JP7409781B2 - Output program, output device and output method - Google Patents

Description

The present invention relates to an output program, an output device, and an output method.

2. Description of the Related Art Conventionally, techniques related to devices that respond to a conversation from a person are known. For example, technologies have been proposed for devices that interact with people using utterances that can be associated with landscapes and landscapes.

Japanese Patent Application Publication No. 2018-147145

However, with the above-mentioned conventional technology, it is not always possible to output appropriate information without depending on the user's input to the dialogue system. For example, the above-mentioned conventional technology merely selects utterances that can be associated with scenery and landscapes from pre-registered utterance contents, and outputs appropriate information without relying on user input to the dialogue system. It doesn't necessarily mean that you can.

The present application has been made in view of the above, and aims to provide an output program, an output device, and an output method that can output appropriate information without relying on user input to an interaction system. .

The output program according to the present application includes an acquisition procedure for acquiring reaction information that is character information generated as a reaction to second information other than the first information input by the user to the dialogue system, and the reaction information acquired by the acquisition procedure. The present invention is characterized by causing a computer to execute an output procedure of outputting reaction information to the user.

According to one aspect of the embodiment, it is possible to output appropriate information without relying on user input to the dialogue system.

FIG. 1 is a diagram illustrating an example of output processing according to an embodiment. FIG. 2 is a diagram illustrating a configuration example of an output device according to an embodiment. FIG. 3 is a diagram illustrating an example of a sensor information storage unit according to the embodiment. FIG. 4 is a diagram illustrating an example of a content storage unit according to the embodiment. FIG. 5 is a diagram illustrating an example of a reaction information storage unit according to the embodiment. FIG. 6 is a diagram illustrating an example of a model information storage unit according to the embodiment. FIG. 7 is a diagram showing an example of a model according to the embodiment. FIG. 8 is a diagram illustrating an example of a learning data storage unit according to the embodiment. FIG. 9 is a flowchart showing the output processing procedure according to the embodiment. FIG. 10 is a diagram illustrating a configuration example of a content distribution device according to an embodiment. FIG. 11 is a diagram illustrating an example of a user information storage unit according to the embodiment. FIG. 12 is a diagram illustrating an example of a content storage unit according to the embodiment. FIG. 13 is a diagram illustrating an example of output processing according to a modification. FIG. 14 is a diagram illustrating an example of output processing according to a modified example. FIG. 15 is a hardware configuration diagram showing an example of a computer that implements the functions of the output device.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Below, embodiments (hereinafter referred to as "embodiments") for implementing the output program, output device, and output method according to the present application will be described in detail with reference to the drawings. Note that this embodiment does not limit the output program, output device, and output method according to the present application. Further, in each of the embodiments below, the same parts are given the same reference numerals, and redundant explanations will be omitted.

[1. Example of output processing]
First, an example of output processing according to the embodiment will be described using FIG. 1. FIG. 1 is a diagram illustrating an example of output processing according to an embodiment. As illustrated in FIG. 1, the output processing system 1 includes a content distribution device 20 and an output device 100. These various devices are communicably connected via a network N (for example, the Internet) by wire or wirelessly. Note that the output processing system 1 shown in FIG. 2 may include any number of content distribution devices 20 and any number of output devices 100.

The content distribution device 20 is an information processing device that distributes content. In the example shown in FIG. 1, the content distribution device 20 distributes a news article, which is an example of content displayed on the output device 100, in response to a distribution request from the output device 100.

The output device 100 is a non-task-oriented dialogue system, which is a system for performing non-task-oriented dialogues such as small talk and casual conversations with the user. Note that the task-oriented dialogue system refers to a system that performs tasks such as searching, reading news, and making reservations in response to user requests. The output device 100 is a type of chatbot, which is a program that automatically interacts with the user through text and voice. The output device 100 performs information processing that acquires reaction information that is character information generated as a reaction to second information other than first information input by the user into the dialogue system, and outputs the acquired reaction information to the user. It is a device.

Further, the output device 100 is an input/output device that has an acquisition device such as a microphone that acquires surrounding sounds, and an input/output device that has an output device such as a speaker that can output any sound. It is a device called a speaker. Further, the output device 100 is a device that can output music and provide information by voice. Further, the output device 100 has a reception function that accepts input of sound, and an output function that outputs a sound according to the content of the acquired voice when the voice uttered by the user is acquired. Note that the output device 100 may be various information processing devices such as a smartphone, a tablet, etc., in addition to a smart speaker. FIG. 1 shows an example in which the output device 100 is a smartphone. Furthermore, the output device 100 does not necessarily need to include an output device such as a speaker. For example, the output device 100 may output various information to the user by displaying information such as character strings instead of audio.

Further, the output device 100 includes a physical sensor that is a sensor that detects the physical condition of the user who uses the output device 100. For example, the output device 100 includes any physical sensor such as a gyro sensor, an image sensor, an acceleration sensor, a temperature sensor, a volume sensor, and a brightness sensor. The output device 100 detects the physical state of the user using the output device 100 using a physical sensor. Furthermore, the output device 100 acquires sensor information, which is information regarding the physical state of the user, using a physical sensor. The output device 100 determines the timing to output reaction information to the user based on the acquired sensor information. In the example shown in FIG. 1, the output device 100 determines whether or not it is the timing to output reaction information to the user by voice, based on the acquired sensor information.

From here, the flow of output processing will be explained using FIG. 1. FIG. 1 is a diagram illustrating an example of output processing according to an embodiment. In the example shown in FIG. 1, the output device 100 uses the volume sensor to acquire audio information regarding the user's voice indicating the user's utterance (step S11). Upon acquiring the audio information, the output device 100 determines whether it is time to output reaction information to the user based on the acquired audio information (step S12). Specifically, the output device 100 determines whether or not a predetermined period of time has elapsed since the user's utterance to the dialogue system. For example, when the output device 100 determines that a predetermined period (for example, 30 seconds) has elapsed since the user's utterance to the dialogue system based on the acquired audio information, the output device 100 outputs reaction information to the user. It is determined that the timing is right. Subsequently, when the output device 100 determines that it is the timing to output the reaction information to the user, it transmits a content distribution request to the content distribution device 20 (step S13).

The content distribution device 20 receives a content distribution request from the output device 100. When the content distribution device 20 receives a content distribution request, the content distribution device 20 identifies the user who uses the output device 100 that is the request source. Subsequently, upon identifying the user, the content distribution device 20 selects content according to the user (step S14). When content is selected, the content distribution device 20 distributes the selected content to the output device 100. Specifically, the content distribution device 20 distributes content #1, which is a news article regarding the formation of a labor union at Company A, which is an online mail order company, to the output device 100.

The output device 100 acquires content #1 from the content distribution device 20 (step S15). When the output device 100 acquires the content #1, the output device 100 inputs the acquired content #1 into the model M1 to generate reaction information #1, which is text information indicating a reaction to the content #1 (step S16). Specifically, the output device 100 generates summary information R11 that is text information corresponding to the summary of content #1, ``It appears that a labor union has been formed at company A.''. Furthermore, the output device 100 generates trigger information R12, which is character information that triggers a dialogue with the user related to content #1. More specifically, the output device 100 generates trigger information R12 that is ``Will there be a strike?'' which is text information requesting the user's opinion regarding content #1. The output device 100 generates reaction information #1 including summary information R11 and trigger information R12. When the output device 100 generates the reaction information #1, the output device 100 acquires the generated reaction information #1. In FIG. 1, an example has been described in which the reaction information #1 is character information including both the summary information R11 and the trigger information R12, but the reaction information #1 does not need to include the summary information R11. That is, the reaction information #1 may be the trigger information R12.

