Retrieval support device, and program

Abstract

To enable desired time-series contents suitable for a user or its specific portion to be easily retrieved.SOLUTION: A retrieval support device makes a terminal device execute retrieval support processing for receiving a second utterance for extracting a subset of utterance guidance information from a set of utterance guidance information for guiding a first utterance for selecting time-series contents or a specific portion of the time-series contents, and displaying the extracted subset of utterance guidance information on the basis of user specific information representing characteristics of a user performing the second utterance and the second utterance, and makes the terminal device execute retrieval result output processing for outputting the time-series contents selected by the first utterance or the specific portion of the time-series contents.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a technology for supporting searches, and more particularly to a technology for supporting searches of time-series content. [Background technology]

[0002] BACKGROUND ART There is known a method of accepting an input of a search word, searching for video content corresponding to the input search word, and providing the found video content (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2014-74942 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the past, unless a user inputs appropriate search terms, it was not possible to find video content suitable for the user. In particular, when a user tries to find a specific scene in video content, it is difficult for the user to select appropriate search terms, making it difficult to find the specific scene in video content suitable for the user. This problem is common not only when searching for video content or a specific scene within it, but also when searching for any time-series content or a specific part thereof.

[0005] The present invention provides a technique that enables a user to easily search for desired time-series content or a specific part thereof that is suitable for the user. [Means for solving the problem]

[0006] The search support device receives a second utterance for extracting a subset of speech guidance information from a set of speech guidance information that induces a first utterance for selecting time-series content or a specific part of the time-series content, causes a terminal device to execute a search support process for displaying the subset of speech guidance information extracted based on user characteristic information that represents the characteristics of the user making the second utterance and the second utterance, and causes the terminal device to execute a search result output process for outputting the time-series content or the specific part of the time-series content selected by the first utterance. [Effects of the Invention]

[0007] This allows the user to easily search for desired time-series content or a specific part thereof that is suitable for the user. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram illustrating the functional configuration of a search system according to an embodiment. [Figure 2] FIG. 2 is a block diagram illustrating a functional configuration of the search support device according to the embodiment. [Figure 3] FIG. 3 is a block diagram illustrating a functional configuration of the terminal device according to the embodiment. [Figure 4] Fig. 4A is a diagram illustrating an example of time-series content according to an embodiment, and Fig. 4B is a diagram illustrating a data structure of time-series content information according to an embodiment. [Figure 5] FIG. 5 is a diagram illustrating an example of a data configuration of user characteristic information according to the embodiment. [Figure 6] FIG. 6 is a diagram illustrating an example of the data structure of the search support information according to the embodiment. [Figure 7] FIG. 7 is a flow diagram illustrating the search support process according to the embodiment. [Figure 8] 8A and 8B are conceptual diagrams illustrating a subset of utterance elicitation information according to an embodiment. [Figure 9]9A and 9B are diagrams illustrating how a subset of utterance elicitation information is displayed according to the embodiment. [Figure 10] 10A and 10B are diagrams illustrating how a subset of utterance elicitation information is displayed according to the embodiment. [Figure 11] 11A and 11B are diagrams illustrating how time-series content is output according to the embodiment. [Figure 12] FIG. 12 is a diagram illustrating a method for extracting a subset of utterance elicitation information according to the embodiment. [Figure 13] 13A and 13B are diagrams illustrating how a subset of utterance elicitation information is displayed according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. [First embodiment] In the first embodiment, the terminal device is caused to execute a search support process to receive a second utterance for extracting a subset of utterance guide information from a set of utterance guide information for inducing a first utterance for selecting time-series content or a specific part of the time-series content, and to display the subset of utterance guide information extracted based on the second utterance and user characteristic information representing characteristics of the user making the second utterance.Furthermore, the terminal device is caused to execute a search result output process to output the time-series content or the specific part of the time-series content selected by the first utterance.

[0010] <Overall structure> 1, a search system 1 of this embodiment includes a search support device 11 and N terminal devices 12-1, ..., 12-N. Here, N is an integer equal to or greater than 1, n=1, ..., N, and each of the terminal devices 12-1, ..., 12-N is referred to as a terminal device 12-n, and the terminal devices 12-1, ..., 12-N are collectively referred to as a terminal device 12. The search support device 11 and the N terminal devices 12-1, ..., 12-N are configured to be able to communicate with each other via a network such as the Internet.

[0011] <Search support device> As illustrated in FIG. 2 , the search support device 11 of this embodiment includes a control unit 110, a memory 111, storage units 112-114, a communication unit 115, a search support processing unit 116, and a content providing unit 117. Although not described below, the search support device 11 executes each process under the control of the control unit 110. Data input to the communication unit 115 and data obtained by the processing of the search support processing unit 116 and the content providing unit 117 are stored in the memory 111 one by one and read and used as needed. The search support device 11 is configured by a general-purpose or dedicated computer that includes, for example, a processor (hardware processor) such as a central processing unit (CPU), memories such as random-access memory (RAM) and read-only memory (ROM), storage devices such as a hard disk and a solid-state drive (SSD), and a communication device, executing a predetermined program. This computer may include one processor and memory, or multiple processors and memories. This program may be installed on the computer or may be pre-recorded on a ROM or the like. Furthermore, instead of electronic circuits that implement functional configurations by loading programs like a CPU, some or all of the processing units may be configured using electronic circuits that implement processing functions without using programs. Also, an electronic circuit that constitutes one device may include multiple CPUs. Specific examples of hardware that constitutes the search support device 11 include a server device or a collection of server devices (e.g., a cloud).

