Information processing program, information processing method, and information processing device

Abstract

To provide an information processing program, an information processing method and an information processing unit, which provide information by estimating information and a method desired by a user.SOLUTION: An information processing unit 1 includes: restriction time estimating means 100 for estimating a restriction time that an article can be reproduced on the basis of at least one of a content of the article of content information 111, user information 112 of a user hearing voice which reproduces the article and a space being an object of the article; interest degree estimation means 101 for estimating an interest degree to the article and a sentence included in the article on the basis of the user information 112; content selection means 102 for selecting one or a plurality of articles on the basis of restriction time and the interest degree of the user from the plurality of articles; and scenario information generating means 104 for generating scenario information 114 on the basis of the restriction time and the interest degree of the user by using the sentences of the one or the plurality of articles selected by the content selection means 102.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an information processing program, an information processing method, and an information processing device. [Background technology]

[0002] As a conventional technology, an information processing program has been proposed that selects articles of high interest from a group of articles in descending order of interest based on conditions such as total audio time length, speech speed, and required number of articles to be played back, summarizes the selected articles to an appropriate amount, and plays them back in synthetic voice (see, for example, Patent Document 1).

[0003] The information processing program disclosed in Patent Document 1 comprises a parameter setting means for setting the desired speech rate, speaker type, and duration of the synthesized speech, a keyword setting means for setting keywords and a level of interest, an interest level determination means for determining a level of interest based on the set keywords and level of interest, a summary level determination means for determining a summary level of the text based on the set speech rate, listening time, and the determined level of interest, a text summarization means for summarizing the text based on the summary level, and a text-to-speech synthesis means for synthesizing speech based on the speech rate and speaker type set by the parameter setting means, and summarizes and plays back an article based on the set information. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2008-164759 Summary of the Invention [Problem to be solved by the invention]

[0005] According to the above information processing program, articles are summarized and played back based on set information such as speaking speed, speaker type, length of time, and keywords, but the length of time, article selection, degree of summarization (policy), etc. must be set according to information about the user. This is not easy to set, especially if the user does not understand the appropriate length of time or degree of summarization, and there is a problem in that it cannot be said to be fully optimized for individuals.

[0006] An object of the present invention is to provide an information processing program, an information processing method, and an information processing device that estimate the information and method desired by a user and provide the information. [Means for solving the problem]

[0007] In order to achieve the above object, one aspect of the present invention provides the following information processing program, information processing method, and information processing device.

[0008] [1] Computer, a time constraint estimation means for estimating a time constraint for playing the article based on at least one of the content of the article, information on a user who listens to the audio of the article being played, and a space targeted by the article; an interest level estimation means for estimating a degree of interest in the article and sentences contained in the article based on the user information; an article selection means for selecting one or more articles from the plurality of articles based on the time limit and the interest level of the user; an information processing program that functions as a generation means for generating scenario information using sentences from one or more articles selected by the article selection means, based on the time constraint and the interest level of the user; [2] The information processing program according to [1], further functioning as article presentation order determination means for determining the presentation order of the plurality of articles selected by the article selection means. [3] The interest level estimation means changes the estimation result of the interest level based on the user's reaction to the reproduction of the scenario information, The information processing program according to [1] or [2], wherein the article selection means changes the selection of one or more articles based on the estimation result. [4] The information processing program described in [3], wherein the generation means regenerates the scenario information based on the user's response to the playback of the scenario information and / or based on the changed interest level and the changed selection content. [5] An information processing program according to any one of [1] to [4], wherein the generation means selects one or more sentences to be played to the user from a plurality of sentences contained in the article, and one or more sentences to respond to the user's reaction, and generates the scenario information. [6] a time constraint estimation step of estimating a time constraint for playing the article based on at least one of the content of the article, information on users who listen to the audio of the article, and the space targeted by the article; an interest level estimation step of estimating a degree of interest in the article and sentences contained in the article based on the user information; an article selection step of selecting one or more articles from the plurality of articles based on the time limit and the interest level of the user; an information processing method comprising: a generating step of generating scenario information based on the time constraint and the interest level of the user using sentences from one or more articles selected in the article selecting step. [7] A time constraint estimation means for estimating a time constraint for playing the article based on at least one of the content of the article, information on the user who listens to the audio of the article, and the space targeted by the article; an interest level estimation means for estimating a degree of interest in the article and sentences contained in the article based on the user information; an article selection means for selecting one or more articles from the plurality of articles based on the time limit and the interest level of the user; an information processing device comprising: a generation means for generating scenario information based on the time limit and the interest level of the user, using sentences from one or more articles selected by the article selection means;

[0009] According to the present invention, it is possible to provide information by estimating the information and method desired by the user. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a schematic diagram illustrating an example of a configuration of an information processing system according to an embodiment. [Figure 2] FIG. 2 is a block diagram illustrating an example of the configuration of an information processing device according to an embodiment. [Figure 3] FIG. 3 is a schematic diagram showing an example of the configuration of content information. [Figure 4] FIG. 4 is a schematic diagram showing an example of content information expressed in a tree structure. [Figure 5] FIG. 5 is a schematic diagram showing the relationship between user information and interest level information. [Figure 6] FIG. 6 is a schematic diagram showing an example of the configuration of scenario information. [Figure 7] 7(a) and (b) are schematic diagrams for explaining the scenario reproducing operation. [Figure 8] FIG. 8 is a flowchart showing an example of the operation of the information processing device. DETAILED DESCRIPTION OF THE INVENTION

[0011] [Embodiment Mode] (Configuration of information processing system) FIG. 1 is a schematic diagram illustrating an example of a configuration of an information processing system according to an embodiment.