Upon acquiring the reaction information #1, the output device 100 outputs the reaction information #1 as voice (step S17). Specifically, the output device 100 outputs the trigger information R12 as voice after outputting the summary information R11 as voice.

As described above, the output device 100 acquires reaction information that is character information generated as a reaction to second information other than first information input by the user into the dialogue system. Furthermore, the output device 100 outputs the acquired reaction information to the user. Thereby, the output device 100 can spontaneously output character information that becomes an opportunity for dialogue related to news articles, etc., without any action from the user on the dialogue system. That is, the output device 100 allows the device to spontaneously and appropriately talk to a person. Therefore, the output device 100 can output appropriate information without depending on the user's input to the dialogue system.

[2. Output device configuration]
Next, the configuration of the output device 100 according to the embodiment will be described using FIG. 2. FIG. 2 is a diagram showing a configuration example of the output device 100 according to the embodiment. As shown in FIG. 2, the output device 100 includes a communication section 110, a storage section 120, an input section 130, a display section 140, a physical sensor 150, and a control section 160.

(Communication Department 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card). The communication unit 110 is connected to the network N by wire or wirelessly, and transmits and receives information to and from the content distribution device 20.

(Storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 14 stores, for example, output programs and the like. Furthermore, as shown in FIG. 2, the storage unit 14 includes a sensor information storage unit 121, a content storage unit 122, a reaction information storage unit 123, a model information storage unit 124, and a learning data storage unit 125.

(Sensor information storage unit 121)
The sensor information storage unit 121 stores various information related to detection by physical sensors. For example, the sensor information storage unit 121 stores sensor information detected by a physical sensor. An example of the sensor information storage unit according to the embodiment will be described using FIG. 3. FIG. 3 is a diagram illustrating an example of a sensor information storage unit according to the embodiment. In the example shown in FIG. 3, the sensor information storage unit 121 has items such as "log ID", "type", "sensor information", "date and time", and "position".

"Log ID" indicates identification information for identifying acquired sensor information. “Type” indicates information regarding the type of detected sensor information. "Sensor information" indicates detected sensor information. Figure 3 shows an example where conceptual information such as "SN11" is stored in "sensor information", but in reality, it is actually a file indicating specifically detected information (sensor data) or its storage location. A path name, etc. may also be stored.

For example, if the type of sensor information is "image", "sensor information" stores a still image or a moving image of the user, or a file path name indicating a storage location thereof. For example, if the type of sensor information is "utterance", "sensor information" includes audio data indicating the user's utterances, character information after converting the audio data, or a file path name indicating the storage location of these data. etc. are stored.

“Date and time” indicates the date and time when the sensor information was detected. In FIG. 3, the "date and time" is illustrated abstractly such as "dt11", but a specific date and time such as "March 23, 2019 22:31:52" may be stored. Note that the "date and time" is not limited to a time, but may be a predetermined period such as "March 23, 2019 22:31" or "March 23, 2019 22:30-31". Moreover, "position" indicates the position where the corresponding action was performed. In addition, in FIG. 3, the "position" is illustrated with an abstract code such as "LC11", but information indicating latitude and longitude and information indicating an address such as "B intersection, Z town, Y city, X prefecture" are also stored. You can.

(Content storage unit 122)
The content storage unit 122 stores various information regarding content. For example, the content storage unit 122 stores information regarding content acquired from the content distribution device 20. For example, the content storage unit 122 stores information related to content such as news articles, posted information on SNS (Social Networking Service), advertisements, and reviews of movies and the like. For example, the content storage unit 122 stores HTML files forming the content, text information, still images, and moving images displayed in the content. An example of the content storage unit according to the embodiment will be described using FIG. 4. FIG. 4 is a diagram illustrating an example of a content storage unit according to the embodiment. In the example shown in FIG. 4, the content storage unit 122 has items such as "content ID", "content", and "acquisition date and time".

"Content ID" indicates identification information for identifying content. "Content" indicates character information included in content such as news articles, posted information on SNS, advertisements, and movie reviews. Although FIG. 4 shows an example in which conceptual information such as "Content #11" is stored in "Content," it actually indicates the specific content (the HTML file that forms it) or its storage location. File path names, etc. may also be stored. Further, "content" may store still images and moving images displayed in news articles, posted information on SNS, advertisements, movie reviews, etc. Alternatively, "content" may store character information extracted from image information as a result of analyzing image information displayed in news articles, posted information on SNS, advertisements, reviews of movies, etc. “Acquisition date and time” indicates the date and time when the content was acquired.

(Reaction information storage unit 123)
The reaction information storage unit 123 stores various information related to reaction information. For example, the reaction information storage unit 123 stores reaction information generated by the output device 100. An example of the reaction information storage unit according to the embodiment will be described using FIG. 5. FIG. 5 is a diagram illustrating an example of a reaction information storage unit according to the embodiment. In the example shown in FIG. 5, the reaction information storage unit 123 has items such as "reaction ID", "reaction information", "response by user", "generation date and time", and "content ID".

"Reaction ID" indicates identification information for identifying reaction information. "Reaction information" indicates character information indicating a reaction to the content identified by the content ID. "Content ID" indicates identification information for identifying content that is a target of a reaction. “Response by user” indicates whether there is a response from the user to the reaction information. Specifically, "1" is stored when there is a response from the user to the reaction information, and "0" is stored when there is no response from the user to the reaction information. “Generation date and time” indicates the date and time when the reaction information was generated.

(Model information storage unit 124)
The model information storage unit 124 stores various information regarding a model that generates reaction information that is character information indicating a reaction to second information other than the first information input by the user into the dialogue system. For example, the model information storage unit 124 stores a model that, when second information other than the first information input by the user is input to the dialogue system, outputs reaction information that is character information indicating a reaction to the second information. An example of the model information storage unit according to the embodiment will be described using FIG. 6. FIG. 6 is a diagram illustrating an example of a model information storage unit according to the embodiment. In the example shown in FIG. 6, the model information storage unit 124 has items such as "model ID" and "model data."

"Model ID" indicates identification information for identifying a model that generates reaction information. "Model data" indicates model data of a model. For example, "model data" stores data for generating reaction information, which is character information indicating a reaction to the second information, from second information other than the first information input by the user into the dialogue system.

Here, an example of the model according to the embodiment will be described using FIG. 7. FIG. 7 shows an example of a model according to the embodiment. In FIG. 7, an example is shown in which the model M1 that generates reaction information, which is character information indicating a reaction to the second information, from second information other than the first information input by the user to the dialogue system is Seq2Seq (Sequence to Sequence Model). I will explain about it. Seq2Seq is a model that learns rules for inputting a word sequence (character string) and outputting another word sequence (character string). Specifically, Seq2Seq is an encoder-decoder model whose constituent elements are LSTM (Long Short Term Memory), which is a type of RNN (Recurrent Neural Network).

In the example shown in FIG. 7, a character string "ABC" is input and a character string "WXYZ" is output. Applying the example shown in Figure 1, ABC reports, ``A labor union has been formed at major online shopping company A. It is calling for the abolition of ``arbitrary personnel evaluations''. This is the text information of content #1, which is a news article. Furthermore, "WXYZ" is text information indicating a reaction to content #1, "It seems that a labor union has been formed at Company A. Will there be a strike?" If there is such a pair of "ABC" (content #1 which is a news article) and "WXYZ" (reaction information #1 indicating a response to content #1), the model M1 which is Seq2Seq will be " The RNN is trained to vectorize "ABC" on the encoder side to which "ABC" is input, and to generate "WXYZ" on the decoder side to which "WXYZ" in FIG. 7 is output.