[0012] <Terminal Device> 3, each terminal device 12-n includes a control unit 120-n, a memory 121-n, an input receiving unit 122-n, an utterance receiving unit 123-n, an utterance guide information display unit 124-n, a search result output unit 125-n, a search processing unit 126-n, and a communication unit 127-n. Although not described below, each terminal device 12-n executes each process under the control of the control unit 120-n. Data input to the input receiving unit 122-n, the utterance receiving unit 123-n, and the communication unit 127-n, as well as data obtained by the processes of the utterance receiving unit 123-n, the utterance guide information display unit 124-n, the search result output unit 125-n, and the search processing unit 126-n, are stored in the memory 121-n one by one and are read out and used as needed. For example, the search support device 11 is configured by a general-purpose or dedicated computer equipped with a processor, a storage device, a memory, a communication device, and user interfaces such as a mouse, a keyboard, a microphone, a display or touch screen, and a speaker, executing a predetermined program. The computer may have one processor and memory, or multiple processors and memories. The program may be installed on the computer or pre-recorded on a ROM or the like. Furthermore, some or all of the processing units may be configured using electronic circuits that realize processing functions without using a program, rather than electronic circuits that realize functional configuration by loading a program like a CPU. Furthermore, the electronic circuits that make up a single device may include multiple CPUs. Specific examples of hardware that make up the search support device 11 include a personal computer, a tablet terminal device, a smartphone, a smartwatch, or any combination thereof. Each terminal device 12-n is used by a corresponding user 1000-n. Users 1000-1, ..., 1000-N are collectively referred to as user 1000.

[0013] <Pre-processing> As a preprocessing step, time-series content information is stored in the storage unit 112 of the search support device 11 (FIG. 2). The time-series content information includes content data representing time-series content or a specific portion of the time-series content. The time-series content may be any content that includes time-series information. For example, the time-series content may be video content, audio content such as audio content or music content, or a combination of these. Examples of time-series content include movies, program videos, streaming videos, game videos, VR videos, simulation videos, audiobooks, music, and audio corpora. A specific portion of the time-series content may be, for example, a specific time-series range of the time-series content. For example, a specific portion of the time-series content may be a specific scene of video content or a specific time interval of audio content. FIG. 4A illustrates an example of time-series content. The time-series content illustrated in FIG. 4A is a specific portion (specific scene) between a specific start time and end time of video content, and includes content related to "delicious ramen in Yokohama." FIG. 4B illustrates an example of time-series content information. 4B includes link information of content data representing time-series content or a specific part of time-series content, the content data, and the start time and end time of the content data, which are associated with each other.

[0014] In addition, user characteristic information is stored in the memory unit 113 of the search support device 11 (Figure 2). User characteristic information is information that represents the characteristics of each user 1000-n. The characteristics of user 1000 may be anything that characterizes user 1000. Specific examples of the characteristics of user 1000 include the attributes and search history of user 1000. The attributes of user 1000 include, for example, gender, age, place of residence, marital status, occupation, hobbies, tastes, psychological tendencies, etc. Figure 5 shows an example of user characteristic information. The user characteristic information illustrated in Figure 5 includes a user ID that identifies user 1000, attribute information, and search history, which are associated with each user ID.

[0015] Furthermore, search support information is stored in the storage unit 114 of the search support device 11 (FIG. 2). The search support information includes a collection of utterance guide information that guides the user 1000 to make an utterance (first utterance) to select time-series content or a specific portion of the time-series content. The utterance guide information may be any information that guides such an utterance (first utterance), and may be text information, image information, or audio information. The utterance guide information is information that corresponds to the time-series content or a specific portion of the time-series content. For example, the utterance guide information includes a word (keyword), a combination of words, a sentence, a title, a summary, a figure, a symbol, a sound, music, or at least any combination thereof, that corresponds to the time-series content or a specific portion of the time-series content. For example, the utterance guidance information may be information representing the time-series content or a specific portion of the time-series content, information identifying the time-series content or a specific portion of the time-series content, information explaining the time-series content or a specific portion of the time-series content, information summarizing the time-series content or a specific portion of the time-series content, or other information related to the time-series content or a specific portion of the time-series content. A specific example of search support information is shown in FIG. 6. The search support information shown in FIG. 6 includes a set of utterance guidance information corresponding to the time-series content or a specific portion of the time-series content, a description of the time-series content or the specific portion of the time-series content, attribute information of the time-series content or the specific portion of the time-series content, and link information of content data representing the time-series content or the specific portion of the time-series content. The speech-guiding information illustrated in FIG. 6 is a word or a combination of words corresponding to the time-series content or a specific portion of the time-series content, the explanation is a word or a combination of words that explains the time-series content or the specific portion of the time-series content, the attribute information is a collection of words or combinations of words that represent the attributes of the time-series content or the specific portion of the time-series content, and the link information is information that represents a link to the content data illustrated in FIG. 4B that represents the time-series content or the specific portion of the time-series content.In the example of Fig. 6, multiple pieces of utterance leading information are set for one piece of time-series content or a specific part of the time-series content. However, this does not limit the present invention, and one piece of utterance leading information may be set for one piece of time-series content or a specific part of the time-series content.

[0016] <Search process> A user 1000-n performing a search performs an operation input to the input receiving unit 122-n of the terminal device 12-n (FIG. 3) performing the search to instruct the start of a search, or performs a speech input to the speech receiving unit 123-n to instruct the start of a search. The operation input or speech input instructing the start of a search is sent to the search processing unit 126-n. In response, the search processing unit 126-n generates a search start request including the user ID of the user 1000-n and sends it to the communication unit 127-n. In this embodiment, the search start request includes the user ID of the user 1000-n. The communication unit 127-n transmits the search start request to the search support device 11 via the network. The communication unit 115 of the search support device 11 (FIG. 2) receives the search start request and sends it to the search support processing unit 116. As a result, the search support device 11 starts the following process.