[0012] As an example, this information processing system is installed to provide guidance to an art gallery, museum, etc. where exhibits 4a to 4d are displayed, and is configured by connecting an information processing device 1 that generates information for guiding around the exhibits 4a to 4d and a guide robot 2 that guides a user 3 around the exhibits 4a to 4d based on the guidance information generated by the information processing device 1 so that the two devices can communicate with each other via a network (not shown). Note that while the following description focuses on facilities such as art galleries and museums that exist in real life, this embodiment may also be applied to facilities that exist in a virtual reality space. Furthermore, the location where the system is installed is not limited to art galleries, museums, etc., and any type of facility or location that requires guidance or explanation may be used.

[0013] The information processing device 1 is a server-type information processing device that operates in response to requests from the user 3 via the guide robot 2. It includes electronic components such as a central processing unit (CPU) for processing information and a flash memory for storing information. The information processing device 1 stores content information for explaining the exhibits 4a-4d, such as the creators, creation dates, and anecdotes of the exhibits 4a-4d. The information processing device 1 estimates the time required for the guide robot 2 to provide a tour (time constraint) based on user information indicating the user 3's attributes, such as age, hometown, occupation, industry, hobbies, and interest in the article genre. It determines the content to be guided and its order, generates the content of the tour, and outputs this information to the guide robot 2. If the user 3 responds to the guide robot 2 by asking a question or other response during the tour, the information processing device 1 changes the time constraint, the content to be guided and its order, the content of the tour, and other information based on the response. The content information may also be acquired from an external source.

[0014] The guide robot 2 is a robot that can move under control or autonomously, and is capable of at least outputting and listening to voices, and is equipped with electronic components such as a CPU with functions for processing information, a flash memory, a speaker, a microphone, and a camera within the body.Instead of the guide robot 2, the user 3 may carry a portable device such as a smartphone, a tablet terminal, VR (Virtual Reality) goggles, or AR (Augumented Reality) glasses, or one or more built-in devices such as smart speakers or network-enabled televisions may be placed in each of the exhibits 4a to 4d or in each area.

[0015] The scenario information played back by the guide robot 2 will be described as being primarily audio information in the embodiment, but it may also be video information, text information, image information, or a combination of these.

[0016] Furthermore, the guide robot 2 acquires, via a microphone, mainly voices uttered by the user 3 as a reaction to the contents of the scenario information being reproduced, and transmits the acquired voices to the information processing device 1 as reaction information. If the guide robot 2 is able to respond within the scope of the scenario information, it analyzes the contents of the reaction information and responds from the information included in the scenario information 114 according to the analysis results. Note that, although the embodiment will be described mainly assuming that the acquired reaction is voice information, the reaction of the user 3 is not limited to voice information, and if the guide robot 2 of the user 3 has an operation unit, it may also be video information capturing the operation of the operation unit, the facial expressions, gaze, gestures, etc. of the user 3.

[0017] The network 4 is a communication network capable of high-speed communication, and is, for example, a wired or wireless communication network such as the Internet, an intranet, or a LAN (Local Area Network).

[0018] In the above configuration, as an example, the information processing device 1 estimates the time (time constraint) for the guide robot 2 to guide the user 3 based on the user information of the user 3, selects exhibits that fit within the guide time and are appropriate for the user information, and generates scenario information for guiding the user through the exhibits from the content information. Note that the scenario information includes not only information for explaining the exhibits but also information for responding to the user 3's reactions. The guide robot 2 responds to the user 3's reactions while guiding the user 3 based on the scenario information. The configuration will be explained in more detail below.

[0019] (Configuration of information processing device) FIG. 2 is a block diagram showing an example of the configuration of the information processing device 1 according to the embodiment.

[0020] The information processing device 1 comprises a control unit 10 which is composed of a CPU and the like and controls each part and executes various programs, a memory unit 11 which is composed of a storage medium such as a flash memory and stores information, and a communication unit 12 which communicates with the outside via a network.

[0021] The control unit 10 executes the information processing program 110 described later to function as a constraint time estimation means 100, an interest level estimation means 101, a content selection means 102 (article selection means), a content presentation order determination means 103, a scenario information generation means 104, an output means 105, and a response acquisition means 106, etc.

[0022] The constraint time estimation means 100 estimates the constraint time during which the guide robot 2 can provide guidance based on the content of the article in the content information 111, user information 112 which is information about the user who listens to the audio of the reproduced article, information about the space that the article targets (the size of the museum, the positions of the exhibits 4a to 4d, and the distance between the exhibits 4a to 4d), and the time set by the user 3 or the administrator of the information processing device 1. Note that if movement is not required, there is no need to take into account the space information, and the movement time of the guide robot 2 does not need to be included in the constraint time (the constraint time becomes the speaking time).