The model data MDT1 includes an input layer into which second information is input, an output layer, a first element belonging to any layer from the input layer to the output layer other than the output layer, and a first element. and a second element whose value is calculated based on the weight of the first element, reaction information indicating a reaction to the second information input to the input layer according to the second information input to the input layer. The output device 100 may be operated to output from the output layer.

Here, it is assumed that the model data MDT1 is realized by a regression model expressed as "y=a1*x1+a2*x2+...+ai*xi". In this case, the first element included in the model data MDT1 corresponds to input data (xi) such as x1 and x2. Further, the weight of the first element corresponds to the coefficient ai corresponding to xi. Here, the regression model can be regarded as a simple perceptron having an input layer and an output layer. When each model is regarded as a simple perceptron, the first element corresponds to any node that the input layer has, and the second element can be regarded as a node that the output layer has.

Further, it is assumed that the model data MDT1 is realized by a neural network having one or more intermediate layers, such as a DNN (Deep Neural Network). In this case, the first element included in the model data MDT1 corresponds to any node included in the input layer or the intermediate layer. Further, the second element corresponds to a next-stage node that is a node to which a value is transmitted from the node corresponding to the first element. Furthermore, the weight of the first element corresponds to a connection coefficient that is a weight that is considered for the value transmitted from the node corresponding to the first element to the node corresponding to the second element.

The output device 100 outputs reaction information using a model having an arbitrary structure, such as the regression model or neural network described above. Specifically, coefficients are set in the model data MDT1 so that, when the second information is input, reaction information indicating a reaction to the second information is output. The output device 100 uses such model data MDT1 to output reaction information indicating a reaction to the second information.

In the above example, the model data MDT1 is a model (hereinafter referred to as model X) that outputs reaction information indicating a reaction to the second information when the second information is input. However, the model data MDT1 according to the embodiment may be a model generated based on the results obtained by repeatedly inputting and outputting data to the model X. For example, model data MDT1 is a model (hereinafter referred to as model Y) that is learned by inputting reaction information indicating a reaction to the second information outputted by model X when second information is input. Good too. Alternatively, the model data MDT1 may be a model learned to input the second information and output the output value of the model Y.

Furthermore, when the output device 100 performs estimation processing using GAN (Generative Adversarial Networks), the model data MDT1 may be a model that forms part of the GAN.

(Learning data storage unit 125)
The learning data storage unit 125 stores various information regarding the learning data (correct data) of the model. For example, the learning data storage unit 125 stores content and reaction information that is text information indicating a reaction to the content. An example of the learning data storage unit according to the embodiment will be described using FIG. 8. FIG. 8 is a diagram illustrating an example of a learning data storage unit according to the embodiment. In the example shown in FIG. 8, the learning data storage unit 125 has items such as "learning data ID", "content", "reaction information", and "date and time".

"Learning data ID" indicates identification information for identifying learning data. "Content" indicates character information included in content such as news articles, posted information on SNS (Social Networking Service), advertisements, and movie reviews. Although FIG. 8 shows an example in which conceptual information such as "Content #1-1" is stored in "Content", in reality, the specific content (the HTML file forming it) or its storage location is stored in "Content". A file path name indicating the file name, etc. may also be stored. Further, "content" may store still images and moving images displayed in news articles, posted information on SNS, advertisements, movie reviews, etc. Alternatively, "content" may store character information extracted from image information as a result of analyzing image information displayed in news articles, posted information on SNS, advertisements, reviews of movies, etc. "Reaction information" indicates text information indicating a reaction to the content. "Date and time" indicates the date and time when the model learned the learning data.

(Input section 130)
The input unit 130 is an input device that accepts various operations from the user. For example, the input unit 130 is realized by a keyboard, a mouse, operation keys, or the like.

(Display section 140)
The display unit 140 is a display device for displaying various information, that is, a screen. For example, the display unit 140 is realized by a liquid crystal display or the like. Note that when a touch panel is adopted as the output device 100, the input section 130 and the display section 140 are integrated. Furthermore, in the following description, the display unit 140 may be referred to as a screen.

(Physical sensor 150)
The physical sensor 150 is a sensor that detects the physical condition of the user who uses the output device 100. For example, the physical sensor 150 can be any sensor such as a gyro sensor, an image sensor, an acceleration sensor, a temperature sensor, a volume sensor, a brightness sensor, etc. that measure the inclination of the output device 100 in three axial directions.

The physical sensor 150 acquires sensor information that is information regarding the physical state of the user who uses the output device 100. Upon acquiring the sensor information, the physical sensor 150 stores the acquired sensor information in the sensor information storage section 121. Specifically, the physical sensor 150, which is a volume sensor, acquires audio information regarding the user's voice indicating the user's utterance. When the physical sensor 150, which is a volume sensor, acquires audio information, it stores the acquired audio information in the sensor information storage unit 121. Further, the physical sensor 150, which is an image sensor, acquires image information regarding the user's image. When acquiring image information, the physical sensor 150 that is an image sensor stores the acquired image information in the sensor information storage unit 121. Further, the physical sensor 150, which is an acceleration sensor, acquires acceleration information regarding acceleration accompanying the user's movement. Upon acquiring acceleration information, the physical sensor 150 that is an acceleration sensor stores the acquired acceleration information in the sensor information storage unit 121 .

(Control unit 160)
The control unit 160 executes various programs (corresponding to an example of an output program) stored in a storage device inside the output device 100 by using a RAM as a work area by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. This is achieved by For example, these various programs correspond to an application program called a web browser. Further, the control unit 160 is realized by, for example, an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

As shown in FIG. 2, the control unit 160 includes a determination unit 161, a request unit 162, a generation unit 163, an acquisition unit 164, and an output unit 165, and has information processing functions and operations described below. to realize or carry out. Note that the internal configuration of the control unit 160 is not limited to the configuration shown in FIG. 2, and may be any other configuration as long as it performs information processing to be described later.

(Determination unit 161)
The determining unit 161 determines whether it is the timing to output reaction information to the user. Specifically, the determination unit 161 refers to the sensor information storage unit 121 and acquires sensor information that is information regarding the physical state of the user who uses the output device 100. Subsequently, upon acquiring the sensor information, the determination unit 161 determines whether it is the timing to output reaction information to the user based on the acquired sensor information.

Further, the determining unit 161 determines whether the timing is other than when the user requests the second information to the dialogue system. More specifically, the output unit 165 determines whether the timing is based on the user's context. In the example shown in FIG. 1, the output unit 165 determines whether or not a predetermined period of time has elapsed since the user's utterance to the dialogue system. In the example shown in FIG. 1, the determination unit 161 refers to the sensor information storage unit 121 and acquires audio information such as log LG1. Subsequently, upon acquiring the voice information, the determination unit 161 determines whether a predetermined period (for example, 30 seconds) has elapsed since the user's utterance to the dialogue system, based on the acquired voice information. . If the determining unit 161 determines that it is the timing when a predetermined period (for example, 30 seconds) has elapsed since the user's utterance to the dialogue system, it determines that it is the timing to output reaction information to the user.

Note that the determination unit 161 may determine whether it is the timing to output reaction information to the user based on the user's location information. For example, the determination unit 161 refers to the position information stored in “position” of the sensor information storage unit 121 and acquires the position information of the user who uses the output device 100. Next, the determining unit 161 determines whether the user is indoors or outdoors based on the acquired position information. For example, when determining that the user is at the user's home, the determination unit 161 may determine that it is the timing to output reaction information to the user. Furthermore, when determining that the user is in the park, the determination unit 161 may determine that it is the timing to output reaction information to the user.

(Request unit 162)
The request unit 162 requests distribution of content. Specifically, when the determining unit 161 determines that it is the timing to output the reaction information to the user, the requesting unit 162 transmits a content distribution request to the content distribution device 20. Further, the requesting unit 162 receives content from the content distribution device 20. Upon receiving the content, the requesting unit 162 stores the received content in the content storage unit 122.