[0017] As illustrated in FIG. 7, the search support processing unit 116 of the search support device 11 (FIG. 2) extracts a subset of utterance leading information (e.g., "eat and drink," "noodles," "ramen," "movies") from a set of utterance leading information (e.g., "eat and drink," "noodles," "ramen," "Yokohama ramen," "movies," "foreign movies") included in the search support information (FIG. 6) stored in the storage unit 114 (FIG. 8A). In other words, the search support processing unit 116 narrows down the set of utterance leading information to the subset of utterance leading information. For example, the search support processing unit 116 randomly extracts the subset of utterance leading information from the set of utterance leading information. Alternatively, for example, the search support processing unit 116 may extract the subset of utterance leading information from the set of utterance leading information based on the user characteristic information (FIG. 5) of the user 1000-n. In the latter case, the search support processing unit 116 may, for example, search the user identification information stored in the storage unit 113 using the user ID included in the search start request, extract user characteristic information, which is attribute information and / or search history corresponding to the user ID, and extract a subset of utterance leading information from the set of utterance leading information based on the user characteristic information. The extraction of the subset of utterance leading information based on the user characteristic information may be performed based on, for example, the likelihood (likelihood) of the subset of utterance leading information selected by a user with attributes represented by the attribute information of the user characteristic information, or based on the similarity between the search history of the user characteristic information and the extracted subset of utterance leading information, or based on both of these. For example, the subset of utterance leading information with the highest likelihood of selection by a user with attributes represented by the attribute information may be extracted, or the subset of utterance leading information with the highest similarity to the search history may be extracted, or the subset of utterance leading information with the highest weighted sum of these may be extracted. The likelihood of a subset of utterance prompting information selected by a user of an attribute represented by the attribute information of the user characteristic information may be determined based on a machine-learned model, a weighted sum of scores of utterance prompting information determined for each attribute, or other methods. The similarity between the search history and the subset of utterance prompting information may be determined based on a machine-learned model, or using a known similarity measure or distance measure.Such a function can be implemented, for example, by a function or algorithm that outputs a subset of utterance leading information in response to input of user characteristic information. The search support processing unit 116 sends the extracted subset of utterance leading information to the communication unit 115, and the communication unit 115 transmits the subset of utterance leading information to the terminal device 12-n via the network. The communication unit 127-n of the terminal device 12-n (FIG. 3) receives the subset of utterance leading information and transmits it to the search processing unit 126-n. The search processing unit 126-n sends the subset of utterance leading information to the utterance leading information display unit 124-n, and the utterance leading information display unit 124-n displays the subset of utterance leading information. For example, as illustrated in FIG. 9A, the utterance leading information display unit 124-n displays an utterance leading information 1231a-n included in the subset of utterance leading information moving across the display screen 1231-n in a specific direction (e.g., rightward). For example, the utterance leading information 1231a-n appears (frames in) from one side (e.g., the left side) of the display screen 1231-n, moves in a specific direction (e.g., to the right) within the display screen 1231-n, and then disappears (frames out) to the other side (e.g., the right side) of the display screen 1231-n. The display order and movement speed of the utterance leading information 1231a-n may or may not be constant. In the latter case, the display order and movement speed of the utterance leading information 1231a-n may be determined randomly. Note that this display method is merely an example and does not limit the present invention. For example, a subset of utterance guidance information may be displayed together on the display screen 1231-n, or the subset of utterance guidance information may be displayed aligned on the display screen 1231-n, or the subset of utterance guidance information may be displayed at random positions on the display screen 1231-n, or the subset of utterance guidance information may move in random directions within the display screen 1231-n. The user 1000-n can view the subset of utterance guidance information displayed in this manner. That is, the search support processing unit 116 of the search support device 11 (FIG. 2) causes the terminal device 12-n (FIG. 3) to execute a process (initial search support process) of displaying a subset of utterance guidance information selected from the set of utterance guidance information (for example, a randomly selected subset of utterance guidance information) (step S116a).

[0018] The user 1000-n can view the subset of speech guide information displayed on the speech guide information display unit 124-n and make an utterance based on it. That is, the user 1000-n can make an utterance related to the displayed subset of speech guide information. The utterances are roughly divided into an utterance S1 (first utterance) for selecting the time-series content or a specific part of the time-series content, an utterance S2 (second utterance) for extracting a subset of speech guide information from the set of speech guide information that guides the utterance S1, and other utterances. Specific examples of the utterance S1 and the utterance S2 are shown below.