[0023] The interest level estimation means 101 estimates the interest of the user 3 in the content information 111 based on the user information 112, and stores the estimation result as interest level information 113 in the storage unit 11. Specifically, for example, profile features obtained by a questionnaire or the like are used as the user information 112, and the interest level of the user 3 in each sentence is estimated based on the profile features and features of the sentences in the content information 111. In addition to the questionnaire, the age and gender estimated from a video of the user 3, and a persona implicitly estimated from the interaction between the guide robot 2 and the user 3 obtained before the content information 111 is explained may also be used as profile features.

[0024] The content selection means 102 selects content to be transmitted to the user 3 from the content information 111 based on the time limit and interest level information 113. For example, the content is selected so that the sum of the interest levels is maximized within the time limit.

[0025] The content presentation order determination means 103 determines the presentation order of the contents selected by the content selection means 102. Specifically, in the example of a museum, the content presentation order determination means 103 basically selects an order that will take visitors around the works in the shortest time. However, an order (route) other than the shortest time may be selected, taking into consideration the story of the selected content and the congestion status of the museum. Furthermore, the order may be selected in order of the user 3's interest, by genre, or by chronological order.

[0026] The scenario information generation means 104 checks the content selected by the content selection means 102 and the presentation order determined by the content presentation order determination means 103, and generates scenario information 114 by assigning accompanying information (playback instruction information) such as whether or not to play and the order of playback to the selected content sentences (playback units). It also modifies the wording, for example, to a colloquial tone, as needed. For example, accompanying information may be assigned to units with a high interest level, setting them to play and not to play (to be used for responding to the user's 3's reactions). (The structure that ensures consistency of the sentences by the playback order and provides units for playback and units for responses is hereinafter referred to as a "discourse structure.") For example, the scenario information generation means 104 selects units to play so as to maximize the sum of the interest levels within a time limit. In addition to automatically assigning additional information to sentences, it can also receive and assign additional information, such as definitions, edited by an operator using an authoring tool (not shown). Here, the scenario information 114 is information configured to not only explain the exhibits 4a-4d but also respond to the user's 3's reactions to the explanation. Furthermore, after the colloquialization process, speech synthesis is performed to generate a speech file, and the speech time can be obtained from the playback time of the speech file. As a modified example, speech times may be defined in advance for each document, paragraph, sentence, clause, etc., and speech synthesis may be performed to adhere to these times. Details will be described later.

[0027] In addition, an example will be described in which the scenario information generation means 104 generates an audio file by converting a set of sentences (units) into audio information consisting of, for example, one or more speech clauses, and has link information to the audio file as an item in the scenario information 114, but the scenario information 114 may not have a link to the audio file as an item, but may instead hold a set of units as text information, and play back audio from the text information in the guide robot 2, which is a playback device.

[0028] Note that, although an example in which the scenario information generating means 104 generates one piece of scenario information 114 from one piece of content information 111 will be described below, one piece of scenario information 114 may be generated from multiple pieces of content information 111, or multiple pieces of scenario information 114 may be generated from one piece of content information 111. Furthermore, the scenario information 114 may not only use the units contained in the content information 111 as they are (extraction type), but may also be created by reconstructing the units using the units or by paraphrasing the units into other units (abstraction type), or a combination of these. In other words, the scenario information 114 may be something like a summary in which communication information and response information are selected and discarded from the content information 111, or it may be information different from the content information 111 that is newly generated by reconstructing the content information 111. In the case of the extraction type, requirements for a summary include high interest, lack of redundancy, and consistency of content.

[0029] The output means 105 outputs the scenario information 114 to the guide robot 2. In addition to the scenario information 114, the output means 105 also outputs the travel route, travel speed, etc. of the guide robot 2, but the travel route, travel speed, etc. may be determined by the guide robot 2.

[0030] The reaction acquisition means 106 acquires, via the guide robot 2, the reaction of the user 3 to the content reproduced by the guide robot 2 (such as an interjection or question).

[0031] The storage unit 11 stores an information processing program 110 that causes the control unit 10 to operate as each of the above-mentioned means 100-106, content information 111, user information 112, interest level information 113, scenario information 114, reaction information 115, and the like.

[0032] FIG. 3 is a schematic diagram showing an example of the configuration of the content information 111. As shown in FIG.

[0033] The content information 111 includes a work ID for identifying a work, a work title, a sentence ID for identifying a sentence describing the work, a dependent sentence ID indicating a parent node in the tree structure, a chunk sentence ID indicating a grouping with other sentences, attributes classifying the content of the sentence, attribute / discourse relationships indicating the discourse relationships indicating the semantic relationships between the dependent sentences, and the content of the sentence. One or more sentences are prepared for each work, and the multiple sentences are assumed to be categorized into an outline of the work, an author, a background of the work, etc., as described in the following Figure 4, but this is not a required configuration. While the content information 111 has been exemplified as being divided into multiple sentences in advance, undivided sentences may also be divided by content information division means (not shown). Furthermore, sentences may be further divided into units such as phrases, clauses, and words, or may be larger units than sentences. It should be noted that the content information 111 is not limited to text information, but may be content other than text, such as video, audio, or images, and in the case of information other than text, it is assumed to be divided into units appropriate for the type of content (chapters, scenes, frames, etc. for video, tracks, measures, etc. for audio, image ranges, pixels, etc. for images). The attributes and discourse relations are used as features when estimating the interest level, and are also used as decision materials when answering questions from user 3.

