Information processing device

Abstract

In a voice dialogue system, provide information of appropriate content. [Solution] The system stores first data for estimating the content of knowledge each of multiple users possesses, acquires conversations between the multiple users, acquires first information which is information related to the conversation, estimates whether at least one of the multiple users possesses knowledge about the first information based on the first data, and determines whether or not to output the first information based at least on the result of the estimation.

Description

Technical Field

[0001] The present disclosure relates to interaction technology using machine learning models.

Background Art

[0002] Systems that enable voice-based interaction are provided in vehicles. In this regard, for example, Patent Document 1 discloses a conversation system in which an AI agent actively participates in a conversation conducted by a plurality of people.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] It is expected that interaction services using natural language will further increase in the future along with the development of machine learning.

[0005] An objective of the present disclosure is to provide information of appropriate content in a voice interaction system.

Means for Solving the Problems

[0006] One aspect of an embodiment of the present disclosure is: An information processing apparatus comprising a storage unit that stores first data for estimating the content of knowledge that each of a plurality of users has, and a control unit that executes: acquiring a conversation held between the plurality of users; acquiring first information that is information related to the conversation; estimating, based on the first data, whether at least any one of the plurality of users has knowledge related to the first information; and determining whether to output the first information at least based on a result of the estimation.

[0007] Other embodiments include a method performed by the above-mentioned device, a program for causing a computer to perform the method, or a computer-readable storage medium that non-temporarily stores the program. [Effects of the Invention]

[0008] According to this disclosure, it is possible to provide information of appropriate content in a voice dialogue system. [Brief explanation of the drawing]

[0009] [Figure 1] A schematic diagram of the dialogue system in the first embodiment. [Figure 2] Configuration diagram of the in-vehicle device in the first embodiment. [Figure 3] An example of a feature database and a knowledge database. [Figure 4] A diagram illustrating the processing flow executed by the control unit of the in-vehicle device. [Figure 5] A flowchart of the processes executed by the control unit of the in-vehicle device. [Figure 6] The second flowchart of the process performed by the control unit of the in-vehicle device. [Modes for carrying out the invention]

[0010] In recent years, with the advancement of machine learning, the number of products incorporating language models has increased. For example, By incorporating a Large Language Model (LLM), whose accuracy has been improved through machine learning using large datasets, into a product, it is possible to add natural language dialogue functionality to the target product.

[0011] Automobiles are a prime example of the usefulness of natural language interaction. For instance, by equipping in-car systems with LLM (Language Learning Module), information can be obtained without operating touch panels or other similar devices. Such functionality is particularly useful when the driver's hands are occupied, such as while driving.

[0012] LLM can not only output responses to questions, but can also participate in conversations among passengers and proactively provide relevant information. For example, if passengers are discussing future destinations in the vehicle, LLM can search for potential locations in the background and proactively provide the results.

[0013] However, when such actions are performed by in-vehicle devices, there is a risk of providing information that the occupants already know. For example, if a conversation about visiting a specific facility is taking place in the vehicle, the system might proactively suggest directions to that facility. However, if the driver is already familiar with the facility, this could result in providing unnecessary information and potentially bothering the driver.

[0014] The information processing device relating to the first aspect of this disclosure is: The system includes a storage unit that stores first data for estimating the content of knowledge possessed by multiple users, and a control unit that performs the following: acquiring conversations between the multiple users, acquiring first information which is information related to the conversation, estimating whether or not at least one of the multiple users has knowledge of the first information based on the first data, and determining whether or not to output the first information based at least on the result of the estimation.

[0015] The information processing device can be, for example, a device mounted in a vehicle (in-vehicle device). The information processing device can, for example, acquire information related to a conversation (first information) based on the content of a conversation taking place inside the vehicle and actively output it. The first information may be, for example, additional information about a facility that is the topic of conversation (e.g., information about routes, information about business operations, etc.).

[0016] Further, the control unit estimates, based on first data for estimating the content of knowledge possessed by each of a plurality of users, whether at least any one of the plurality of users possesses the information to be output as knowledge. Further, the control unit determines whether to output the first information at least based on the estimation result. For example, when it is estimated that at least any one of the plurality of users has knowledge of the information to be output, the control unit may suppress the output of the first information.