[0019] <Example of utterance S1 (first utterance)> (Example 1-1) An utterance for extracting a subset of speech leading information in which the number of elements is equal to or less than the threshold T2. For example, an utterance that specifies a specific field or specific speech leading information from the displayed subset of speech leading information, where the number of speech leading information in the specified field or the number of speech leading information related to the specified speech leading information is equal to or less than the threshold T2, or an utterance that excludes a specific field or specific speech leading information from the displayed subset of speech leading information, where the number of remaining speech leading information is equal to or less than the threshold T2. T2 is an integer equal to or greater than 1. T2 may be a constant or a variable determined based on an operation input or other processing. Here, the extraction of the subset of speech leading information in this embodiment is based on the utterance of the user 1000-n and the user characteristic information (FIG. 5). For example, the extraction of the subset of speech leading information based on the utterance and the user characteristic information may be based on the likelihood of the subset of speech leading information selected when a user with attributes represented by the attribute information of the user characteristic information makes the utterance, or on the similarity between the search history of the user characteristic information and the utterance and the extracted subset of speech leading information, or on both of these. For example, the subset of speech leading information that is most likely to be selected when a user with attributes represented by the attribute information makes the utterance may be extracted, or the subset of speech leading information that is most similar to the search history and the utterance may be extracted, or the subset of speech leading information that is most similar to the search history and the utterance may be extracted, or the subset of speech leading information that is most weighted sum of these may be extracted. The likelihood of a subset of utterance prompting information selected by a user of an attribute represented by the attribute information of the user characteristic information may be determined based on a machine-learned model, a weighted sum of scores of utterance prompting information determined for each attribute, or other methods. The similarity between the search history and the utterance and the subset of utterance prompting information may be determined based on a machine-learned model, or using a known similarity measure or distance measure.Such a function may be implemented, for example, by a function or algorithm that outputs a subset of speech leading information to be extracted in response to an input of an utterance and user characteristic information; or by a function or algorithm that outputs a subset of speech leading information to be extracted in response to an input of a field or speech leading information specified by an utterance and user characteristic information; or by a function or algorithm that outputs a subset of speech leading information to be extracted in response to an input of a probability that each field or each piece of speech leading information is specified by an utterance and user characteristic information; or by a function or algorithm that outputs a subset of speech leading information to be extracted in response to an input of a field or speech leading information excluded by an utterance and user characteristic information; or by a function or algorithm that outputs a subset of speech leading information to be extracted in response to an input of a probability that each field or each piece of speech leading information is excluded by an utterance and user characteristic information. The user 1000-n may designate or exclude a specific field or specific speech prompt by reading out the field or speech prompt, or by uttering a demonstrative term such as "this." In the latter case, a criterion may be established as to which of the displayed speech prompts the uttered demonstrative term refers to, and the designated or excluded speech prompt may be determined according to the criterion. For example, when a demonstrative term such as "this" is uttered, it may be determined that all of the speech prompts 1231a-n displayed on the display screen 1231-n (FIG. 9A) have been designated or excluded. Alternatively, when the demonstrative term is uttered, it may be determined that the speech prompt 1231a-n displayed in a specific display position (e.g., the central region) on the display screen 1231-n has been designated or excluded. Alternatively, it may be determined that the speech prompt 1231a-n has been designated or excluded with a probability that depends on the display position of the speech prompt 1231a-n when the demonstrative term is uttered. For example, it may be determined that the closer the utterance guide information 1231a-n is displayed in a position closer to a specific area (e.g., the central area) of the display screen 1231-n at the time when the demonstrative term is uttered, the higher the probability that it has been designated or excluded by the user 1000-n. In this case, a new set of utterance guide information is extracted based on the display position, utterance, and user characteristic information of the subset of the displayed utterance guide information. Furthermore, the field of the utterance guidance information specified or excluded by the utterance may be determined based on the attribute information of the search support information (FIG. 6). For example, if the attribute information of the search support information (FIG. 6) includes a word representing a field (e.g., "eating and drinking") and the uttered word matches or is similar to any word included in this attribute information (e.g., "ramen") (e.g., if the degree of match or similarity exceeds a threshold), the field specified or excluded by the utterance may be determined to be the field (e.g., "eating and drinking") represented by the word included in this attribute information. For example, utterance guidance information that includes the same word representing a field (e.g., "eating and drinking") in its attribute information may be determined to belong to the same field. In Example 1-1, in the process of step S116g described later, for example, the time-series content or a specific part of the time-series content corresponding to the utterance guidance information narrowed down to equal to or less than the threshold T2 is selected. The time-series content or the specific part of the time-series content corresponding to the utterance guidance information is identified based on, for example, the time-series content information (FIG. 4B) and the search support information (FIG. 6). For example, link information corresponding to the search support information (FIG. 6) can be identified, and content data representing the time-series content or the specific part of the time-series content corresponding to the utterance guidance information can be identified based on the identified link information and the time-series content information (FIG. 4B). (Example 1-2) An utterance whose degree of match with any of the displayed utterance leading information exceeds the threshold T1. For example, this is an utterance in which any of the displayed utterance leading information is read out as is. In this case, in the process of step S116g described below, for example, a time-series content or a specific part of the time-series content corresponding to the utterance leading information whose degree of match with the utterance exceeds the threshold T1 is selected. T1 is a positive real number. T1 may be a constant or a variable determined based on an operation input or other processing. (Example 1-3) An utterance whose similarity or relevance to any of the displayed utterance guides exceeds the threshold T1. For example, this is an utterance that reads out any of the displayed utterance guides, or its synonyms, thesaurus, synonyms, or equivalents. In this case, in the process of step S116g described below, for example, a time-series content or a specific part of the time-series content corresponding to the utterance guides whose similarity or relevance to the utterance exceeds the threshold T1 is selected. The similarity or relevance may be an index representing the similarity of characters or pronunciation, or an index representing the similarity of meaning or concept. (Example 1-4) An utterance whose probability of representing any of the displayed utterance leading information exceeds the threshold T1. For example, an utterance that reads out any of the displayed utterance leading information as is, or an utterance that reads out a short phrase to the utterance. In this case, in the process of step S116g described below, for example, a time-series content or a specific part of the time-series content corresponding to the utterance leading information whose probability of representing the utterance exceeds the threshold T1 is selected. (Example 1-5) An utterance that satisfies the conditions of any one or more of Examples 1-1 to 1-4.

[0020] <Example of utterance S2 (second utterance)> (Example 2-1) Utterances for extracting a subset of speech leading information in which the number of elements exceeds the threshold T2. For example, an utterance that specifies a specific field or speech leading information from the displayed subset of speech leading information, in which the number of speech leading information in the specified field or the number of speech leading information related to the specified speech leading information exceeds the threshold T2, or an utterance that excludes a specific field or speech leading information from the displayed subset of speech leading information, in which the number of remaining speech leading information exceeds the threshold T2. As explained in Example 1-1, the extraction of a subset of utterance guidance information in this embodiment is based on the utterance of user 1000-n and user characteristic information (Fig. 5). The specification or exclusion of a specific field or specific utterance guidance information may be performed by reading out the field or utterance guidance information, or by uttering a demonstrative such as "this." The field of the utterance guidance information specified or excluded by the utterance may be determined based on the attribute information of the search support information (Fig. 6). Specific examples of these are as shown in Example 1-1. In the case of Example 2-1, in the process of step S116e described later, for example, a new subset of utterance leading information is extracted, which has a smaller number of elements than the displayed subset of utterance leading information. The newly extracted subset of utterance leading information includes, for example, a part of the displayed subset of utterance leading information. For example, the displayed subset of utterance leading information and the newly extracted subset of utterance leading information partially overlap with each other.

[0021] (Example 2-2) An utterance whose degree of similarity with any of the displayed utterance leading information does not exceed the threshold T1. In this case, for example, in the process of step S116e described below, a new subset of utterance leading information different from the subset of the displayed utterance leading information is extracted. The newly extracted subset of utterance leading information includes, for example, utterance leading information not included in the subset of the displayed utterance leading information. As explained in Example 1-1, the extraction of the subset of utterance leading information in this embodiment is based on the utterance of user 1000-n and user characteristic information. A specific example is as shown in Example 1-1. The subset of the displayed utterance leading information and the newly extracted subset of utterance leading information may or may not overlap partially. These points also apply to Examples 2-3, 2-4, 2-5, and 2-6 below. Explanation of these points will be omitted in Examples 2-3, 2-4, 2-5, and 2-6. (Example 2-3) An utterance whose similarity to any of the displayed prompts does not exceed the threshold T1. (Example 2-4) An utterance in which the probability of representing any of the displayed utterance prompting information does not exceed the threshold T1. (Example 2-5) Utterances that negate a subset of the displayed prompts. For example, "I don't want to see this" or "This isn't it." (Example 2-6) An utterance instructing a change of the displayed subset of prompts. For example, an utterance such as "change it" or "show me something else."