[0034] FIG. 4 is a schematic diagram showing an example of content information 111 expressed in a tree structure.

[0035] Content information 111a is an example of structuring work ID "001" of content information 111 when used as scenario information, and includes sentence 111a1, sentences 111a2 and 111a3 that supplement sentence 111a1, sentence 111a4 that supplements sentence 111a3, sentence 111a5 that further supplements sentence 111a4, and sentences 111a6 to 111a9 at the same level as sentences 111a2 and 111a3. This structuring may be performed in advance on content information 111, or may be performed by scenario information generating means 104 when the content information 111 is used.

[0036] FIG. 5 is a schematic diagram showing the relationship between the user information 112 and the interest level information 113. As shown in FIG.

[0037] User information 1121, 1122, 112 62 , 112 63 ... is information about the user, such as gender and age, prepared for each user, and as an example, multiple items are prepared and parameterized in the illustration, but this is not limited to this and it may be expressed by any feature expression such as a vector expression. The interest level information 113 is, for example, the interest level of each user calculated based on the user information 112 for each sentence and the whole (overall) of the content information 111a.

[0038] (Operation of information processing device) Next, the operation of this embodiment will be explained by dividing it into (1) basic operation, (2) scenario information generation operation, and (3) scenario information reproduction operation.

[0039] (1) Basic operation First, when a user 3 visits a facility such as an art gallery or a museum where an information processing system is installed, the guide robot 2 approaches the user 3 and identifies the user 3. The method of identification may be to register the user 3 in advance and compare the user 3 with registered information such as an image or a combination of the user 3 ID and password, or to newly register the user 3 when the user visits.

[0040] Furthermore, user information 112 showing the profile of user 3 is generated by, for example, conducting a questionnaire for user 3 at the time of registration, or based on the history of past guidance, if any. The user information 112 may also be acquired from an external service. In addition to the questionnaire, the age and gender estimated from a video image of user 3, or a persona implicitly estimated from the interaction between guide robot 2 and user 3 obtained before explaining the content information 111, may also be used as profile features.

[0041] Next, the user 3 requests guidance from the guide robot 2. Upon receiving the guidance request, the guide robot 2 communicates with the information processing device 1 and requests the information processing device 1 to generate scenario information. The information processing device 1 receives the request to generate scenario information and starts the "(2) Scenario information generation operation" described below. Note that the scenario information generation operation may be performed in advance based on the user 3's visit reservation, etc.

[0042] In the above operation, the guide robot 2 reproduces audio to the user 3, for example, as follows. "Hello. Welcome to the XX Museum. Is this your first time visiting the main building?" To this, User 3 answers, "Yes, that's right." "Thank you for visiting us for the first time. In the main building, we have a variety of Japanese national treasures and important cultural properties on display. We'll recommend some pieces for you and show you around."

[0043] (2) Scenario information generation operation FIG. 8 is a flowchart showing an example of the operation of the information processing device.

[0044] First, the time constraint estimation means 100 of the information processing device 1 estimates the time (including travel time) during which the guide robot 2 can speak based on the content of the article (the artwork and a sentence explaining the artwork) in the content information 111, user information 112 which is information about the user who listens to the audio of the article, information about the space covered by the article (the size of the museum, the positions of the exhibits 4a to 4d, and the distance between the exhibits 4a to 4d), the time set by the user 3 or the administrator of the information processing device 1, etc. (S10). Note that if the time constraint is set in advance, this step may be omitted.

[0045] 5, the interest level estimation means 101 estimates the interest of the user 3 in each work and each sentence in the content information 111 based on the user information 112, and stores the result as interest level information 113 in the storage unit 11 (S11). Specifically, the interest level of the user 3 in each sentence is estimated based on the profile feature amount obtained by a questionnaire or the like and the feature amount of the sentence in the content information 111.

[0046] Next, the content selection means 102 selects content to be transmitted to the user 3 from the content information 111 based on the time constraint and the interest level information 113 (S12). Note that the amount of information that can be transmitted is selected taking into consideration the travel time of the guide robot 2 between exhibits.

[0047] Next, the content presentation order determination means 103 determines the presentation order of the selected articles (S13). Specifically, in the example of a museum, the content presentation order determination means 103 basically selects an order that will allow the works to be toured in the shortest time. However, an order (route) other than the shortest time may be selected, taking into consideration the story of the selected content and the congestion status of the museum. Furthermore, an order based on the degree of interest of the user 3, an order grouped by genre, or a chronological order may also be selected.