[0017] The first data may be, for example, past action histories of the plurality of users. By referring to the past action histories, for example, it is possible to make a determination such as "since a user who has visited a specific facility in the past and is considered to have knowledge of the facility is riding in the vehicle, there is no need to output additional information (the first information) related to the facility". The predetermined timing may be, for example, a timing at which the conversation between users is interrupted.

[0018] The first data may be stored in the information processing apparatus, or may be acquired from terminals respectively associated with the plurality of users. For example, when the first data relates to past action histor ies, the control unit can also acquire the data related to the action histories from a portable terminal or the like possessed by the user.

[0019] Further, the first information may include information related to a first facility, and the control unit may estimate, based on the action histories, whether at least any one of the plurality of users has knowledge of the first facility as knowledge related to the first information. When the first information is information related to a facility, it is possible to estimate whether the target user has knowledge of the facility based on the presence or absence of a history of visiting the facility.

[0020] Furthermore, the control unit may output the first information if a predetermined length of time or longer has passed during which none of the users have spoken, even if it is presumed that at least one of the users has knowledge of the first information. If, after a certain amount of time has elapsed, no utterance regarding the first piece of information has actually been made, it is preferable for the information processing device to output the first piece of information.

[0021] Furthermore, if a conversation between the multiple users following the conversation contains erroneous information that can be corrected by the first information, the control unit may output the first information regardless of the result of the estimation.

[0022] Thus, when an utterance containing incorrect information is made, and it can be corrected by the first information, it is preferable for the information processing device to correct the incorrect information by outputting the first information.

[0023] The following describes specific embodiments of this disclosure with reference to the drawings. Unless otherwise specified, the hardware configurations, module configurations, functional configurations, etc., described in each embodiment are not intended to limit the technical scope of the disclosure to those configurations alone.

[0024] (First embodiment) [System Overview] An overview of the dialogue system according to the first embodiment will be described. The dialogue system according to this embodiment consists of an in-vehicle device 10 mounted on a vehicle 1. The in-vehicle device 10 is a device that can provide a natural language dialogue service using a virtual agent.

[0025] A virtual agent is a program that can be executed by the in-vehicle device 10 and is capable of natural language interaction. By running and keeping the virtual agent resident in the in-vehicle device 10, the vehicle's occupants can receive information services in natural language at any time they wish.

[0026] The virtual agent may behave as a different agent by loading a different language model for each of the multiple users riding in vehicle 1. For example, if there are four users who ride in vehicle 1 on a daily basis, the in-vehicle device 10 may store language models for all four users. In this case, the in-vehicle device 10 can provide information using the language model corresponding to the user riding in the vehicle. Each language model may contain data on the user's personal information and preferences, and may be configured to provide personalized suggestions.

[0027] The in-vehicle device 10 is installed in a connected vehicle that can communicate with any device via wireless communication. The in-vehicle device 10 may include a data communication module (DCM) for connecting the vehicle's components (e.g., ECUs, in-vehicle terminals, etc.) to a network. A virtual agent operating in the in-vehicle device 10 communicates with external devices, etc. via the network. It is also possible to gather information from there.

[0028] In this embodiment, the virtual agent running on the in-vehicle device 10 can provide information in response to requests from occupants, as well as monitor conversations taking place inside the vehicle and proactively provide information as needed. For example, suppose a discussion about a destination is taking place inside a car, and among the several potential stores mentioned, one is currently closed. In this case, the virtual agent can generate and output information such as "It appears that XX is closed today" based on the store information acquired in the background.

[0029] On the other hand, there are cases where the vehicle occupants already possess the information that is being output. In this case, the occupants may find it annoying to see information they already know being output. Therefore, in this embodiment, the in-vehicle device 10 acquires data on the knowledge of multiple occupants riding in the vehicle (the first data in this disclosure) and provides information only when it is estimated that "the occupants of the vehicle do not have knowledge of the information to be output."

[0030] [Device configuration] Next, the configuration of the devices that make up the system will be described. Figure 2 is a schematic diagram showing an example of the configuration of vehicle 1 according to this embodiment.

[0031] The in-vehicle device 10 can be configured as a computer having a processor (CPU, GPU, etc.), main memory (RAM, ROM, etc.), and auxiliary storage (EPROM, hard disk drive, removable media, etc.). The auxiliary storage contains an operating system (OS), various programs, various tables, etc., and by executing the programs stored therein, various functions (software modules) that match a predetermined purpose, as described later, can be realized. However, some or all of the functions may be realized as hardware modules by hardware circuits such as ASICs and FPGAs.

