 133 controls the dialogue with the speaker so that it spontaneously speaks the episode indicated by the acquired episode data D1.

[0132] As a result, the information processing device 100 can acquire episode data D1 that satisfies acquisition conditions, regardless of the speaker's utterance, and spontaneously provide episodes based on said episode data D1. Furthermore, since the information processing device 100 can acquire diverse episode data D1 depending on the acquisition conditions, it can spontaneously provide diverse episodes. Consequently, the information processing device 100 can provide episodes of the speaker even when not interacting with the speaker, thereby fostering a sense of familiarity with the speaker.

[0133] In the information processing device 100, if the acquisition unit 132 has not acquired episode data D1, the dialogue control unit 133 controls the dialogue with the speaker based on dialogue data D5 which is different from the episode.

[0134] As a result, if the information processing device 100 cannot obtain episode data D1 that is suitable for the speaker's utterance, it can implement a dialogue based on dialogue data D5 that does not include an episode. Consequently, if there is no episode suitable for the speaker's utterance, the information processing device 100 can prevent the dialogue with the speaker from being interrupted by engaging in a dialogue that is different from the episode.

[0135] The information processing device 100 further comprises a generation unit 134 that generates episode data D1 based on the analysis results of speech data D30, and a storage unit 120 that stores the episode data D1 in association with the speaker.

[0136] As a result, the information processing device 100 generates episode data D1 based on the analysis results of the speaker's speech data D30, and stores the episode data D1 in the storage unit 120 in association with the speaker. Consequently, the information processing device 100 can enrich the episodes included in the dialogue with the speaker by constructing episode data D1 according to the content of the speaker's speech, thereby enabling a dialogue that the speaker will feel even closer to.

[0137] In the information processing device 100, the dialogue control unit 133 controls the dialogue with the speaker based on the question data D4 that asks the speaker, and the generation unit 134 generates episode data D4 based on the analysis results of the utterance data D30 corresponding to the question data D4 and the question data D4.

[0138] As a result, the information processing device 100 can ask the speaker questions based on the question data D4, and generate episode data D1 from the analysis results of the utterance data D30 corresponding to the questions and the question data D4, thereby constructing diverse episode data D1 of the speaker. Consequently, the information processing device 100 can maintain the freshness of the episodes provided to the speaker by enriching the episodes that can be provided to the speaker.

[0139] The information processing device 100 further includes a collection unit 135 that collects relevant information about the speaker's past circumstances, and a generation unit 134 generates episode data D1 based on the relevant information.

[0140] As a result, the information processing device 100 can collect relevant information about the speaker and generate episode data D1 based on that relevant information, thereby constructing episode data D1 that corresponds to the speaker's past circumstances. Consequently, the information processing device 100 can further enrich the episodes available to the speaker, thereby maintaining the freshness of the episodes provided to the speaker.

[0141] In the information processing device 100, the generation unit 134 generates generalized episode data d1 based on common data of the episode data D1 in the storage unit 120.

[0142] As a result, if the information processing device 100 finds that the speaker's episode data D1 contains common data, it can generate episode data D1 that generalizes the common data. Consequently, the information processing device 100 can realize a dialogue that includes the generalized speaker's episodes, thus providing a more intimate dialogue without interruption with the speaker.

[0143] The information processing system 1 comprises an agent device 10 for collecting speaker utterance data D30 and an information processing device 100, wherein the information processing device 100 includes an acquisition unit 132 for acquiring episode data D1 relating to topic information contained in the utterance data D30 from a storage unit 120 that stores the speaker's episode data D1, and a dialogue control unit 133 for controlling the dialogue with the speaker so as to include an episode based on the episode data D1.

[0144] As a result, when the information processing system 1 obtains episode data D1 related to topic information contained in the speaker's utterance data D30, it can realize a dialogue that includes the speaker's episodes based on the episode data D1. The information processing system 1 can provide episodes based on the speaker's episode data D1 within the dialogue with the speaker. Consequently, by including the speaker's episodes in the dialogue, the information processing system 1 can realize a dialogue that makes the speaker feel a sense of familiarity.