[0048] Next, the scenario information generating means 104 generates, as scenario information 114, the amount of information that can be reproduced within the time obtained by subtracting the travel time of the guide robot 2 from the time constraint. Specifically, the scenario information generating means 104 adds accompanying information (reproduction instruction information) such as whether to reproduce and the order of reproduction based on the interest level information 113 to the sentences of the content selected by the content selecting means 102, and also changes the wording as necessary, for example, to colloquial language, thereby generating the scenario information 114, for example, as shown in Fig. 6 (S14).

[0049] Specifically, the scenario information generating means 104 generates scenario information 114 by adding accompanying information such as whether or not to play and the order of playback to units divided into paragraphs, sentences, phrases, words, and the like.

[0050] In addition, the summarized content is played back as accompanying information by playing back the units set to be played back in the order in which the clauses are presented, and the units not set to be played back are used to respond to the reaction of the user 3 to the played back content, and are also used for later playback by changing the scenario information 114.

[0051] FIG. 6 is a schematic diagram showing an example of the configuration of scenario information 114, which is an example of the configuration of scenario information 114 related to content information 111 after extracting sentences with a high level of interest so as not to violate the time limit or discourse structure.

[0052] As an example, scenario information 114a1 is information created for user 3 based on content information 111a in FIG. 4, and is generated by selecting sentences 111a1 to 111a5 from sentences 111a1 to 111a9 based on the numerical value of "U1" in interest level information 113 shown in FIG. 5 (by selecting sentences with a numerical value of 5 or more).

[0053] Like scenario information 114a1, scenario information 114a2 is information created for user 3 based on content information 114b with work ID "002", and is generated from sentences selected based on the value of "U1" in interest level information 113.

[0054] 6 does not show the accompanying information indicating whether or not to play and the order of playback, but it is assumed that it is set as described below. For example, in the above-mentioned scenario information 114a1, playback is set for sentences 111a1 to 111a3, but playback is not set for sentences 111a4 and 111a5, and sentences 111a4 and 111a5 are used as responses to the reaction of user 3, and are also used for later playback by subsequent changes to the accompanying information.

[0055] Furthermore, in order to respond to the reaction of user 3, the scenario information generation means 104 may, if necessary, obtain information from an external information database (not shown) or an information provider site on the Internet for units including proper nouns or terms that are thought to have low recognition, and add information such as definitions and trivia of the proper nouns, terms, etc.

[0056] Next, the output means 105 outputs the scenario information 114 to the guide robot 2. Note that the output means 105 also outputs the movement route, movement speed, etc. of the guide robot 2 in addition to the scenario information 114, but the movement route, movement speed, etc. may be determined by the guide robot 2.

[0057] (3) Scenario information playback operation Next, the guide robot 2 receives the scenario information 114 from the information processing device 1, and plays back the scenario information 114 to guide the user 3 (S15). The following examples show a case where a user 31 is guided and a case where a user 32 is guided.

[0058] 7(a) and (b) are schematic diagrams for explaining the scenario reproducing operation.

[0059] For example, when guiding user 31, as shown in Figure 7(a), the time limit is determined to be, for example, 5 minutes based on user information 1121 of user 31, exhibits 4a and 4b are selected as content, and the order of presentation is determined to be exhibits 4a, 4b.

[0060] First, guide robot 2 plays an introduction to the guide, such as "Today, I would like to introduce the Sixth Volume of the Ancient Book of Documents and the painting "Tsuchigumo Soushi Emaki." Let's take a look at them one by one." Then guide robot 2 guides user 31 to exhibit 4a (1). Next, it plays scenario information 114a1 for user 31, which is an explanation of exhibit 4a from scenario information 114, and explains exhibit 4a, such as "The Book of Documents is the classic book that records the history from Yao and Shun to Duke Mu of Qin. It is said to have been compiled by Confucius. The Book of Documents was hidden during the Qin book burning, but was discovered inside the walls during the Han dynasty." (2).

[0061] As described above, if user 3 responds to the explanation of exhibit 4a with a nod or a question, guide robot 2 responds to the response based on scenario information 114a1. For example, if user 3 requests additional explanation, such as "Tell me more," guide robot 2 plays prepared response sentences 111a4 and 111a5 and adds, "This was rewritten in the common characters of the Han Dynasty so that people of the time could read it, and is therefore called the 'Modern Book of Confucius'. On the other hand, what King Gong of Lu obtained from within the walls of Confucius' former residence during the reign of Emperor Jing of the Han Dynasty was written in ancient characters, and is therefore called the 'Ancient Book of Confucius'." Furthermore, if user 3 responds with a "Huh?", guide robot 2 may determine whether the user is interested based on intonation, etc.

[0062] Next, the guide robot 2 guides the user 31 to the exhibit 4b (3) and plays back the scenario information 114b1 that explains the exhibit 4b from the scenario information 114 (3). After that, the guide robot 2 guides the user 31 to the starting position (4) and ends the guidance.

[0063] Furthermore, when moving from exhibit 4a to exhibit 4b, a transition such as "That's an interesting work, isn't it?" may be played. If user 3 responds to this transition with "Yes, it's interesting," a transition such as "You seemed interested in scriptures, so would you like me to introduce another scripture before the next work?" may be played. If user 3 responds with "Please introduce one," the original content selection and presentation order will be changed to introduce another scripture. Note that the candidate for the next work to be introduced here is changed as a result of a comprehensive judgment based on user 31's reaction, and the above reaction does not necessarily change the candidate for the next work to be introduced.