(Generation unit 163)
The generation unit 163 generates reaction information from the second information. Specifically, the generation unit 163 acquires second information other than the first information input by the user into the dialogue system. More specifically, the generation unit 163 acquires second information other than the first information input by the user into the dialogue system, which is a non-task-oriented dialogue system. The generation unit 163 also acquires second information selected according to the user. For example, the generation unit 163 acquires second information that is predetermined content. In the example shown in FIG. 1, the generation unit 163 acquires second information that is a news article. More specifically, the generation unit 163 refers to the content storage unit 122 and obtains content #1, which is a news article.

Subsequently, upon acquiring the second information, the generation unit 163 generates reaction information by inputting the acquired second information into a predetermined model. After generating reaction information, the generation unit 163 stores the generated reaction information in the reaction information storage unit 123. More specifically, upon acquiring the second information, the generation unit 163 refers to the model information storage unit 124 and acquires the model M1. For example, when the generation unit 163 inputs second information other than the first information input by the user to the dialogue system, it acquires the model M1 that outputs reaction information that is textual information indicating a reaction to the second information. For example, when the generation unit 163 inputs second information other than the first information input by the user to the dialogue system, the generation unit 163 acquires a model M1 that is a Seq2Seq that outputs reaction information that is character information indicating a reaction to the second information. do. Subsequently, upon acquiring the model M1, the generation unit 163 inputs the second information to the acquired model M1. The generation unit 163 acquires the output information of the model M1 as reaction information. In this way, the generation unit 163 generates reaction information that is character information indicating a reaction to the second information other than the first information input by the user into the dialogue system. More specifically, the generation unit 163 generates reaction information for second information other than the first information input by the user into the dialog system, which is a non-task-oriented dialog system. Furthermore, the generation unit 163 generates reaction information for the second information selected according to the user. For example, the generation unit 163 generates reaction information for second information that is predetermined content. In the example shown in FIG. 1, the generation unit 163 generates reaction information for the second information, which is a news article. More specifically, the generation unit 163 generates reaction information indicating a reaction to content #1, which is a news article selected by the content distribution device 20 according to the user.

Furthermore, the generation unit 163 generates summary information that is a summary of the second information. In the example shown in FIG. 1, the generation unit 163 generates summary information R11, which is text information corresponding to the summary of content #1, which is a news article, ``It seems that a labor union has been formed at company A.''. Further, the generation unit 163 generates reaction information that becomes an opportunity for dialogue related to the second information with the user. For example, the generation unit 163 acquires reaction information that requests the user's opinion on the second information. In the example shown in FIG. 1, the generation unit 163 generates opportunity information R12, which is text information that requests the user's opinion on content #1, which is a news article, ``Will there be a strike?''. The generation unit 163 generates reaction information #1 including summary information R11 and trigger information R12. Although FIG. 1 describes an example in which the generation unit 163 generates reaction information #1 that includes both summary information R11 and trigger information R12, it is also possible to generate reaction information #1 that does not include summary information R11. good. That is, the generation unit 163 may generate reaction information #1, which is the trigger information R12.

(Acquisition unit 164)
The acquisition unit 164 acquires the reaction information generated by the generation unit 163. Specifically, the acquisition unit 164 acquires reaction information that is character information generated as a reaction to second information other than the first information input by the user into the dialogue system. More specifically, the acquisition unit 164 acquires reaction information to second information other than the first information input by the user to the dialogue system, which is a non-task-oriented dialogue system. do. For example, the acquisition unit 164 acquires reaction information for second information that is predetermined content. The acquisition unit 164 also acquires reaction information for the second information selected according to the user. In the example shown in FIG. 1, the acquisition unit 164 acquires reaction information for the second information, which is a news article. More specifically, the acquisition unit 164 acquires reaction information indicating a reaction to content #1, which is a news article selected by the content distribution device 20 according to the user.

The acquisition unit 164 also acquires summary information that is a summary of the second information. In the example shown in FIG. 1, the acquisition unit 164 acquires summary information R11, which is text information corresponding to the summary of content #1, which is a news article, ``It seems that a labor union has been formed at company A.''. The acquisition unit 164 also acquires reaction information that becomes a trigger for dialogue with the user regarding the second information. For example, the acquisition unit 164 acquires reaction information that requests the user's opinion regarding the second information. In the example shown in FIG. 1, the acquisition unit 164 acquires opportunity information R12, which is text information that requests the user's opinion on content #1, which is a news article, ``Will there be a strike?''. The acquisition unit 164 acquires reaction information #1 including summary information R11 and trigger information R12. Although FIG. 1 describes an example in which the acquisition unit 164 acquires reaction information #1 that includes both summary information R11 and trigger information R12, it is possible to acquire reaction information #1 that does not include summary information R11. good. That is, the acquisition unit 164 may acquire reaction information #1, which is the trigger information R12.

(Output section 165)
The output unit 165 outputs the reaction information acquired by the acquisition unit 164 to the user. Specifically, the output unit 165 outputs the reaction information acquired by the acquisition unit 164 to the user when the determination unit 161 determines that it is the timing to output the reaction information to the user. More specifically, the output unit 165 outputs the reaction information at a timing other than the user's request for the second information to the dialogue system. For example, the output unit 165 outputs reaction information at a timing based on the user's context. In the example shown in FIG. 1, the output unit 165 outputs reaction information at a timing when a predetermined period of time has elapsed since the user's utterance to the dialogue system.

Furthermore, the output unit 165 outputs summary information and reaction information. More specifically, the output unit 165 outputs the reaction information after outputting the summary information. Furthermore, the output unit 165 outputs the reaction information as voice. In the example shown in FIG. 1, the output unit 165 outputs the summary information R11 in voice, and then outputs the trigger information R12 in voice.

[3. Output processing flow]
Next, the output processing procedure according to the embodiment will be described using FIG. 9. FIG. 9 is a flowchart showing the output processing procedure according to the first embodiment. In the example shown in FIG. 9, the output device 100 determines whether it is the timing to output reaction information based on the sensor information (step S101). If the output device 100 determines that it is not the timing to output the reaction information (step S101; No), it ends the process.

On the other hand, when the output device 100 determines that it is the timing to output the reaction information (step S101; Yes), it acquires predetermined content from the content distribution device 20 (step S102). Subsequently, upon acquiring the predetermined content, the output device 100 generates reaction information for the predetermined content (step S103). Subsequently, after generating reaction information, the output device 100 outputs the generated reaction information (step S104).

[4. Configuration of content distribution device]
Next, the configuration of the content distribution device 20 according to the embodiment will be described using FIG. 10. FIG. 10 is a diagram illustrating a configuration example of the content distribution device 20 according to the embodiment. As shown in FIG. 10, the content distribution device 20 includes a communication section 21, a storage section 22, and a control section 23.

(Communication Department 21)
The communication unit 21 is realized by, for example, a NIC or the like. The communication unit 21 is connected to the network N by wire or wirelessly, and transmits and receives information to and from the output device 100.

(Storage unit 22)
The storage unit 22 is realized by, for example, a semiconductor memory device such as a RAM or a flash memory, or a storage device such as a hard disk or an optical disk. Furthermore, the storage unit 14 includes a user information storage unit 221 and a content storage unit 222, as shown in FIG.

(User information storage unit 221)
The user information storage unit 221 stores various information regarding users. An example of the user information storage unit according to the embodiment will be described using FIG. 11. FIG. 11 is a diagram illustrating an example of a user information storage unit according to the embodiment. In the example shown in FIG. 11, the user information storage unit 221 has items such as "user ID" and "user information."