[0032] The in-vehicle device 10 is configured as hardware and includes a control unit 11, a storage unit 12, a communication unit 13, and an input / output unit 14. The in-vehicle device 10 is also connected to a group of sensors 15 of the vehicle 1.

[0033] The control unit 11 is a computing unit that realizes various functions of the in-vehicle device 10 by executing a predetermined program. The control unit 11 can be implemented by a hardware processor such as a CPU. The control unit 11 may also be configured to include RAM, ROM (Read Only Memory), cache memory, etc.

[0034] In this embodiment, the control unit 11 of the in-vehicle device 10 is configured with three software modules: an interactive input / output unit 111, an agent unit 112, and a determination unit 113. Each software module may be implemented by the control unit 11 (CPU, etc.) executing a program stored in the storage unit 12. The information processing performed by the software modules is synonymous with the information processing performed by the control unit 11 (CPU, etc.).

[0035] The dialogue input / output unit 111 acquires utterances made by the vehicle occupants via the input / output unit 14. The dialogue input / output unit 111 performs predetermined processing on the acquired voice data and performs speech recognition. As a result, the content of the utterances is converted into text. Unit 111 transmits the text obtained as a result of speech recognition to the agent unit 112.

[0036] Furthermore, the dialogue input / output unit 111 converts the utterance transmitted from the agent unit 112 into speech and outputs it via the input / output unit 14.

[0037] Agent unit 112 provides a dialogue service using a language model. By providing a dialogue service using a language model, agent unit 112 can behave as a virtual agent. A virtual agent may be implemented, for example, by a combination of a program and a language model. In this embodiment, a different language model is used for each occupant. For example, the agent unit 112 can behave as a virtual agent for user A by using the language model corresponding to user A, and can behave as a virtual agent for user B by using the language model corresponding to user B.

[0038] In this embodiment, these language models are stored in the storage unit 12, which will be described later. The agent unit 112 can set up a virtual agent, for example, by obtaining a language model corresponding to the driver from the storage unit 12.

[0039] Furthermore, if specialized knowledge is required in the dialogue, the agent unit 112 may collect information from one or more external devices available via the network. For example, if it is determined that the occupant who made the utterance is seeking information about facilities or shops, the agent unit 112 may access an external language model that is available via the network and capable of providing local information, and obtain the information.

[0040] The agent unit 112 can provide information in response to requests made through speech, and also has the function of monitoring conversations taking place in the vehicle, acquiring additional information, and outputting it spontaneously. For example, based on the content of the conversation, the agent unit 112 estimates that the vehicle occupant needs information about a specific subject (e.g., information about a facility they are considering visiting) and searches for that information in the background. The agent unit 112 then obtains information for identifying the vehicle occupants from the determination unit 113 (described later), searches the knowledge database (described later) based on this information, and estimates the knowledge that each of the vehicle occupants is presumed to possess.

[0041] A knowledge database is a collection of data used to estimate the knowledge possessed by the occupants of a vehicle. Knowledge databases may include personal information. For example, if user A has an educational or professional background in a particular field, it can be inferred that "user A possesses knowledge related to that field." Knowledge may also be inferred based on past experience. For example, if a knowledge database includes a history of user A's past visits to a particular location, it can be inferred that "user A possesses knowledge about the visited location." The agent unit 112 decides whether or not to output the target information based on the results of the estimation, and then outputs the information.

[0042] The determination unit 113 identifies the person riding in vehicle 1. For example, the determination unit 113 recognizes multiple people riding in the vehicle 1 and obtains their identifiers based on sensor data output by multiple sensors included in the sensor group 15 (described later) and a feature database (described later) that records the characteristics of people.

[0043] For example, if user A is in the vehicle, the determination unit 113 obtains user A's identifier based on biometric information, image information, etc., acquired from the sensors. The identification of individuals occupying the vehicle may be performed based on biometric information or facial images acquired via sensors, or by utilizing other additional information. For example, if each occupant of vehicle 1 is carrying a mobile device, the identification of individuals occupying the vehicle may be performed based on the results of short-range wireless communication with these mobile devices. The identification results are transmitted to agent unit 112.

[0044] The memory unit 12 is a means for storing information and is composed of storage media such as RAM, magnetic disks, and flash memory. The memory unit 12 stores programs executed by the control unit 11, data used by those programs, and so on.