[0064] If user 3 is gazing at a part of exhibit 4b while the explanation of exhibit 4b is being played back, the explanation of the part being watched may be played back even if it is a sentence in scenario information 114b that is not specified for playback. Also, if user 3 does not seem interested while the explanation of exhibit 4b is being played back, it may be determined that user 3 is not interested in the genre of exhibit 4b, and subsequent content selection or modification of scenario information may be performed.

[0065] Also, when the tour finishes, Guide Robot 2 may play a closing message such as, "It's a fascinating piece, isn't it? That concludes the pieces we'll be introducing today. Today, you were able to come into contact with a variety of national treasures and important cultural properties. What did you think of touring with us? Japan has many valuable pieces of art. Next time, we'd like to introduce some other pieces. Please come and see us again. We'll be waiting for you."

[0066] 7(b), when guiding user 32, the time limit is determined to be, for example, 15 minutes based on user information 1122 of user 32, and exhibits 4b, 4d, and 4c are selected as the content. Basically, to guide the user over the shortest distance (shortest travel time), exhibits 4b, 4c, and 4d are presented in this order, but taking into consideration interest level information 1132 of user 32 and the story flow of the selected content, the initial plan may be to present exhibits 4b, 4d, and 4c in this order, for example.

[0067] First, the guide robot 2 guides the user 32 to exhibit 4b (1) and plays back scenario information 114b2 for the user 32, which explains exhibit 4b, from the scenario information 114 (2). Next, the guide robot 2 guides the user 32 to exhibit 4d (3) and plays back scenario information 114d2, which explains exhibit 4d, from the scenario information 114 (4). Next, the guide robot 2 guides the user 32 to exhibit 4c (5) and plays back scenario information 114c2, which explains exhibit 4c, from the scenario information 114 (6). Thereafter, the guide robot 2 guides the user 32 to the starting position (7) and ends the guidance.

[0068] Furthermore, if a reaction is obtained from the user 32 while the exhibits 4b and 4d are being presented and the interest level information 1132 is changed, the initial plan of presenting the exhibits 4b, 4d, and 4c in that order may be changed so that a new exhibit 4e is presented, such as exhibits 4b, 4d, and 4e.

[0069] Furthermore, during the above-mentioned guidance by the guide robot 2, in addition to the examples described above, users 31 and 32 (hereinafter referred to as user 3) listen to the played audio and, if there is a backchannel, question, doubt, operation command, etc. in the middle of the utterance clause being played, they speak as an interruption response. Also, user 3 speaks as a response if there is a backchannel, question, doubt, operation command, etc. between utterance clauses. The response is received via the microphone of the guide robot 2. Specific examples of the content of the reactions and responses will be explained in more detail below.

[0070] The guide robot 2 transmits the content of the reaction to the information processing device 1. The reaction acquisition means 106 of the information processing device 1 receives the content of the reaction and stores it in the storage unit 11 as reaction information 115 (S16).

[0071] The content selection means 102 and / or scenario information generation means 104 of the information processing device 1 calculates the interest level information 113, selects content, determines the presentation order, and / or modifies the contents of the scenario information 114 (including accompanying information) based on the reaction information 115 (S17). The modification may be performed each time a reaction is received from the user 3, for each set of a predetermined number of reactions, at predetermined time intervals, for each content, for each sentence, or at other intervals.

[0072] The above-mentioned corrections are made, for example, in the following cases: The scenario information 114 is updated appropriately (for each piece of content) taking into consideration the reaction of the user 3. Furthermore, when the user 3 shows a reaction, it is estimated from the reaction whether the user seemed interested in the work or not, and based on that result, the level of interest in the content is re-estimated, the work is selected again, the route is decided, and the scenario is corrected.

[0073] Furthermore, if the usage situation allows the time limit determined in step S10 to be changed, the time limit may be extended or shortened as appropriate, taking into consideration the convenience of user 3. If it is determined that user 3 still wants to listen and has time to spare, the time limit is extended and the scenario is revised based on the updated time limit and interest level. Furthermore, if it is determined that user 3 does not seem interested in any of the works, or if it is determined that user 3 wants to end the explanation early for his or her own convenience, the time limit is shortened and the scenario is revised so that the explanation can be completed early.

[0074] On the other hand, the guide robot 2 can respond autonomously and within the range of the received scenario information 114, and in this case, operates as follows.

[0075] The response analysis means of the guide robot 2 recognizes the user's speech, analyzes the response based on the speech content and timing of the response, and classifies the analysis result into a predetermined response type. If the classification result of the analysis result information indicates that a response is required, the robot responds using information corresponding to the answer from the scenario information 114.

[0076] For example, when the user 3 advances or rewinds the speech clauses with an operation such as "next sentence," "previous sentence," or "say it again," the speech clause to be played back is switched to "n=n+1," "n=n-1," or "n=n" even while the nth document is being played back. Furthermore, the speech clause to be played back may be switched by jumping to a specific speech clause, rather than just to an adjacent speech clause.