[0022] (Example 2-7) An utterance that satisfies the conditions of any one or more of the utterances in Examples 2-1 to 2-6.

[0023] The utterance receiving unit 123-n of the terminal device 12-n (FIG. 3) receives an input of an utterance from the user 1000-n. The utterance receiving unit 123-n sends information representing the input utterance to the search processing unit 126-n. The information representing the utterance is, for example, text information obtained by speech recognition of the utterance. The search processing unit 126-n sends the information representing the utterance to the communication unit 127-n, and the communication unit 127-n transmits the information representing the utterance to the search support device 11 via the network. The communication unit 115 of the search support device 11 (FIG. 2) receives the information representing the utterance and sends it to the search support processing unit 116. The search support processing unit 116 receives the information representing the utterance. At this time, the search support processing unit 116 may update attribute information of the user characteristic information (FIG. 5) stored in the storage unit 113 based on the received utterance. For example, the search support processing unit 116 may add words identified in the received utterance to the attribute information (attribute information corresponding to the user ID of user 1000-n) of the user characteristic information (Figure 5) stored in the memory unit 113 (step S116b).

[0024] The search support processing unit 116 determines whether the received utterance is utterance S1 (first utterance). For example, the search support processing unit 116 determines whether the received utterance corresponds to the utterance exemplified in <Example of utterance S1 (first utterance)>. For example, the search support processing unit 116 may determine whether the received utterance satisfies Example 1-1, or may determine whether the received utterance satisfies both Example 1-1 and Example 1-2. Here, if it is determined that the received utterance is not utterance S1, the process proceeds to step S116d. On the other hand, if it is determined that the received utterance is utterance S1, the process proceeds to step S116g (step S116c).

[0025] In the process of step S116d, the search support processing unit 116 determines whether the received utterance is utterance S2 (second utterance). For example, the search support processing unit 116 determines whether the received utterance corresponds to the utterance exemplified in <Example of utterance S2 (second utterance)>. For example, the search support processing unit 116 may determine whether the received utterance satisfies Example 2-1, or may determine whether the received utterance satisfies both Example 2-1 and Example 2-2. Here, if it is determined that the received utterance is not utterance S2, the process returns to the process of step S116b. On the other hand, if it is determined that the received utterance is utterance S2, the process proceeds to the process of step S116e (step S116d).

[0026] In the process of step S116e, the search support processing unit 116 of the search support device 11 (FIG. 2) extracts a new subset of speech guidance information from the set of speech guidance information included in the search support information (FIG. 6) stored in the storage unit 114 based on the user characteristic information of the user 1000-n and the received utterance (second utterance) (FIG. 8B). For example, the search support processing unit 116 searches for user identification information stored in the storage unit 113 using the user ID included in the search start request, extracts user characteristic information, which is attribute information and / or search history corresponding to the user ID, and extracts a new subset of speech guidance information from the set of speech guidance information based on the user characteristic information and the received utterance. Specific examples of extracting a subset of speech guidance information in this embodiment are as shown in <Example of Utterance S2 (Second Utterance)> and <Example of Utterance S1 (First Utterance)>. The new subset of speech guidance information is sent to the communication unit 115 of the search support device 11 (FIG. 2). The communication unit 115 transmits the new subset of speech guide information to the terminal device 12-n via the network. The communication unit 127-n of the terminal device 12-n (FIG. 3) receives the new subset of speech guide information and transmits it to the search processing unit 126-n. The search processing unit 126-n transmits the new subset of speech guide information to the speech guide information display unit 124-n, and the speech guide information display unit 124-n displays the new subset of speech guide information. For example, as illustrated in FIGS. 9B, 10A, and 10B, the speech guide information display unit 124-n displays an image of speech guide information 1231a-n included in the subset of speech guide information moving across the display screen 1231-n in a specific direction (e.g., rightward). Here, FIG. 9B shows an example in which the subset of speech guide information displayed as in FIG. 9A is narrowed down to a subset of speech guide information related to eating and drinking. Fig. 10A shows an example in which the subset of utterance guides in the food and drink category displayed as in Fig. 9B is narrowed down to a subset of utterance guides in the ramen category. Fig. 10B shows an example in which the subset of utterance guides in the Yokohama ramen category displayed as in Fig. 10A is negated, and a subset of utterance guides in the Shibuya ramen category is displayed.As described above, the search support processing unit 116 of the search support device 11 (FIG. 2) of this embodiment receives an utterance S2 (second utterance) for extracting a subset of utterance leading information from a set of utterance leading information that guides an utterance S1 (first utterance) for selecting time-series content or a specific part of the time-series content, and causes the terminal device 12-n (FIG. 3) to execute a search support process that displays the subset of utterance leading information extracted based on the utterance S2 (second utterance) and user characteristic information that indicates the characteristics of the user 1000-n who made the utterance S2 (second utterance). Then, the process returns to step S116a (step S116e).