[0045] Furthermore, the memory unit 12 stores the aforementioned feature database and knowledge database. Figure 3(A) shows an example of a feature database. The feature database stores features obtained from sensor data in association with personal identifiers (user IDs). Figure 3(B) shows an example of a knowledge database. The knowledge database stores personal identifiers (user IDs) in association with attribute information and behavioral history. Attribute information is personal information that can be used to estimate knowledge, and may include, for example, information on educational background, work history, or field of expertise. Behavioral history is information representing places visited in the past. Behavioral history may also be a collection of location information represented in chronological order.

[0046] The communication unit 13 includes a communication module that performs wireless communication with a predetermined network (for example, a wide-area network such as the Internet). In this embodiment, the communication unit 13 is configured to communicate with a predetermined mobile communication network. The communication unit 13 can be configured to have an eUICC (for example, a SIM card). Through the communication unit 13, the in-vehicle device 10 can communicate with an external device via the predetermined mobile communication network.

[0047] The input / output unit 14 is a unit that receives input from the user of the device and presents information to the user. Typically, the input / output unit 14 includes devices for inputting and outputting sound, such as a microphone and a speaker. The input / output unit 14 may also include devices that provide visual information (such as a display).

[0048] Sensor group 15 is a collection of multiple sensors installed in vehicle 1. The multiple sensors are not limited to specific types, as long as they can acquire data to identify people riding in the vehicle, such as image sensors or biometric sensors. In addition, means of personal identification by communicating with a mobile terminal inside the vehicle may also be included in sensor group 15.

[0049] Note that the configuration shown in Figure 2 is just one example, and all or part of the illustrated functions may be performed using specially designed circuits. Furthermore, program storage and execution may be performed using combinations of main memory and auxiliary memory other than those shown.

[0050] [Process Flow] Next, an overview of the processing performed by the control unit 11 will be described. Figure 4 is a diagram showing the flow of processing performed by the control unit 11 of the in-vehicle device 10.

[0051] The dialogue input / output unit 111 acquires speech from the vehicle occupants via the input / output unit 14. For example, the input / output unit 14 converts speech acquired via a microphone or the like into voice data, which the dialogue input / output unit 111 acquires. The dialogue input / output unit 111 performs a predetermined speech recognition process on the acquired voice data and converts the voice data into text. The system, while withholding a response to the utterance, transmits the text obtained as a result of speech recognition to the agent unit 112. This text will hereafter be referred to as the "utterance."

[0052] For example, if the content of the utterance requests the activation of a virtual agent, the agent unit 112 responds to the utterance by activating the virtual agent using the language model corresponding to the crew member who made the utterance.

[0053] If the content of the utterance requests interaction with a virtual agent, the agent unit 112 generates a response using the acquired language model. For example, if the currently running virtual agent corresponds to user A, the agent unit 112 forwards the utterance to the language model corresponding to user A. If the agent unit 112 determines that the language model corresponding to user A cannot provide sufficient information, it may forward the utterance to an external language model (remote language model) via the network. The agent unit 112 obtains the response (provided information) generated in response to the utterance from the target language model.

[0054] Furthermore, the agent unit 112 can also provide information spontaneously. For example, the agent unit 112 may acquire utterances corresponding to a conversation taking place between multiple occupants inside the vehicle from the dialogue input / output unit 111 and understand the gist and content of the conversation. In this case, the agent unit 112 determines that it is necessary to provide the occupants of the vehicle with additional information related to the conversation (hereinafter referred to as related information) and acquires the related information (for example, through a language model stored in its own device or an external device).

[0055] Now, let's explain the processing of the determination unit 113. The determination unit 113 identifies multiple people riding in the vehicle 1 based on sensor data output by multiple sensors included in the sensor group 15. Person identification can be performed, for example, by searching a feature database using features obtained by transforming the sensor data. The identification result is transmitted to the agent unit 112.

[0056] The agent unit 112 estimates, based on the identification result obtained from the determination unit 113, whether or not there is a crew member who has knowledge about the information to be provided.

[0057] For example, the agent unit 112 obtains knowledge-related data for each passenger from the knowledge database based on the identification result obtained from the determination unit 113 (for example, multiple identifiers corresponding to each of the multiple passengers). An example of knowledge-related data might be: "User A and User B are in the vehicle; User A specializes in electrical engineering, and User B is a student." Knowledge-related data may also include data on the past behavioral history of Users A and B, as it can be assumed they possess a certain level of knowledge about places they have visited.