[0077] Based on the classification results of the analysis result information, the guide robot 2 can respond with one of the following: a command response, a template response, a scenario response, a supplementary response, an open response, or a default response. A command response is a response when the response type is an operation command such as "Wait a moment," "Go back," or "Go to the next article." The command response is a response indicating that the response has been accepted and performing an operation according to the operation command. A template response is a response when the response conforms to a predetermined rule and is a response that conforms to the rule. For example, if the response includes an adjective, the robot responds by repeating the same expression, such as "It's interesting, isn't it?" in response to the response "Interesting." If the response is an adjective with confirmation, the robot responds with "I think so." If the response is self-disclosure, the robot responds with "I see." If the response is an opinion such as "I think...," the robot responds with "Yes, that's true." If the response is an opinion such as "You should...," the robot responds with "Yes, I think that's how it should be." If the response is an opinion such as "You shouldn't...," the robot responds with "Yes, well, that's true."

[0078] Supplementary responses include responses based on similarity, responses with additional information, responses with noun repetition, and responses with error handling. A similarity-based response responds with the sentence with the highest similarity among the sentences in the scenario information 114 whose similarity between the content words included in the reaction and the sentences is equal to or exceeds a predetermined threshold. A response with additional information responds by adding and replaying another sentence from the same paragraph as the reaction if the reaction includes an adjective that indicates interest in the reaction (e.g., "interesting") or if the reaction is a word that encourages replay, such as "and then" or "the rest." A noun repetition response responds with "Yes, ~" if the reaction is a noun included in the reaction. A response with error handling responds by repeating or defining the term if the reaction is a misspelling of a term in the reaction. For example, in response to the question "Glucola" or "What's Glucola?", it responds with "Glucolaphasatin," and in response to the question "Progress?", it responds with "Kobe Steel is..." An open response is a response to a question whose answer is information not included in the scenario information 114, and is a response that searches for an answer using an external service. A default response is a response that says "I don't really know about that" when the response is a question or confirmation but the answer cannot be given. A default response is a response that says "I don't really know about that" when the response is not a question or confirmation and none of the above responses are given, and it responds by saying "Yes."

[0079] When a scenario response is made in the above case, the response content is determined as follows using the scenario information 114.

[0080] Based on the analysis result, the guide robot 2 searches for a unit that will be an answer to the reaction of the user 3 from the units that make up the scenario information 114. The search is performed by tracing the history of the units that have been played and searching from the units that have been played.

[0081] For example, if a user responds with "What is Company Alpha?", the first pattern is to search for a definition of "Company Alpha" by tracing back the playback history from the unit (phrase, etc.) that was being played at the time of the response.

[0082] As a second pattern, the playback history is traced back from the unit (phrase, etc.) that was being played back at the time of the response, and a unit that includes a description of "Company α" is searched for.

[0083] If no results are found in the above search, the entire scenario information 114 is searched because information that should respond to a reaction may be included in an untransmitted unit.

[0084] For example, if a user responds with "What is Alpha Company?", the third pattern is to search for an explanation about "Alpha Company" that is included in an uncommunicated unit that comes after the unit (phrase, etc.) that was being played at the time of the response.

[0085] When selecting, candidate units are judged based on the playback history to determine whether they have already been played, and units that have already been played are selected as a non-priority unit only if there are no other candidates. Note that units that have already been played may be excluded from the selection, and this is also considered a type of non-priority selection. On the other hand, if the user 3's response is something that requests a unit that has already been played, such as "What was that again?" or "Tell me about that again," then it may be selected.

[0086] If no search results are found in the above search, an external information database or the like may be searched as an information source other than the scenario information 114 .

[0087] The guide robot 2 rewrites the selected phrases and presentation order of the scenario information 114 based on the selected unit or based on search results outside the scenario information 114. The rewriting of the selected phrases and presentation order is performed, for example, so as not to transmit the same content as the selected unit redundantly, and so as to optimize the transmission order of the selected unit and the content that follows it. The rewriting of the selected phrases and presentation order may also be performed by canceling the selection according to priority so as not to exceed the upper limit of the transmission time.

[0088] The above responses specifically include, for example, the following reactions and responses:

[0089] If the response does not include a predicate, such as "Who?", the subject of the phrase closest to the currently played utterance is used as the answer. If the response refers to a noun, such as "Who is that?" or "What is that?", the answer is determined taking into account the low frequency of occurrence and the named entity class.

[0090] Furthermore, if the response is a question about a named entity such as "when" or "who," the system responds as an answer based on the results of extracting the named entity. If the response is a "Why?" type question such as "Why?", the system focuses on clue words such as "because" or "for" to extract an expression of the reason from the scenario information 114 and responds as an answer. Furthermore, discourse structure analysis may be used to identify discourse relationships (such as "cause" and "result") between distant sentences, and the system may respond to why-type questions based on the results.

[0091] Furthermore, if the response is a question about a definition, such as "What kind of company is Company α?", the system responds with a definition or trivia by referring to the scenario information 114. If the response includes information not included in the scenario information 114, the system may respond by searching an external information database.