"User ID" indicates identification information for identifying a user. "User information" indicates information regarding the user. For example, "user information" is keywords related to genres, topics, hobbies, and preferences that the user is interested in. Note that the "user information" may include demographic information of the user such as the user's age, gender, place of residence, family life cycle, income, and occupation. Further, the "user information" may include psychographic information such as the user's values, lifestyle, personality, and preferences.

(Content storage unit 222)
The content storage unit 222 stores various information regarding content. For example, the content storage unit 222 stores information related to content such as news articles, posted information on SNS, advertisements, and reviews of movies and the like. For example, the content storage unit 222 stores HTML files forming the content, and still images and moving images displayed in the content. An example of the content storage unit according to the embodiment will be described using FIG. 12. FIG. 12 is a diagram illustrating an example of a content storage unit according to the embodiment. In the example shown in FIG. 12, the content storage unit 222 has items such as "content ID", "content", and "type".

"Content ID" indicates identification information for identifying content. "Content" indicates character information included in content such as news articles, posted information on SNS, advertisements, and movie reviews. Although FIG. 12 shows an example in which conceptual information such as "Content #1" is stored in "Content," it actually indicates the specific content (the HTML file forming it) or its storage location. File path names, etc. may also be stored. Further, "content" may store still images and moving images displayed in news articles, posted information on SNS, advertisements, movie reviews, etc. Alternatively, "content" may store character information extracted from image information as a result of analyzing image information displayed in news articles, posted information on SNS, advertisements, reviews of movies, etc. “Type” indicates the type of content. For example, "type" indicates the type of content, such as a news article, posted information on SNS, an advertisement, or a movie review.

(Control unit 23)
The control unit 23 is realized by, for example, a CPU, an MPU, or the like executing various programs stored in a storage device inside the content distribution device 20 using the RAM as a work area. Further, the control unit 23 is realized by, for example, an integrated circuit such as an ASIC or an FPGA.

As shown in FIG. 10, the control unit 23 includes a reception unit 231, a selection unit 232, and a distribution unit 233, and implements or executes information processing functions and operations described below. Note that the internal configuration of the control unit 23 is not limited to the configuration shown in FIG. 10, and may be any other configuration as long as it performs information processing to be described later.

(Reception Department 231)
The receiving unit 231 receives a content distribution request. The reception unit 231 receives a content distribution request. Specifically, the reception unit 231 receives a content distribution request from the output device 100.

(Selection section 232)
The selection unit 232 selects content according to the user. Specifically, when the accepting unit 231 accepts a content distribution request, the selecting unit 232 identifies the user who uses the requesting output device 100 . For example, the selection unit 232 refers to the user information storage unit 221 to identify the user who uses the requesting output device 100. Subsequently, upon specifying the user, the selection unit 232 refers to the user information storage unit 221 and acquires user information associated with the specified user. Upon acquiring the user information, the selection unit 232 selects content appropriate for the user based on the acquired user information. For example, the selection unit 232 acquires keywords related to genres, topics, hobbies, and preferences that the user is interested in, based on the user information. Subsequently, the selection unit 232 refers to the content storage unit 222 and selects the content set with the acquired keyword as the content appropriate for the user.

In the example shown in FIG. 1, when the selection unit 232 specifies the user U1, the selection unit 232 refers to the user information storage unit 221 and acquires user information #1 associated with the specified user U1. Subsequently, upon acquiring the user information #1, the selection unit 232 acquires keywords related to genres, topics, hobbies, and preferences that the user U1 is interested in, based on the acquired user information #1. Subsequently, the selection unit 232 refers to the content storage unit 222 and selects content #1, which is a news article to which the acquired keyword is set, as the content corresponding to the user U1.

(Distribution department 233)
The distribution unit 233 distributes the content selected by the selection unit 232. Specifically, the distribution unit 233 distributes the content selected by the selection unit 232 to the output device 100. In the example shown in FIG. 1, the distribution unit 233 distributes content #1 selected by the selection unit 232 to the output device 100.

[5. Modified example]
The output processing system 1 according to the embodiment described above may be implemented in various different forms other than the embodiment described above. Therefore, other embodiments of the output processing system 1 will be described below. Note that the same parts as those in the embodiment are given the same reference numerals and the description thereof will be omitted.

[5-1. Output processing of reaction information showing impressions of the dialogue system]
In FIG. 1, an example has been described in which the output device 100 outputs the trigger information R12, "Will there be a strike?", which is text information requesting the user's opinion on content #1. The information is not limited to text information requested from the user. In FIG. 13, an example will be described in which the output device 100 outputs reaction information indicating impressions on the second information of the dialogue system. In the example shown in FIG. 13, the acquisition unit 164 acquires reaction information indicating impressions regarding the second information of the dialogue system.

An example of output processing according to the modified example will now be described using FIG. 13. FIG. 13 is a diagram illustrating an example of output processing according to a modification. In the example shown in FIG. 13, the determination unit 161 refers to the sensor information storage unit 121 and acquires image information regarding the user's image (step S21). Upon acquiring the image information, the determining unit 161 determines whether it is time to output reaction information to the user based on the acquired image information (step S22). Specifically, if the determination unit 161 determines that the user is in a stationary state for a predetermined period of time or more (for example, 5 minutes or more) based on the acquired image information, the determination unit 161 determines at the timing of outputting reaction information to the user. It is determined that there is. Subsequently, when the determining unit 161 determines that it is the timing to output the reaction information to the user, the requesting unit 162 transmits a content distribution request to the content distribution device 20 (step S23).

The content distribution device 20 acquires a content distribution request from the output device 100. Upon acquiring the content distribution request, the content distribution device 20 identifies the user who uses the output device 100 that is the request source. Subsequently, upon identifying the user, the content distribution device 20 selects content according to the user (step S24). When content is selected, the content distribution device 20 distributes the selected content to the output device 100. Specifically, the content distribution device 20 distributes content #2, which is a news article about the government distributing tablet terminals to all evacuation centers in the disaster area, to the output device 100.

The generation unit 163 acquires content #2 from the content distribution device 20 (step S25). Upon acquiring content #2, the generating unit 163 generates reaction information #2, which is text information indicating a reaction to content #2, by inputting the acquired content #2 into model M1 (step S26). Specifically, the generation unit 163 generates summary information R21 that is text information corresponding to the summary of content #2, ``It seems that the government will distribute tablet terminals to all evacuation centers.'' The generation unit 163 also generates trigger information R22, which is text information that triggers a dialogue with the user related to content #2. More specifically, the generation unit 163 generates trigger information R22, which is text information indicating the impression of content #2 of the dialogue system, "I hope it can be used even during a power outage." The generation unit 163 generates reaction information #2 including summary information R21 and trigger information R22. The acquisition unit 164 acquires reaction information #2 generated by the generation unit 163. Although FIG. 13 describes an example in which the generation unit 163 generates reaction information #2 that includes both summary information R21 and trigger information R22, it is also possible to generate reaction information #2 that does not include summary information R21. good. That is, the generation unit 163 may generate reaction information #2, which is the trigger information R22.

The output unit 165 outputs reaction information #2 acquired by the acquisition unit 164 as audio (step S27). Specifically, the output device 100 outputs the trigger information R22 as voice after outputting the summary information R21 as voice.

[5-2. Output processing of reaction information that prompts user actions]
In FIG. 1, an example has been described in which the output device 100 outputs reaction information #1 including trigger information R12 "Will there be a strike?", which is text information requesting the user's opinion on content #1. The information is not limited to text information that requests the user's opinion on the content. In FIG. 14, an example will be described in which the output device 100 outputs reaction information that prompts the user to take an action related to the second information. In the example shown in FIG. 14, the acquisition unit 164 acquires reaction information that prompts the user to take action related to the second information. Specifically, the acquisition unit 164 acquires reaction information that prompts the user to take an action other than responding to the reaction information.