[0027] In the process of step S116g, the content providing unit 117 of the search support device 11 (FIG. 2) extracts the time-series content or a specific part of the time-series content selected by the utterance (first utterance) received by the search support processing unit 116 (see <Example of utterance S1 (first utterance)> above). Here, the extraction of the time-series content or the specific part of the time-series content corresponding to the utterance guide information is performed, for example, based on the search support information (FIG. 6) and the time-series content information (FIG. 4B). For example, the search support processing unit 116 searches the search support information (FIG. 6) stored in the storage unit 114 and extracts link information associated with the utterance guide information corresponding to the received utterance. Alternatively, for example, the search support processing unit 116 may extract link information associated with description or attribute information including the utterance guide information corresponding to the received utterance. Alternatively, for example, the search support processing unit 116 may extract link information associated with speech guidance information whose degree of match with the speech guidance information corresponding to the received utterance exceeds a threshold, or may extract link information associated with description or attribute information including an element whose degree of match with the speech guidance information corresponding to the received utterance exceeds a threshold. The search support processing unit 116, for example, sends the extracted link information to the content providing unit 117. The content providing unit 117, for example, searches the time-series content information ( FIG. 4B ) stored in the storage unit 112 and extracts content data corresponding to the sent link information. The extracted content data represents the time-series content selected by the utterance S1 (first utterance) or a specific part of the time-series content. The content providing unit 117 sends the extracted content data to the search support processing unit 116, and the search support processing unit 116 sends the received content data to the communication unit 115. The communication unit 115 transmits the content data to the terminal device 12-n via the network. The communication unit 127-n of the terminal device 12-n (FIG. 3) receives the content data and sends it to the search processing unit 126-n. The search processing unit 126-n sends the content data to the search result output unit 125-n, which then outputs the time-series content or a specific portion of the time-series content represented by the content data. For example, the search result output unit 125-n displays the time-series content or a specific portion of the time-series content.For example, when the search result output unit 125-n displays multiple pieces of time-series content or specific parts of the time-series content, the search result output unit 125-n may display thumbnail images 1241a-n of the multiple pieces of time-series content or specific parts of the time-series content on the display screen 1241-n, as illustrated in Fig. 11A. In this case, the search result output unit 125-n may also display a selection window 1241b-n for selecting a thumbnail image 1241a-n on the display screen 1241-n. In this case, the search result output unit 125-n may move the display position of the selection window 1241b-n under the control of the search processing unit 126-n based on an input to the input receiving unit 122-n or the utterance receiving unit 123-n. Furthermore, based on a selection input to the input receiving unit 122-n or the utterance receiving unit 123-n, the search processing unit 126-n may send the time-series content or the video 1241c-n of the specific part of the time-series content corresponding to the specific thumbnail image 1241a-n to the search result output unit 125-n, and the search result output unit 125-n may display the video 1241c-n of the time-series content or the specific part of the time-series content on the display screen 1241-n, as illustrated in Fig. 11B. As described above, the content providing unit 117 of the search support device 11 (Fig. 2) causes the terminal device 12-n (Fig. 3) to execute a search result output process that outputs the time-series content or the specific part of the time-series content selected by the utterance S1 (first utterance) (step S116g).

[0028] <Features of this form> In this embodiment, the search support device 11 (FIG. 2) receives an utterance S2 (second utterance) for extracting a subset of utterance leading information from a set of utterance leading information that guides an utterance S1 (first utterance) for selecting time-series content or a specific part of the time-series content, causes a terminal device 12-n (FIG. 3) to execute a search support process for displaying the subset of utterance leading information extracted based on the utterance S2 (second utterance) and user characteristic information that represents the characteristics of the user 1000-n who made the utterance S2 (second utterance), and causes the terminal device to execute a search result output process for outputting the time-series content or the specific part of the time-series content selected by the utterance S1 (first utterance). This allows the user 1000-n to easily search for desired time-series content or a specific part thereof that suits the user, simply by making an utterance in accordance with the displayed utterance leading information. In particular, by extracting a subset of speech-guiding information taking into consideration not only the speech of user 1000-n but also user characteristic information that represents the characteristics of user 1000-n, it is possible to easily search for time-series content or specific parts thereof that are suitable for the characteristics of user 1000-n.

[0029] Alternatively, the search support device 11 may cause the terminal device 12-n to execute the search support process after causing the terminal device 12-n to execute an initial search support process that displays a subset of utterance prompting information randomly selected from a set of utterance prompting information. In this initial search support process, user characteristic information representing the characteristics of the user 1000-n is not taken into consideration, which prevents the subset of utterance prompting information from being excessively limited by the characteristics of the user 1000-n. This prevents only a subset of utterance prompting information biased toward the past characteristics of the user 1000-n from being displayed, thereby preventing time-series content or a specific part of time-series content that is originally suitable for the user 1000-n from being excluded from the search target. As a result, it is possible to encourage the user 1000-n to discover novel time-series content or a specific part of time-series content that is suitable for the user 1000-n.

[0030] [Modification 1 of the First Embodiment] The search support process (step S116e) may be executed multiple times. In this case, the degree (contribution degree) of the user characteristic information (FIG. 5) considered when extracting the subset of the speech guidance information may be increased according to the number of executions of the search support process. That is, the search support processing unit 116 of the search support device 11 (FIG. 2) may cause the terminal device 12-n (FIG. 3) to execute the search support process multiple times. At this time, the relevance between the subset of the speech guidance information extracted in the A-th search support process and the user characteristic information may be lower than the relevance between the subset of the speech guidance information extracted in the B-th search support process executed after the A-th search support process and the user characteristic information. Here, A and B are positive integers, and A < B. In other words, the contribution degree of the user characteristic information when extracting the subset of the speech guidance information in the B-th search support process may be made larger than the contribution degree of the user characteristic information when extracting the subset of the speech guidance information in the A-th search support process. For example, as the number of executions of the search support process increases, the search support processing unit 116 may increase the contribution degree of the user characteristic information considered when extracting the subset of the speech guidance information. Thereby, when the number of executions of the search support process is small, since the past characteristics of the users 1000-n are not considered so much, it is possible to encourage the discovery of new time-series content suitable for the users 1000-n or a specific part of the time-series content. On the other hand, when the search support process increases, since the past characteristics of the users 1000-n are strongly considered, it is possible to suppress a situation where, despite the increase in the search support process, the user 1000-n cannot reach a time-series content or a specific part of the time-series content suitable for the user 1000-n.