[0092] Furthermore, if the response requires a determination of true intent, such as "Does that mean...?", an answer is generated and responded based on the results of entailment contradiction recognition. For example, if an answer can be given by matching based on the predicate-argument structure, an answer such as "That's right, it's..." or "No, it's not, it's..." is generated and responded. If an answer cannot be given by matching based on the predicate-argument structure, the coverage rate of the content words in the response to the target sentence is calculated, and if it is equal to or greater than a predetermined first value, the answer is "Yes," if it is smaller than the first value and equal to or greater than a predetermined second value, the answer is "Yes, maybe," and if it is smaller than the second value, the answer is "Hmm, I wonder."

[0093] Also, if the reaction is a question of degree, such as "what + (numeral particle)", an expression matching "(numeral particle) + (numeral particle)" is searched for in the scenario information 114 and used as the response.

[0094] The information processing device 1 and the guide robot 2 respond as described above, and when all the articles have been reproduced (S18; Yes), the guide ends.

[0095] (Effects of the embodiment) According to the above-described embodiment, the time required to guide the user 3 around the exhibits 4a to 4d is estimated as the time constraint, content information 111 relating to the exhibits 4a to 4d that can be guided around within the time constraint is selected based on the interest level information 113 of the user 3, and scenario information 114 is generated based on the selected content information 111 so that the content fits within the time constraint, making it possible to estimate and provide the information and method desired by the user 3. Furthermore, because scenario information 114 is generated so as to fit within the time constraint and in accordance with the attributes of the user 3, the necessary information can be communicated efficiently.

[0096] Furthermore, the guide robot 2 plays back the content based on the scenario information 114, searches for and selects units that will serve as an answer to the reaction of the user 3 from the scenario information 114, and responds with an answer using the selected units, so that the information desired by the user can be provided as needed.

[0097] Furthermore, the time limit and content selection are reviewed based on the reaction information 115, and the contents of the scenario information 114 are also changed, so that the situation is fed back and the information desired by the user can be provided within an appropriate time limit.

[0098] [Other embodiments] The present invention is not limited to the above-described embodiment, and various modifications are possible without departing from the spirit of the present invention.

[0099] In the above embodiment, the functions of the means 100 to 106 of the control unit 10 are realized by a program, but all or part of the means may be realized by hardware such as an ASIC. The program used in the above embodiment may also be provided by storing it on a recording medium such as a CD-ROM. The steps described in the above embodiment may be replaced, deleted, or added without departing from the spirit and scope of the present invention.

[0100] Furthermore, the functions of the means 100 to 106 of the control unit 10 of the information processing device 1 do not necessarily have to be realized on the information processing device 1, and may be realized on the guide robot 2 within the scope of the present invention. Similarly, the functions of the guide robot 2 do not necessarily have to be realized on the guide robot 2, and may be realized on the information processing device 1 within the scope of the present invention. [Explanation of symbols]

[0101] 1: Information processing equipment 2: Guide robot 3: User 4: Network 4a~4d:Exhibits 10: Control section 11: Storage section 12: Communications Department 100: Constraint time estimation means 101: Interest level estimation means 102: Content selection means 103: Content presentation order determining means 104: Scenario information generation means 105: Output means 106: Response acquisition means 110: Information Processing Program 111: Content information 112: User information 113: Interest level information 114: Scenario Information 115:Reaction Information 202: Scenario information generation means

Claims

1. Computer, a time constraint estimation means for estimating a time constraint for playing the article based on at least one of the content of the article, information on a user who listens to the audio of the article being played, and a space targeted by the article; an interest level estimation means for estimating a degree of interest in the article and sentences contained in the article based on the user information; an article selection means for selecting one or more articles from the plurality of articles based on the time limit and the interest level of the user; an information processing program that functions as a generation means for generating scenario information using sentences from one or more articles selected by the article selection means, based on the time limit and the interest level of the user;

2. 2. The information processing program according to claim 1, further functioning as article presentation order determining means for determining an order in which the articles selected by the article selecting means are to be presented.

3. the interest level estimation means changes the estimation result of the interest level based on the user's reaction to the reproduction of the scenario information; 3. The information processing program according to claim 1, wherein the article selection means changes the selected content of one or more articles based on the estimation result.

4. 4. The information processing program according to claim 3, wherein the generation means regenerates the scenario information based on the user's reaction to the playback of the scenario information and / or based on the changed interest level and the changed selection content.

5. An information processing program according to any one of claims 1 to 4, wherein the generation means generates the scenario information by selecting one or more sentences to be played to the user from a plurality of sentences contained in the article and one or more sentences to respond to the user's reaction.

6. a time constraint estimation means for estimating a time constraint for playing the article based on at least one of the content of the article, information on a user who listens to the audio of the article being played, and a space targeted by the article; an interest level estimation means for estimating a degree of interest in the article and sentences contained in the article based on the user information; an article selection means for selecting one or more articles from the plurality of articles based on the time limit and the interest level of the user; a generating means for generating scenario information based on the time limit and the interest level of the user, using sentences from one or more articles selected by the article selecting means.

Images