WO2017212689A1 - Responding device, method for controlling responding device, and control program - Google Patents

Responding device, method for controlling responding device, and control program Download PDF

Info

Publication number
WO2017212689A1
WO2017212689A1 PCT/JP2017/006333 JP2017006333W WO2017212689A1 WO 2017212689 A1 WO2017212689 A1 WO 2017212689A1 JP 2017006333 W JP2017006333 W JP 2017006333W WO 2017212689 A1 WO2017212689 A1 WO 2017212689A1
Authority
WO
WIPO (PCT)
Prior art keywords
phrase
word
response message
input
concept
Prior art date
Application number
PCT/JP2017/006333
Other languages
French (fr)
Japanese (ja)
Inventor
慎哉 佐藤
和典 森下
弘康 伊神
直起 江角
Original Assignee
シャープ株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by シャープ株式会社 filed Critical シャープ株式会社
Priority to JP2018522313A priority Critical patent/JP6644141B2/en
Priority to CN201780035585.2A priority patent/CN109313899A/en
Publication of WO2017212689A1 publication Critical patent/WO2017212689A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/16—Sound input; Sound output
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/40—Processing or translation of natural language
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L13/00—Speech synthesis; Text to speech systems
    • G10L13/08—Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G10L15/10—Speech classification or search using distance or distortion measures between unknown speech and reference templates
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue

Definitions

  • the present invention relates to a response device that recognizes input contents and responds based on the recognized contents.
  • Patent Document 1 discloses a conversation scenario editing apparatus that generates and edits a conversation scenario used in an automatic conversation apparatus that is an apparatus that can establish a conversation with a user.
  • the conventional technology as described above is configured to respond to the recognized result based on a conversation scenario prepared in advance, but the number of scenarios is limited. For this reason, when a conversation scenario that matches the recognized result cannot be found, only the contents defined in advance in ⁇ Others> can respond. Furthermore, if the number of scenarios is small, the response contents are similar and there is a problem that the conversation becomes boring.
  • the present invention has been made in view of the above problems, and its purpose is to respond to a response message based on a phrase included in the input content even when a conversation scenario that matches the input content cannot be found. It is to implement a response device that generates.
  • a response device is a response device that outputs a response message for an input content input by a user, and the first phrase and the first related phrase are related to each other.
  • a storage unit that stores that the second phrase and the second related phrase are associated with each other, and a response message generation unit that generates a response message for the input content,
  • the first phrase or the second phrase is included in the input content, and the storage unit is configured such that the second phrase and the first related phrase are related to each other, or the first phrase and the second phrase
  • the response message generating unit When it is stored that the second word / phrase belongs to the same concept, the response message generating unit generates the response message having a content in which the first word / phrase and the second related word / phrase are associated with each other.
  • a control method for a response device is a control method for a response device that outputs a response message with respect to input content input by a user, wherein the response device includes: A storage unit for storing that the first word and the first related word are related to each other and that the second word and the second related word are related to each other; A search step of searching the storage unit for storage relating to the first phrase or the second phrase included, and a response message generation step of generating a response message for the input content, the first phrase or the second A phrase is included in the input content, and the storage unit determines that the second phrase and the first related phrase are related to each other, or that the first phrase and the second phrase are If Flip stores that belong to the concept, in the response message generation step, generating a response message content associated with each other and the first word and the second related phrases.
  • Embodiment 1 Hereinafter, embodiments of the present invention will be described in detail with reference to FIGS.
  • FIG. 1 is a block diagram illustrating an example of a main configuration of a robot 1 according to this embodiment.
  • the robot 1 includes a voice input unit 11, a voice output unit 12, a storage unit 13, and a control unit 20.
  • the voice input unit 11 detects a user's utterance.
  • the sound input unit 11 may be a sound collecting device such as a microphone.
  • the voice input unit 11 sends the detected user utterance as voice data to a voice recognition unit 21 described later.
  • the voice input unit 11 specifies one utterance (an utterance that is a single sentence or sentence) from the user's utterance (time during which no voice is uttered), and the voice for each utterance. It is desirable to transmit the data to the voice recognition unit 21.
  • the audio output unit 12 functions as an output unit that outputs audio data received from the audio synthesis unit 27 described later as audio. Specifically, the audio output unit 12 is realized by a speaker or the like provided in the robot 1. In the example of FIG. 1, the audio output unit 12 is built in the robot 1, but the audio output unit 12 may be an external device attached to the robot 1.
  • the storage unit 13 stores various data handled by the robot 1.
  • the storage unit 13 according to the present embodiment stores at least word / phrase information, input / output history, and a response message template to be described later.
  • FIG. 2 is an example of word / phrase information stored in the storage unit 13 by the robot 1 according to the present embodiment
  • FIG. 3 is an input / output history stored in the storage unit 13 by the robot 1 according to the present embodiment. It is an example.
  • FIG. 4 shows an example of a response message template stored in the storage unit 13 by the robot 1 according to the first embodiment of the present invention and an example sentence using the template.
  • the phrase information stored in the storage unit 13 manages the attribute to which the phrase belongs and the relationship with other phrases, and includes the phrase, concept, and related phrase items. , Part of speech, classification, usage status.
  • the phrase indicates not only nouns such as “apple” and “Osaka”, but also adjectives such as “sweet” and “round”, and a verb having the meaning of “buy”. Words such as particles that do not make sense only with the words are not included.
  • the part of speech indicates the type of part of speech to which each phrase belongs, such as “noun”, “adjective”, “verb”, and the like.
  • the classification indicates a broader category to which each word belongs, such as “fruit” and “domestic region”.
  • the usage status indicates a status in which each word / phrase such as “What” or “Where” should be used. That is, the response message generation unit 26 can determine that a plurality of words / phrases belong to the same concept when the combinations of the concepts (parts of speech, classification, usage status) are the same.
  • a related phrase indicates, for each phrase, another phrase that is associated with the phrase and that can be combined with the phrase and / or a template of the response message to form a response message. That is, the memory
  • the input / output history provided in the storage unit 13 manages history information related to input / output of the robot 1 and includes items of number, date / time, input / output, contents, and words used.
  • the number manages the input contents of the user and the output contents of the robot 1 in chronological order, and the date and time records the date and time when the user or the robot 1 inputs / outputs.
  • the input / output indicates whether the history information is from the user or the robot 1.
  • “input” indicates a user input
  • “output” indicates a robot output.
  • the contents indicate input / output contents by the user or the robot 1.
  • the used phrase indicates a phrase included in the input / output contents by the user or the robot 1. In the illustrated example, up to three words 1 to 3 are used, but the present invention is not limited to this.
  • FIG. 4 shows a response message template provided in the storage unit 13, an outline thereof, and an example sentence using the template.