[0031] [Modification Example 2 of the First Embodiment] When performing a search, even if a user cannot think of a word to use as a search key, it is easy to exclude unnecessary words from the displayed words. Therefore, by excluding utterances related to the user 1000 from the displayed subset of utterance guide information, it is possible to easily narrow down the subset of utterance guide information to one that is suitable for the user 1000. That is, the search support processing unit 116 of the search support device 11 (FIG. 2) may cause the terminal device 12-n (FIG. 3) to repeatedly execute a search support process in which the terminal device 12-n (FIG. 3) sets a set obtained by excluding utterance guide information related to utterance S2 (second utterance) from the displayed subset of utterance guide information as a new set of utterance guide information. For example, the search support processing unit 116 may cause the terminal device 12-n to perform an initial search support process in which a subset of speech leading information randomly selected from a set of speech leading information is displayed, and then the search support processing unit 116 may cause the terminal device 12-n to repeatedly perform a search support process in which a new subset of speech leading information is displayed by excluding speech leading information related to the utterance S2 (second utterance) from the displayed subset of speech leading information, and the new subset of speech leading information is used as a new set of speech leading information. Here, the speech leading information related to the utterance S2 (second utterance) may be speech leading information expressed in the utterance S2, speech leading information whose degree of agreement with the utterance S2 exceeds a threshold, speech leading information belonging to a field excluded by the utterance S2, or speech leading information obtained by applying information representing the utterance S2 to a machine-learned model. The speech leading information related to the utterance S2 (second utterance) may be speech leading information extracted based on user characteristic information (e.g., attribute information) of the user 1000-n and the utterance S2 (second utterance). The speech leading information related to the utterance S2 (second utterance) may be speech leading information extracted based on the display position of the displayed subset of speech leading information, the utterance S2 (second utterance), and the user characteristic information. To extract specific speech leading information based on the user characteristic information and utterance of the user 1000-n, or to extract specific speech leading information based on the display position of the displayed subset of speech leading information, the utterance, and the user characteristic information, for example, the methods described in <Example of utterance S1 (first utterance)> and <Example of utterance S2 (second utterance)> may be used.Furthermore, as exemplified in the first modification of the first embodiment, the degree (degree of contribution) of the user characteristic information ( FIG. 5 ) taken into consideration when extracting utterance guidance information may be increased depending on the number of times the search support process is executed. That is, the relevance between the utterance guidance information and the user characteristic information extracted in the Ath search support process may be lower than the relevance between the utterance guidance information and the user characteristic information extracted in the Bth search support process executed after the Ath search support process.

[0032] FIG. 12 illustrates how a subset of utterance leading information is narrowed down by repeating such search support processing. In the example of FIG. 12, first, a subset of utterance leading information, "dining, news, sports, beauty, real estate, business, travel...," is randomly extracted from the set of utterance leading information and displayed (step S2231, FIG. 13A). In response to this, assume that the user 1000 makes an utterance S2 related to the field "dining, news, beauty, real estate, business...," which he does not want to see. As a result, a subset of utterance leading information, "sports, real estate, travel," is displayed, excluding the utterance leading information "dining, news, beauty, real estate, business..." related to utterance S2 from the subset of utterance leading information, "dining, news, beauty, real estate, business..." (step S2232). This subset of utterance leading information, "sports (baseball, soccer, swimming...), real estate (urban, rural), travel (domestic, overseas)," is used as the set of utterance leading information (step S2231). Next, suppose that the user 1000 makes an utterance S2 related to the fields "Sports (baseball, soccer, etc.), Real Estate (urban, rural), Travel (overseas)" that the user does not want to see. As a result, a subset of utterance leading information "Sports, Real Estate, Travel" is extracted and displayed from the set of utterance leading information "Sports (baseball, soccer, swimming, etc.), Real Estate (urban, rural), Travel (domestic, overseas)" by excluding the utterance leading information "Food & Drink, News, Beauty, Real Estate, Business, etc." (Step S2232, FIG. 13B). In this way, by repeatedly performing the search support process in which the set obtained by excluding the utterance leading information related to utterance S2 from the displayed subset of utterance leading information is used as a new set of utterance leading information, the subset of utterance leading information suitable for the user 1000 can be easily narrowed down.

[0033] [Modification 3 of the First Embodiment] Contrary to the second modification of the first embodiment, the displayed subset of speech leading information may be updated to one that is appropriate for the user 1000 by adding utterances related to the user 1000 to the displayed subset of speech leading information. That is, the search support processing unit 116 of the search support device 11 (FIG. 2) may repeatedly cause the terminal device 12-n (FIG. 3) to execute a search support process in which a set is created by adding utterance leading information related to the utterance S2 (second utterance) to the displayed subset of speech leading information, and the new set of speech leading information is created. Alternatively, the search support processing unit 116 may repeatedly cause the terminal device 12-n (FIG. 3) to execute a search support process in which a set is created by adding utterance leading information belonging to the same or similar field as the utterance leading information related to the utterance S2 (second utterance) to the displayed subset of speech leading information, and the new set of speech leading information is created. The rest is as described in the second modification of the first embodiment.

[0034] [Hardware configuration] The functions performed by the components described herein may be implemented in circuitry or processing circuitry, including general-purpose processors, application-specific processors, integrated circuits, ASICs (Application Specific Integrated Circuits), a CPU (a Central Processing Unit), conventional circuits, and / or combinations thereof, programmed to perform the described functions. A processor includes transistors and other circuits and is considered to be circuitry or processing circuitry. A processor may also be a programmed processor that executes programs stored in memory.

[0035] In this specification, a circuitry, unit, or means is hardware that is programmed to realize or performs the described functions, which may be any hardware disclosed herein or any hardware known to be programmed to realize or perform the described functions.

[0036] If the hardware is a processor considered to be a type of circuitry, the circuitry, means, or unit is a combination of the hardware and software used to configure the hardware and / or processor.

[0037] The program describing the above-described processing contents can be recorded on a computer-readable recording medium. Examples of computer-readable recording media are non-transitory recording media. Examples of such recording media include magnetic recording devices, optical disks, magneto-optical recording media, and semiconductor memories.

[0038] The program may be distributed, for example, by selling, transferring, lending, etc. a portable recording medium such as a DVD or CD-ROM on which the program is recorded. Furthermore, the program may be stored in a storage device of a server computer, and then transferred from the server computer to another computer via a network, thereby distributing the program.

[0039] A computer that executes such a program may first temporarily store the program recorded on a portable recording medium or transferred from a server computer in its own storage device. Then, when executing a process, the computer reads the program stored on its own recording medium and executes the process in accordance with the read program. Alternatively, the computer may read the program directly from a portable recording medium and execute the process in accordance with the program. Furthermore, the computer may execute the process in accordance with the program each time a program is transferred from a server computer to the computer. The server computer may not transfer the program to the computer, but may instead execute the process through a so-called ASP (Application Service Provider) service, which realizes the processing function by issuing an execution instruction and obtaining the results. Furthermore, the server computer may execute the process on a terminal using a so-called SaaS (Software as a Service) service, which allows users to use part of the server computer along with the program. In this embodiment, the program includes information used for computer processing that is equivalent to a program (such as data that is not a direct instruction to the computer but has properties that define computer processing).