Further, in FIG. 1, an example has been described in which the output device 100 acquires reaction information #2 for content #1 that is a news article, but the output device 100 may acquire reaction information for second information that is an advertisement. . In the example shown in FIG. 14, the acquisition unit 164 acquires reaction information for the second information, which is an advertisement.

An example of output processing according to the modified example will now be described using FIG. 14. FIG. 14 is a diagram illustrating an example of output processing according to a modified example. In the example shown in FIG. 14, the determination unit 161 refers to the sensor information storage unit 121 and acquires acceleration information regarding acceleration accompanying the user's movement (step S31). Upon acquiring the acceleration information, the determination unit 161 determines whether it is the timing to output reaction information to the user based on the acquired acceleration information (step S32). Specifically, if the determination unit 161 determines that the user is in a stationary state for a predetermined period of time or more (for example, 5 minutes or more) based on the acquired acceleration information, the determination unit 161 determines that the user is in a stationary state for a predetermined period of time or more (for example, 5 minutes or more). It is determined that there is. Subsequently, when the determining unit 161 determines that it is the timing to output the reaction information to the user, the requesting unit 162 transmits a content distribution request to the content distribution device 20 (step S33).

The content distribution device 20 acquires a content distribution request from the output device 100. Upon acquiring the content distribution request, the content distribution device 20 identifies the user who uses the output device 100 that is the request source. Subsequently, upon identifying the user, the content distribution device 20 selects content according to the user (step S34). When content is selected, the content distribution device 20 distributes the selected content to the output device 100. Specifically, the content distribution device 20 distributes content #3, which is an advertisement regarding the newly released product C, to the output device 100.

The generation unit 163 acquires content #3 from the content distribution device 20 (step S35). Upon acquiring content #3, the generation unit 163 generates reaction information #3, which is text information indicating a reaction to content #3, by inputting the acquired content #3 into model M1 (step S36). Specifically, the generation unit 163 generates summary information R31 that is text information corresponding to the summary of content #3, ``It looks like product C has been released.''. Furthermore, the generation unit 163 generates opportunity information R32, which is text information that becomes an opportunity for interaction with the user regarding content #3. More specifically, the generation unit 163 generates trigger information R32, which is text information that prompts the user to take an action related to content #3. For example, the generation unit 163 generates opportunity information R32 that is text information that prompts the user to take an action other than responding to reaction information #3, ``Would you like to go shopping with me next time?''. The generation unit 163 generates reaction information #2 including summary information R31 and trigger information R32. The acquisition unit 164 acquires reaction information #3 generated by the generation unit 163. Although FIG. 14 describes an example in which the generation unit 163 generates reaction information #3 that includes both summary information R31 and trigger information R32, it is also possible to generate reaction information #3 that does not include summary information R31. good. That is, the generation unit 163 may generate reaction information #3, which is the trigger information R32.

The output unit 165 outputs reaction information #3 acquired by the acquisition unit 164 as audio (step S37). Specifically, the output device 100 outputs the trigger information R32 as voice after outputting the summary information R31 as voice.

Note that although FIG. 14 describes an example in which the output device 100 acquires reaction information #2 for content #2, which is an advertisement, as content other than news articles, content other than news articles is not limited to advertisements. Specifically, the acquisition unit 164 acquires reaction information to the second information, which is posted information on a predetermined SNS (Social Networking Service).

[5-3. Output at timing related to push notifications]
The output unit 165 outputs reaction information at a timing related to push notification in the dialogue system. Specifically, the determination unit 161 acquires information regarding push notifications in the dialog system. For example, the determination unit 161 acquires information that push notifications are to be made at pre-programmed predetermined intervals as information regarding push notifications in the interaction system. Subsequently, upon acquiring the information regarding the push notification, the determination unit 161 determines whether or not it is the timing regarding the push notification in the dialog system based on the acquired information regarding the push notification. The output unit 165 outputs reaction information when the determination unit 161 determines that the timing is related to a push notification in the dialog system.

[5-4. Output at the time of search]
The output unit 165 outputs reaction information at a timing when a user's behavior other than the behavior toward the dialogue system and related to the second information is performed. Specifically, the output unit 165 outputs the reaction information at the timing when the user performs a search related to the second information. More specifically, the determination unit 161 acquires search information regarding the user's search behavior from a search server (not shown). Subsequently, upon acquiring the search information, the determination unit 161 determines whether or not it is the timing at which the user performed a search related to the second information, based on the acquired search information. The output unit 165 outputs reaction information when the determination unit 161 determines that the timing has come when the user performed a search related to the second information.

[5-5. Timing judgment after content acquisition〕
In FIG. 1, an example has been described in which the requesting unit 162 transmits a content distribution request to the content distribution device 20 after the determining unit 161 determines whether it is the timing to output reaction information to the user. It is not limited to this. Specifically, the output unit 165 outputs reaction information every predetermined time (for example, 60 minutes). Further, the requesting unit 162 transmits a content distribution request to the content distribution device 20 at predetermined intervals (for example, 30 minutes). For example, the requesting unit 162 transmits a content distribution request to the content distribution device 20 at a certain time t1. Upon receiving the distribution request, the content distribution device 20 distributes the first content to the output device 100. Subsequently, the requesting unit 162 obtains the first content from the content distribution device 20. Upon acquiring the first content, the requesting unit 162 stores the acquired first content in the content storage unit 122. Next, the generation unit 163 refers to the content storage unit 122 and obtains the first content. Upon acquiring the first content, the generation unit 163 generates first reaction information by inputting the acquired first content into a predetermined model. After generating the first reaction information, the generation unit 163 stores the generated first reaction information in the reaction information storage unit 123.

Subsequently, the requesting unit 162 transmits a content distribution request to the content distribution device 20 at time t2, which is 30 minutes after time t1. Upon receiving the distribution request, the content distribution device 20 distributes the second content to the output device 100. Subsequently, the requesting unit 162 obtains the second content from the content distribution device 20. Upon acquiring the second content, the requesting unit 162 stores the acquired second content in the content storage unit 122. Next, the generation unit 163 generates second reaction information by inputting the second content into a predetermined model. After generating the second reaction information, the generation unit 163 stores the generated second reaction information in the reaction information storage unit 123. Subsequently, after generating the second reaction information, the generation unit 163 refers to the content storage unit 122 and acquires the first content whose distribution was requested at time t1 and the second content whose distribution was requested at time t2. . Subsequently, upon acquiring the first content and the second content, the generation unit 163 determines the degree of similarity between the first content and the second content. When the degree of similarity between the first content and the second content exceeds a predetermined threshold, the generation unit 163 assumes that the first content has become old and has been updated to a new second content, and extracts the second content from the content storage unit 122. 1 Delete content. Subsequently, the determination unit 161 determines whether it is the timing to output reaction information to the user. The output unit 165 outputs the second reaction information when the determination unit 161 determines that it is the timing to output the reaction information to the user.

[5-6. Reaction information to reviews〕
The generation unit 163 generates reaction information, which is character information that the dialogue system can say from the standpoint of the review writer, as reaction information to the content, which is a movie review. For example, the generation unit 163 generates characters such as "XX (movie title) was scary! I don't think I'll be able to sleep tonight." written from the perspective of the review writer, as reaction information for content that is a movie review of a horror movie. Generate reaction information that is information.