[0058] Furthermore, the agent unit 112 performs estimations about the knowledge possessed by each crew member based on this data. For example, if the closing time of a restaurant being considered for a visit is approaching during the conversation, the agent unit 112 is expected to generate information about the restaurant's business hours as relevant information. In this case, if a crew member who has previously visited the restaurant after closing time is participating in the conversation, it is presumed that this crew member has knowledge of the restaurant's business hours.

[0059] Agent unit 112 possesses knowledge of the relevant information it intends to provide. The relevant information will be output only if there are no members present.

[0060] In this embodiment, whether or not a vehicle occupant possesses specific knowledge is estimated based on information stored in the knowledge database, but the estimation method is not limited to a specific technique. For example, the agent unit 112 may obtain information about messages, emails, SNS, etc. that the user has sent or received in the past, or information about past behavioral history, from the user's mobile terminal, etc., and estimate the knowledge based on that information.

[0061] [flowchart] Next, we will describe the details of the processes performed by the in-vehicle device 10. Figure 5 is a flowchart of the processes performed by the in-vehicle device 10. The processes shown in Figure 5 are executed periodically while the virtual agent is running.

[0062] First, in step S11, the determination unit 113 identifies the vehicle occupants. In this step, the determination unit 113 first acquires sensor data. The sensor data may be acquired from multiple sensors included in the sensor group 15. Then, based on the features obtained from the sensor data, it searches the feature database and obtains the occupant identification result (e.g., personal identifier). The identification result is transmitted to the agent unit 112.

[0063] Furthermore, it is also possible to identify vehicle occupants using data other than sensor data. For example, the determination unit 113 can obtain data to identify vehicle occupants by communicating with a terminal carried by the vehicle occupant via a unit or module that performs short-range wireless communication. If individual identification has been completed and no passengers are getting on or off, step S11 may be omitted.

[0064] Next, in step S12, the agent unit 112 acquires the content of the conversation inside the vehicle. The conversation may also be acquired by analyzing audio data collected over a predetermined period in the past. For this reason, the agent unit 112 may perform the process of collecting audio data in the background.

[0065] In step S13, the agent unit 112 generates relevant information (information related to the subject being discussed in the conversation) based on the content of the acquired conversation. Whether or not relevant information needs to be generated can be determined by a rule-based system or by a machine learning model. In the former case, for example, rules such as "generate relevant information if a visit to the facility is being discussed and there is a possibility that the arrival will be outside of business hours" or "generate relevant information if a visit to the facility is being discussed and there is no mention of usage fees in the conversation" can be used. If relevant information is generated, the process proceeds to step S14. If no relevant information is generated, the process terminates.

[0066] In step S14, the agent unit 112 estimates whether there are any occupants who possess the generated related information as knowledge. For example, if the related information is about a facility, and "an occupant is considered to have knowledge about the facility if they have visited the facility before," the agent unit 112 may determine whether a person with a history of visiting the facility is on board vehicle 1. If the determination in this step is negative, the process proceeds to step S15. If the determination in this step is positive, the process proceeds to step S16.

[0067] If the process proceeds to step S15, the agent unit 112 outputs relevant information at a predetermined timing. The predetermined timing is, for example, when none of the vehicle occupants are speaking. It can be set to a good timing.

[0068] In cases where the process transitions to step S16, the agent unit 112 does not output the relevant information because a passenger with knowledge of the relevant information is on board vehicle 1. However, it is preferable to output the relevant information if no utterance containing the relevant information is made after a predetermined period of time has elapsed. Therefore, in step S16, the agent unit 112 waits for an utterance containing the relevant information to be made or for a predetermined period of time, and outputs the relevant information if the period times out (step S16-Yes). If the period does not time out and an utterance containing the relevant information is made (step S16-No), the process ends.

[0069] In addition to the condition "after a specified period of time has elapsed," other conditions can also be set to trigger the output of related information.

[0070] As described above, the in-vehicle device according to the first embodiment generates relevant information based on the conversations of multiple people riding in the same vehicle. Furthermore, if it is presumed that no person with knowledge of the relevant information is riding in the vehicle, the relevant information is output. With this configuration, it is possible to prevent providing unnecessary information and bothering the occupants.

[0071] (Second Embodiment) In the first embodiment, the in-vehicle device 10 suppressed the output of relevant information when it was presumed that a person with knowledge of the relevant information was riding in the vehicle. However, it is possible that the person in question possesses incorrect knowledge. The second embodiment addresses this by correcting utterances based on incorrect knowledge using relevant information.