[0040] Furthermore, in this embodiment, the device is configured by executing a predetermined program on a computer, but at least a part of the processing contents may be realized by hardware.

[0041] [Other variations] The present invention is not limited to the above-described embodiment. For example, in the above-described embodiment, the search result output unit 125-n of the terminal device 12-n displays the selected time-series content or a specific portion of the time-series content. However, the selected time-series content or a specific portion of the time-series content may be projected, output as sound, or reproduced in a virtual space. Furthermore, the various processes described above may not only be executed in time series as described, but may also be executed in parallel or individually depending on the processing capabilities of the device executing the processes or as needed. Needless to say, other modifications are possible within the spirit and scope of the present invention.

[0042] [Note] The items explained above are listed below. [Appendix 1] a search support processing unit that causes the terminal device to execute a search support process to receive a second utterance for extracting a subset of utterance guide information from a set of utterance guide information that guides a first utterance for selecting the time-series content or a specific part of the time-series content, and to display the subset of utterance guide information extracted based on user characteristic information that indicates characteristics of the user who makes the second utterance and the second utterance; a content providing unit that causes the terminal device to execute a search result output process to output the time-series content or a specific part of the time-series content selected by the first utterance; A search support device having the above. [Appendix 2] 10. The search support device of claim 1, The search support processing unit causes the terminal device to execute an initial search support process that displays a subset of utterance guide information randomly selected from the set of utterance guide information, and then causes the terminal device to execute the search support process. [Appendix 3] 3. A search support device according to claim 1 or 2, comprising: the search support processing unit causes the terminal device to execute the search support processing multiple times; a search support device in which the relevance between the subset of utterance guidance information extracted in the search support processing at time A and the user characteristic information is lower than the relevance between the subset of utterance guidance information extracted in the search support processing at time B, which is executed after the search support processing at time A. [Appendix 4] 4. The search support device of claim 1, The search support processing unit causes the terminal device to repeatedly execute the search support process to set a new set of utterance guide information, which is a set obtained by excluding utterance guide information related to the second utterance from the displayed subset of utterance guide information, or a set obtained by adding utterance guide information related to the second utterance to the displayed subset of utterance guide information. [Appendix 5] 10. The search support device of claim 1, the search support processing unit causes the terminal device to execute an initial search support process to display a subset of speech guide information randomly selected from the set of speech guide information, and then causes the terminal device to repeatedly execute the search support process to display a new subset of speech guide information by excluding the speech guide information related to the second utterance from the displayed subset of speech guide information, and to set the new subset of speech guide information as a new set of speech guide information. [Appendix 6] 6. The search support device of claim 4 or 5, The search support device, wherein utterance guide information related to the second utterance is extracted based on the user characteristic information and the second utterance. [Appendix 7] 7. The search support device of any one of appendix 4 to 6, The search support device extracts the utterance guide information related to the second utterance based on a display position of the displayed subset of the utterance guide information, the second utterance, and the user characteristic information. [Appendix 8] 8. The search support device of any one of appendixes 1 to 7, The search support device, wherein the search support processing unit extracts a new set of the utterance guide information based on a display position of the displayed subset of the utterance guide information, the second utterance, and the user characteristic information. [Appendix 9] a speech receiving unit that receives a second utterance for extracting a subset of speech guide information from a set of speech guide information that guides a first utterance for selecting the time-series content or a specific part of the time-series content; an utterance guide information display unit that displays a subset of the utterance guide information extracted based on user characteristic information representing a characteristic of the user who makes the second utterance and the second utterance; a search result output unit that outputs the time-series content or a specific part of the time-series content selected by the first utterance; A terminal device having the above configuration. [Appendix 10] A search support device and a terminal device are included, The search support device a search support processing unit that causes the terminal device to execute a search support process to receive a second utterance for extracting a subset of utterance guide information from a set of utterance guide information that guides a first utterance for selecting time-series content or a specific part of the time-series content, and to display the subset of utterance guide information extracted based on user characteristic information that indicates characteristics of a user who makes the second utterance and the second utterance; a content providing unit that causes the terminal device to execute a search result output process to output the time-series content or a specific part of the time-series content selected by the first utterance; Including, The terminal device an utterance receiving unit that receives the second utterance; a speech guide information display unit that displays a subset of the speech guide information extracted based on the user characteristic information and the second utterance; a search result output unit that outputs the time-series content or a specific part of the time-series content selected by the first utterance; A search system, including: [Appendix 11] A search support device according to any one of claims 1 to 8; a terminal device for using the search support device via a network; A search system having: [Explanation of symbols]

[0043] 1. Search System 11 Search support device 12 Terminal Equipment 116 Search support processing unit 117 Contents Provider 124 Speech guidance information display unit 125 Search result output section

Claims

1. a search support processing unit that causes the terminal device to execute a search support process to receive a second utterance for extracting a subset of utterance guide information from a set of utterance guide information that guides a first utterance for selecting the time-series content or a specific part of the time-series content, and to display the subset of utterance guide information extracted based on user characteristic information that indicates a characteristic of the user who makes the second utterance and the second utterance; a content providing unit that causes the terminal device to execute a search result output process that outputs the time-series content or a specific part of the time-series content selected by the first utterance; A search support device having the above.

2. 2. The search support device of claim 1, The search support processing unit causes the terminal device to execute an initial search support process that displays a subset of utterance guide information randomly selected from the set of utterance guide information, and then causes the terminal device to execute the search support process.

3. 3. A search support device according to claim 1 or 2, the search support processing unit causes the terminal device to execute the search support processing multiple times; A search support device in which the relevance between the subset of speech guidance information extracted in the Ath search support process and the user characteristic information is lower than the relevance between the subset of speech guidance information extracted in the Bth search support process executed after the Ath search support process and the user characteristic information.

4. 3. A search support device according to claim 1 or 2, The search support processing unit causes the terminal device to repeatedly execute the search support process to set a new set of speech guide information, which is a set obtained by excluding speech guide information related to the second utterance from the displayed subset of speech guide information, or a set obtained by adding speech guide information related to the second utterance to the displayed subset of speech guide information.

5. A program for causing a computer to function as the search support device of claim 1.

Images