[0145] In the information processing system 1, the agent device 10 is a mobile robot, and the dialogue control unit 133 controls the dialogue with the speaker via the agent device 10.

[0146] As a result, the information processing system 1 can collect speaker utterance data D30 via the agent device 10 and realize a dialogue that includes the speaker's episodes based on the episode data D1. Consequently, by including the speaker's episodes in the dialogue with the mobile robot, the information processing system 1 can realize a friendly dialogue between the speaker and the robot.

[0147] The information processing method includes the computer obtaining episode data D1 relating to topic information contained in the speaker's utterance data D30 from a storage unit 120 that stores the speaker's episode data D1, and controlling the interaction with the speaker so as to include episodes based on the episode data D1.

[0148] As a result, the information processing method, upon acquiring episode data D1 related to topic information contained in the speaker's utterance data D30, can realize a dialogue by computer that includes the speaker's episodes based on the said episode data D1. The information processing method can provide episodes based on the speaker's episode data D1 within the dialogue with the speaker. Consequently, by including the speaker's episodes in the dialogue, the information processing method can realize a dialogue by computer that fosters a sense of familiarity with the speaker.

[0149] Furthermore, the following configurations also fall within the technical scope of this disclosure. (1) An acquisition unit acquires episode data relating to topic information contained in the speaker's utterance data from a storage unit that stores the speaker's episode data, A dialogue control unit that controls the dialogue with the speaker to include an episode based on the aforementioned episode data, An information processing device equipped with the following features. (2) The dialogue control unit generates dialogue data including the episode based on the episode data and the algorithm for generating dialogue text, and controls the dialogue based on the dialogue data. The information processing device described in (1) above. (3) The algorithm has a template corresponding to the classification of the episode, The dialogue control unit generates the dialogue data by setting the episode data into the template. The information processing device described in (2) above. (4) The acquisition unit acquires the episode data for which the topic information satisfies the dialogue conditions. An information processing device according to any one of (1) to (3) above. (5) If the acquisition unit recognizes the speaker, it acquires the episode data that satisfies the acquisition conditions. The dialogue control unit controls the dialogue with the speaker so that the speaker spontaneously speaks the episode indicated by the acquired episode data. The information processing device described in (4) above. (6) If the acquisition unit has not acquired the episode data, the dialogue control unit controls the dialogue with the speaker based on dialogue data different from the episode. An information processing device according to any of (1) to (5) above. (7) A generation unit that generates the episode data based on the analysis results of the speech data, The storage unit stores the episode data in association with the speaker, To further enhance An information processing device according to any one of (1) to (6) above. (8) The dialogue control unit controls the dialogue with the speaker based on the question data to be asked of the speaker. The generation unit generates the episode data based on the analysis results of the utterance data corresponding to the question data and the question data. The information processing device described in (7) above. (9) The system further includes a collection unit that collects relevant information regarding the speaker's past circumstances, The generation unit generates the episode data based on the related information. The information processing device described in (8) above. (10) The generation unit generates generalized episode data based on the common data of the episode data in the storage unit. The information processing apparatus described in (8) or (9) above. (11) An information processing system comprising an agent device for collecting speaker utterance data and an information processing device, The aforementioned information processing device is An acquisition unit that acquires the episode data relating to topic information contained in the utterance data from a storage unit that stores the speaker's episode data, A dialogue control unit that controls the dialogue with the speaker to include an episode based on the aforementioned episode data, An information processing system equipped with the following features. (12) The agent device is a mobile robot, The dialogue control unit controls the dialogue with the speaker via the agent device. The information processing system described in (11) above. (13) Computers To retrieve the episode data relating to topic information contained in the speaker's utterance data from a storage unit that stores the speaker's episode data, Controlling the dialogue with the speaker to include an episode based on the aforementioned episode data, Information processing methods including (14) On the computer, To retrieve the episode data relating to topic information contained in the speaker's utterance data from a storage unit that stores the speaker's episode data, Controlling the dialogue with the speaker to include an episode based on the aforementioned episode data, A computer-readable recording medium containing a program to achieve this. (15) On the computer, To retrieve the episode data relating to topic information contained in the speaker's utterance data from a storage unit that stores the speaker's episode data, Controlling the dialogue with the speaker to include an episode based on the aforementioned episode data, A program to achieve this. [Explanation of symbols]