[0072] Figure 6 is a flowchart of the process performed by the in-vehicle device 10 (control unit 11) in the second embodiment. The illustrated process is started when a timeout does not occur in step S16, that is, when a vehicle occupant makes an utterance about the relevant information.

[0073] First, in step S21, the agent unit 112 acquires additional utterances from inside the vehicle. For example, if the relevant information was about the store's closing days, an utterance about the store's opening days might be made. In step S22, the agent unit 112 determines whether the spoken content contains misinformation, that is, whether it is inconsistent with the relevant information. For example, if the relevant information indicates that "the store you are trying to visit is closed on Mondays," and the utterance includes the statement "the store is open on Mondays," then this step is judged positively, and the process proceeds to step S23. If this step is judged negatively, the process ends.

[0074] In step S23, the agent unit 112 outputs relevant information to correct the utterance. In this step, it may generate a response that points out that the content of the utterance is incorrect and output it along with the relevant information, or it may output only the relevant information.

[0075] According to the second embodiment, if an utterance made by a vehicle occupant contains an error, the system makes an utterance regardless of the result of the knowledge estimation. With this configuration, the correct information can be conveyed without bothering the occupant.

[0076] (modified version) The embodiments described above are merely examples, and this disclosure may be modified as appropriate without departing from its essence. For example, the processes and means described in this disclosure can be freely combined and implemented, as long as no technical inconsistencies arise.

[0077] Furthermore, while the description of the embodiment included an example where a language model provides information, the source of information provided by the in-vehicle device 10 is not limited to the language model. For example, if any program (user application) is running in the in-vehicle device 10, filtering may be performed on the information output by that user application. In this case, the agent unit 112 may relay and filter the target information.

[0078] Furthermore, in the description of the embodiment, information was provided when it was presumed that the vehicle occupant did not possess knowledge related to the information being offered. Conversely, information may also be provided when it was presumed that the vehicle occupant possessed relevant knowledge. This can also have the effect of facilitating conversation.

[0079] Furthermore, a process described as being performed by a single device may be divided and executed by multiple devices. Conversely, a process described as being performed by different devices may be executed by a single device. In a computer system, the hardware configuration (server configuration) by which each function is implemented can be flexibly changed.

[0080] The present disclosure can also be realized by supplying a computer program implementing the functions described in the embodiments above to a computer, and having one or more processors in the computer read and execute the program. Such a computer program may be provided to the computer by a non-temporary computer-readable storage medium that can be connected to the computer's system bus, or it may be provided to the computer via a network. Non-temporary computer-readable storage mediums include, for example, any type of disk such as magnetic disks (floppy disks, hard disk drives (HDDs), etc.), optical disks (CD-ROMs, DVDs, Blu-ray discs, etc.), read-only memory (ROM), random access memory (RAM), EPROM, EEPROM, magnetic cards, flash memory, optical cards, and any type of medium suitable for storing electronic instructions. [Explanation of Symbols]

[0081] 10...In-vehicle equipment 11. Control Unit 12...Storage section 13. Communications Department 14...Input / output section 15. Sensor group

Claims

1. A storage unit that stores first data for estimating the content of knowledge held by multiple users, To obtain the conversations that took place between the aforementioned multiple users, To obtain the first piece of information which is information related to the aforementioned conversation, Based on the first data, estimate whether at least one of the multiple users has knowledge of the first information, Based at least on the results of the estimation, the decision is made as to whether or not to output the first information, An information processing device having a control unit that performs the following:

2. The first information described above includes information relating to the first facility, The control unit estimates, based on the behavioral history of each of the multiple users, whether or not at least one of the multiple users has knowledge of the first facility as knowledge of the first information. The information processing apparatus according to claim 1.

3. The control unit shall not output the first information if it is presumed that at least one of the multiple users has knowledge of the first information. The information processing apparatus according to claim 1.

4. If a period of time lasting longer than a predetermined length is observed during which none of the users have spoken, the control unit will output the first information, even if it is presumed that at least one of the users has knowledge of the first information. The information processing apparatus according to claim 3.

5. The control unit outputs the first information regardless of the estimation result if the conversations that took place between the multiple users following the conversation contain erroneous information that can be corrected by the first information. The information processing apparatus according to claim 1.

Images