[5-7. Update model〕
The generation unit 163 generates reaction information using the updated model based on feedback from the user. Specifically, the generation unit 163 refers to the “presence or absence of a response by the user” item in the reaction information storage unit 123, and generates reaction information using a model updated according to the presence or absence of a response by the user. . For example, the generation unit 163 performs learning by making the learning weight for set data of reaction information and content to which a user has responded larger than the learning weight for set data of reaction information and content to which no user has responded. Generate reaction information using the model.

[5-8. Summary generation model and reaction generation model]
In the example shown in FIG. 1, an example has been described in which the generation unit 163 inputs content #1 into one model M1 and generates reaction information #1 including summary information R11 and trigger information R12. and the trigger information R12 may be generated using different models. Specifically, upon input of the second information, the generation unit 163 generates summary information using the model M11 that outputs summary information that is a summary of the second information. Moreover, when the generation unit 163 receives the second information, it generates reaction information using the model M12 that outputs reaction information indicating a reaction to the second information. In the example shown in FIG. 1, when content #1 is input, the generation unit 163 generates summary information R11 using model M11 that outputs summary information R11 that is a summary of content #1. Further, when content #1 is input, the generation unit 163 generates trigger information R12 using a model M12 that outputs trigger information R12 indicating a reaction to content #1.

In this case, the output processing system 1 may include a plurality of output devices 100, including an output device 100 for a summary role and an output device 100 for a reaction role. For example, the summary output device 100 outputs summary information R11, which is a summary of content #1, as voice. Subsequently, the output device 100 for the reaction combination outputs the trigger information R12 indicating the reaction to the content #1 using a different voice from the output device 100 for the summary combination. The output processing system 1 includes, in addition to the output device 100 for the summary role and the output device 100 for the reaction role, the output device 100 for the role that advances the dialogue, and the output device 100 for the role that assists the progress of the dialogue. A plurality of output devices 100 may be included.

[5-9. Output to screen]
The output unit 165 may output reaction information to the user by displaying text information on the screen instead of audio. Furthermore, the output unit 165 may output image information instead of audio. In this case, the output unit 165 displays an image that allows the user to understand the reaction information at a glance. The image information may be either a moving image or a still image. Furthermore, when the output device 100 is a humanoid robot, the output unit 165 may output the reaction information using the robot's gestures or hand gestures.

[6. effect〕
As described above, the output device 100 according to the embodiment includes the acquisition section 164 and the output section 165. The acquisition unit 164 acquires reaction information that is character information generated as a reaction to second information other than the first information input by the user into the dialogue system. The output unit 165 outputs the reaction information acquired by the acquisition unit 164 to the user.

Thereby, the output device 100 can spontaneously output character information that becomes an opportunity for dialogue related to news articles, etc., without any action from the user on the dialogue system. That is, the output device 100 allows the device to spontaneously and appropriately talk to a person. Therefore, the output device 100 can output appropriate information without depending on the user's input to the dialogue system.

Furthermore, the acquisition unit 164 acquires reaction information for second information other than the first information input by the user into the dialogue system, which is a non-task-oriented dialogue system.

Thereby, the output device 100 can spontaneously output character information that becomes a trigger for dialogue related to news articles, etc., even without any action from the user on the non-task-oriented dialogue system.

Furthermore, the acquisition unit 164 acquires reaction information for the second information, which is predetermined content.

Thereby, the output device 100 can spontaneously output character information that becomes an opportunity for dialogue related to predetermined content.

The acquisition unit 164 also acquires reaction information to the second information, which is a news article.

Thereby, the output device 100 can spontaneously output character information that becomes an opportunity for dialogue related to a news article.

Further, the acquisition unit 164 acquires reaction information to the second information, which is posted information on a predetermined SNS (Social Networking Service).

Thereby, the output device 100 can spontaneously output character information that becomes an opportunity for dialogue related to posted information on the SNS.

Furthermore, the acquisition unit 164 acquires reaction information regarding the second information, which is an advertisement.

Thereby, the output device 100 can spontaneously output character information that becomes an opportunity for dialogue related to the advertisement.

The acquisition unit 164 also acquires reaction information that becomes a trigger for dialogue with the user regarding the second information.

Thereby, the output device 100 can spontaneously output character information that becomes an opportunity for dialogue related to news articles, etc., without any action from the user on the dialogue system.

The acquisition unit 164 also acquires reaction information indicating impressions regarding the second information of the dialogue system.

Thereby, the output device 100 can spontaneously start an appropriate dialogue using impressions regarding the second information as an opportunity for dialogue, even without any action from the user on the dialogue system.

The acquisition unit 164 also acquires reaction information that requests the user's opinion regarding the second information.

Thereby, the output device 100 can spontaneously start an appropriate dialogue using the request for the user's opinion on the second information as an opportunity for dialogue, even without any action from the user on the dialogue system.

The acquisition unit 164 also acquires reaction information that prompts the user to take action related to the second information. For example, the acquisition unit 164 acquires reaction information that prompts the user to take an action other than responding to the reaction information.

Thereby, the output device 100 can spontaneously start an appropriate dialogue using the prompting for action related to the second information as an opportunity for dialogue, even without any action from the user on the dialogue system.

Further, the output unit 165 outputs the reaction information at a timing other than when the user requests the second information to the dialogue system.

Thereby, the output device 100 can spontaneously start appropriate dialogue without any action from the user on the dialogue system.

Furthermore, the output unit 165 outputs reaction information at a timing related to push notifications in the dialogue system.

Thereby, the output device 100 can spontaneously start an appropriate dialogue at the timing related to the push notification in the dialogue system.

Furthermore, the output unit 165 outputs reaction information at a timing based on the user's context.

Thereby, the output device 100 can spontaneously start an appropriate dialogue depending on the user's context.

Furthermore, the output unit 165 outputs reaction information at a timing when a predetermined period of time has elapsed since the user's utterance to the dialogue system.

Thereby, the output device 100 can spontaneously start appropriate dialogue at the timing when the dialogue with the user is interrupted.

Furthermore, the output unit 165 outputs reaction information at the timing when the user performs an action other than the action regarding the dialogue system and which is related to the second information. For example, the output unit 165 outputs reaction information at the timing when the user performs a search related to the second information.

Thereby, the output device 100 can spontaneously start appropriate dialogue at the timing when the user performs a search.

Further, the output device 100 according to the embodiment further includes a generation unit 163. The generation unit 163 generates reaction information from the second information. The acquisition unit 164 acquires the reaction information generated by the generation unit 163. Specifically, the generation unit 163 generates reaction information by inputting the second information into a predetermined model.

Thereby, the output device 100 can spontaneously and appropriately output reaction information other than reaction information registered in advance.

The acquisition unit 164 also acquires reaction information for the second information selected according to the user.

Thereby, the output device 100 can spontaneously and appropriately start a dialogue with content depending on the user.

Furthermore, the output unit 165 outputs the reaction information as voice. Furthermore, the output unit 165 may output image information instead of audio. In this case, the output unit 165 displays an image that allows the user to understand the reaction information at a glance. The image information may be either a moving image or a still image.

Thereby, the output device 100 can spontaneously and appropriately output timely reaction information.

The acquisition unit 164 also acquires summary information that is a summary of the second information. Furthermore, the output unit 165 outputs summary information and reaction information. Further, the output unit 165 outputs reaction information after outputting the summary information.

This allows the output device 100 to share common topics with users by outputting summary information. Further, by outputting reaction information, the output device 100 can create an opportunity to elicit a user's response to the summary information.

[7. Hardware configuration]
Further, the output device 100 and the content distribution device 20 according to the embodiments described above are realized by, for example, a computer 1000 having a configuration as shown in FIG. 15. FIG. 15 is a hardware configuration diagram showing an example of a computer that implements the functions of the output device 100 and the content distribution device 20. The computer 1000 includes a CPU 1100, a RAM 1200, a ROM 1300, an HDD 1400, a communication interface (I/F) 1500, an input/output interface (I/F) 1600, and a media interface (I/F) 1700.