[0150] 1. Information Processing System 10 Agent devices 11 Sensor section 12 Input section 13 Light source 14 Output section 15 Drive unit 16 Control Unit 17 Communications Department 100 Information Processing Devices 110 Communications Department 120 Storage section 130 Control Unit 131 Recognition part 132 Acquisition Department 133 Dialogue Control Unit 134 Generation part 135 Collection Department 140 Operation Control Unit D1 Episode Data D2 template data D3 User Data D4 Question Data D5 Dialogue Data D10 Knowledge Data D20 Management Data D30 Speech Data

Claims

1. An acquisition unit acquires episode data relating to topic information contained in the speaker's utterance data from a storage unit that stores the speaker's episode data, A dialogue control unit that controls the dialogue with the speaker to include an episode based on the aforementioned episode data, A recognition unit that recognizes the aforementioned speaker, Equipped with, The acquisition unit acquires the episode data that satisfies the acquisition conditions if the speaker is recognized and no dialogue has taken place with the speaker. The dialogue control unit controls the dialogue with the speaker so that the speaker spontaneously speaks the episode indicated by the episode data acquired by the acquisition unit. Information processing device.

2. The dialogue control unit generates dialogue data including the episode based on the episode data and the algorithm for generating dialogue text, and controls the dialogue based on the dialogue data. The information processing apparatus according to claim 1.

3. The algorithm has a template corresponding to the classification of the episode, The dialogue control unit generates the dialogue data by setting the episode data into the template. The information processing apparatus according to claim 2.

4. The acquisition unit acquires the episode data for which the topic information satisfies the dialogue conditions. The information processing apparatus according to claim 3.

5. If the acquisition unit has not acquired the episode data, the dialogue control unit controls the dialogue with the speaker based on dialogue data different from the episode. The information processing apparatus according to claim 1.

6. A generation unit that generates the episode data based on the analysis results of the speech data, The storage unit stores the episode data in association with the speaker, To further enhance The information processing apparatus according to claim 1.

7. The dialogue control unit controls the dialogue with the speaker based on the question data to be asked of the speaker. The generation unit generates the episode data based on the analysis results of the utterance data corresponding to the question data and the question data. The information processing apparatus according to claim 6.

8. The system further includes a collection unit that collects relevant information regarding the speaker's past circumstances, The generation unit generates the episode data based on the related information. The information processing apparatus according to claim 7.

9. The generation unit generates generalized episode data based on the common data of the episode data in the storage unit. The information processing apparatus according to claim 8.

10. An information processing system comprising an agent device for collecting speaker utterance data and an information processing device, The aforementioned information processing device is An acquisition unit that acquires the episode data relating to topic information contained in the utterance data from a storage unit that stores the speaker's episode data, A dialogue control unit that controls the dialogue with the speaker to include an episode based on the aforementioned episode data, A recognition unit that recognizes the aforementioned speaker, Equipped with, The acquisition unit acquires the episode data that satisfies the acquisition conditions if the speaker is recognized and no dialogue has taken place with the speaker. The dialogue control unit controls the dialogue with the speaker so that the speaker spontaneously speaks the episode indicated by the episode data acquired by the acquisition unit. Information processing system.

11. The agent device is a mobile robot, The dialogue control unit controls the dialogue with the speaker via the agent device. The information processing system according to claim 10.

12. Computers To retrieve the episode data relating to topic information contained in the speaker's utterance data from a storage unit that stores the speaker's episode data, Controlling the dialogue with the speaker to include an episode based on the aforementioned episode data, Recognizing the aforementioned speaker, Includes, The acquisition of the aforementioned data means that if the speaker is recognized and no dialogue has taken place with the speaker, the episode data that satisfies the acquisition conditions is acquired. The aforementioned control involves controlling the dialogue with the speaker so that the speaker spontaneously speaks the episode indicated by the episode data acquired. Information processing methods.

Images