CPU 1100 operates based on a program stored in ROM 1300 or HDD 1400, and controls each section. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started, programs depending on the hardware of the computer 1000, and the like.

The HDD 1400 stores programs executed by the CPU 1100, data used by the programs, and the like. Communication interface 1500 receives data from other devices via a predetermined communication network and sends it to CPU 1100, and transmits data generated by CPU 1100 to other devices via a predetermined communication network.

The CPU 1100 controls output devices such as a display and a printer, and input devices such as a keyboard and mouse via an input/output interface 1600. CPU 1100 obtains data from an input device via input/output interface 1600. Further, CPU 1100 outputs the generated data to an output device via input/output interface 1600.

Media interface 1700 reads programs or data stored in recording medium 1800 and provides them to CPU 1100 via RAM 1200. CPU 1100 loads this program from recording medium 1800 onto RAM 1200 via media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or a PD (Phase change rewritable disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. etc.

For example, when the computer 1000 functions as the output device 100 and the content distribution device 20, the CPU 1100 of the computer 1000 realizes the functions of the control unit 160 or the control unit 23 by executing a program loaded onto the RAM 1200. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them, but as another example, these programs may be acquired from another device via a predetermined communication network.

Some of the embodiments of the present application have been described above in detail based on the drawings, but these are merely examples, and various modifications and variations may be made based on the knowledge of those skilled in the art, including the embodiments described in the disclosure section of the invention. It is possible to carry out the invention in other forms with modifications.

[8. others〕
Furthermore, among the processes described in the above embodiments and modified examples, all or part of the processes described as being performed automatically can be performed manually, or may be described as being performed manually. All or part of this processing can also be performed automatically using known methods. In addition, information including the processing procedures, specific names, and various data and parameters shown in the above documents and drawings may be changed arbitrarily, unless otherwise specified. For example, the various information shown in each figure is not limited to the illustrated information.

Furthermore, each component of each device shown in the drawings is functionally conceptual, and does not necessarily need to be physically configured as shown in the drawings. In other words, the specific form of distributing and integrating each device is not limited to what is shown in the diagram, and all or part of the devices can be functionally or physically distributed or integrated in arbitrary units depending on various loads and usage conditions. Can be integrated and configured.

Furthermore, the above-described embodiments and modifications can be combined as appropriate within a range that does not conflict with the processing contents.

Further, the above-mentioned "section, module, unit" can be read as "means", "circuit", etc. For example, the output section can be read as an output means or an output circuit.

1 Output processing system 100 Output device 110 Communication unit 120 Storage unit 121 Sensor information storage unit 122 Content storage unit 123 Reaction information storage unit 124 Model information storage unit 125 Learning data storage unit 130 Input unit 140 Display unit 150 Physical sensor 160 Control unit 161 Judgment unit 162 Request unit 163 Generation unit 164 Acquisition unit 165 Output unit

Claims (21)

Reaction information corresponding to a predetermined content including text information provided to the user by a service different from the dialogue system based on the user's location information acquired by a physical sensor , the predetermined content If it is determined that it is the timing to output the reaction information to the user, the predetermined content is acquired and the text information is output. When input information including the input information is input, summary information is a summary of the character information included in the input information, and trigger information is a sentence that prompts the reference user who referred to the input information to make a comment regarding the input information. The reaction information corresponding to the predetermined content is extracted from the predetermined content using a machine learning model trained to output the reaction information including trigger information different from the summary information as output information. a generation procedure that generates
an output procedure of outputting the reaction information generated by the generation procedure to the user when it is determined that it is time to output the reaction information to the user based on the user's location information; ,
An output program that causes a computer to execute. The generation procedure is
When it is determined that it is the timing to output the reaction information corresponding to the predetermined content provided to the user by a service different from the dialog system that is a non-task-oriented dialog system, the predetermined The output program according to claim 1, wherein the output program acquires content and generates the reaction information . The generation procedure is
The output program according to claim 1 or 2, wherein a news article is acquired as the predetermined content and the reaction information is generated. The generation procedure is
The output program according to claim 1 or 2, wherein the output program acquires posted information in a predetermined SNS (Social Networking Service) as the predetermined content and generates the reaction information . The generation procedure is
The output program according to claim 1 or 2, wherein content that is an advertisement is acquired as the predetermined content and the reaction information is generated. The generation procedure is
The output program according to any one of claims 1 to 5, wherein the output program generates the reaction information that becomes an opportunity for dialogue with the user regarding the predetermined content . The generation procedure is
7. The output program according to claim 1, wherein the output program generates the reaction information that requests the user's opinion regarding the predetermined content . The generation procedure is
The output program according to claim 1, wherein the output program generates the reaction information that prompts the user to take an action related to the predetermined content . The generation procedure is
The output program according to claim 8, wherein the output program generates the reaction information that prompts the user to take an action other than a response to the reaction information. The output procedure is
10. The output program according to claim 1, wherein the reaction information is output at a timing other than when the user requests the predetermined content to the dialog system. The output procedure is
11. The output program according to claim 1, wherein the reaction information is output at a timing related to a push notification in the dialog system. The output procedure is
12. The output program according to claim 1, wherein the reaction information is output at a timing based on the user's context. The output procedure is
13. The output program according to claim 12, wherein the reaction information is output at a timing when a predetermined period of time has elapsed since the user's utterance to the dialogue system. The output procedure is
Claim 12 or claim 12, wherein the reaction information is output at a timing when an action of the user other than an action with respect to the dialogue system and related to the predetermined content is performed. The output program described in 13. The output procedure is
15. The output program according to claim 14, wherein the reaction information is output at a timing when the user performs a search related to the predetermined content . The generation procedure is
16. The output program according to claim 1, wherein the reaction information is generated by inputting the predetermined content into a predetermined model. The generation procedure is
17. The output program according to claim 1, wherein the reaction information is generated based on the predetermined content selected according to the user. The output procedure is
The output program according to any one of claims 1 to 17, characterized in that the reaction information is output as voice. The output procedure is
The output program according to claim 18 , wherein the trigger information is output after outputting the summary information . Reaction information corresponding to a predetermined content including text information provided to the user by a service different from the dialogue system based on the user's location information acquired by a physical sensor , the predetermined content If it is determined that it is the timing to output the reaction information to the user, the predetermined content is acquired and the text information is output. When input information including the input information is input, summary information is a summary of the character information included in the input information, and trigger information is a sentence that prompts the reference user who referred to the input information to make a comment regarding the input information. The reaction information corresponding to the predetermined content is extracted from the predetermined content using a machine learning model trained to output the reaction information including trigger information different from the summary information as output information. a generation unit that generates
an output unit that outputs the reaction information generated by the generation unit to the user when it is determined that it is time to output the reaction information to the user based on the user's location information; ,
An output device comprising: An output method executed by a computer,
Reaction information corresponding to a predetermined content including text information provided to the user by a service different from the dialogue system based on the user's location information acquired by a physical sensor , the predetermined content If it is determined that it is the timing to output the reaction information to the user, the predetermined content is acquired and the text information is output. When input information including the input information is input, summary information is a summary of the character information included in the input information, and trigger information is a sentence that prompts the reference user who referred to the input information to make a comment regarding the input information. The reaction information corresponding to the predetermined content is extracted from the predetermined content using a machine learning model trained to output the reaction information including trigger information different from the summary information as output information. a generation step of generating;
an output step of outputting the reaction information generated in the generation step to the user when it is determined that it is the timing to output the reaction information to the user based on the user's location information; ,
An output method characterized by including.

Images