  • the storage unit 13 only needs to store at least the template item, and may not store the outline and example sentence items.
  • the template is a template of a sentence that is selected by the robot 1 and composes a response message in combination with a phrase.
  • positions where the phrase is incorporated are described as [A] and [B]. For example, in the first template, “[A] is [B] is what?”, A response message is formed by replacing [A] and [B] with words.
  • the outline indicates the content of the response message corresponding to the template.
  • the outline of the template “[A] is [B] is why?” Is “jokes”.
  • the example sentence shows an example of a response message when words / phrases acquired from user input contents are combined with each template.
  • the phrase replaced with [A] is “Kinkakuji”, and the phrase replaced with [B] is “sweet”.
  • the template includes information on the relationship between the word / phrase replaced by [A] and the word / phrase replaced by [B]. For example, the relevance is expressed by word order or particle.
  • the robot 1 may generate a message “I wonder how I heard” described at the end of the figure as a response message.
  • the control unit 20 controls and controls each part of the robot 1.
  • the voice recognition unit 21 performs voice recognition on the voice data of one utterance received from the voice input unit 11.
  • speech recognition refers to a process of obtaining text data indicating speech content (input content) from speech data of speech.
  • the speech recognition method of the speech recognition unit 21 is not particularly limited, and speech recognition may be performed using any conventional method.
  • the phrase / concept acquisition unit 22 divides the text data into words and acquires the concept of each divided phrase from the storage unit 13.
  • the phrase / concept acquisition unit 22 sets a first phrase (or second phrase) and a first related phrase (or second related phrase), which will be described later, from the divided phrases.
  • the phrase / concept acquisition unit 22 may newly register the phrase in the phrase information when each segmented phrase does not exist in the phrase information of FIG. Then, when one of the separated words is not registered as a related word of another word, the word may be registered as a related word of another word.
  • the scenario unit 23 acquires scenario information corresponding to the phrase and the phrase concept acquired by the phrase / concept acquisition unit 22 among the scenario information stored in the storage unit 13.
  • the concept determination unit 24 determines whether the phrase concept acquired by the phrase / concept acquisition unit 22 is the same as the concept of the phrase included in the output content immediately before the robot 1 acquired from the storage unit 13. judge. That is, the concept determination unit 24 determines whether or not the content input by the user is content corresponding to the response message output immediately before being input.
  • the related phrase acquisition unit 25 determines the related phrase that is related to the phrase and is different from the phrase based on the phrase and the concept of the phrase acquired by the phrase / concept acquisition unit 22. Get from. Moreover, the related phrase acquisition part 25 sets the acquired related phrase as a phrase to be used for inference to be described later.
  • the condition for the related phrase acquisition unit 25 to acquire the related phrase may be any. For example, for a specific word / phrase, all the words / phrases registered as related words / phrases in the word / phrase information of FIG. 2 may be acquired, or only words / phrases that are not included in the input content may be acquired. You may preferentially acquire a phrase whose “part of speech” is “noun”.
  • the response message generation unit 26 generates text data that is a response message output by the robot 1 according to the input content of the user, based on the instruction content from each unit. Specifically, at least one of the phrase acquired by the phrase / concept acquisition unit 22 and the related phrase acquired by the related phrase acquisition unit 25, the template stored in the storage unit 13, history information, and the scenario unit 23 are acquired. Text data is generated by combining any of the scenario information as necessary.
  • the response message generation unit 26 may be configured to extract a specific related phrase according to a predetermined condition.
  • the response message generation unit 26 is configured to preferentially extract related words belonging to the same concept as the word acquired by the word / concept acquisition unit 22 and generate a response message having contents associated with these words. May be.
  • the response message generator 26 may have any configuration as long as it generates text data to be a response message.
  • the response message generation unit 26 is based on context information including a concept to which each phrase belongs and a phrase before and after each phrase for a plurality of phrases included in the input content of the user. And a configuration for selecting the input words.
  • the context information is information indicating whether a phrase is classified as a subject, predicate, or other in the input content.
  • the response message generation unit 26 may be configured to preferentially select a phrase whose concept (classification) belongs to a proper noun and a phrase that is a predicate in the input content as an input phrase.
  • the voice synthesizer 27 converts the text data of the response message received from the response message generator 26 into voice data.
  • the voice synthesizer 27 outputs the converted voice data to the voice output unit 12.
  • FIG. 5 shows an image in which the robot 1 according to the present embodiment generates a response message by inference. In the following description, it is assumed that the user inputs an utterance to the robot 1 that “apple is sweet”.
  • the robot 1 converts the utterance into text data by voice recognition, and divides the text data into words. That is, if “apple is sweet”, it is divided into three phrases “apple”, “ha”, and “sweet”. At this time, the robot 1 may register or update the divided words in the word information in FIG.
  • the robot 1 selects a word / phrase as a starting point of inference from the divided words and sets it as the first word / phrase (or second word / phrase).
  • the first phrase (or second phrase) is an input phrase that is a phrase that characterizes the user's input content.
  • the first word (or second word) may be set to any word as long as it is a word having meaning in the same manner as the word information shown in FIG. It is preferable to set. For example, when a phrase that is a predicate includes a noun, this is set as a first word (or second word), and when it does not, a noun that is a subject is set as a first word (or second word) It may be.
  • an appropriate word / phrase may be set as the first word / phrase (or the second word / phrase).
  • the predicate “sweet” is an adjective, and the subject “apple” is a noun. "Is set to the first word (or second word).
  • the first related phrase (or the second phrase) that is recognized as being related to the first phrase (or the second phrase) is used.
  • “sweet” is set as the first related phrase (or second related phrase).
  • the robot 1 After setting the first word (or second word) and the first related word (or second related word), the robot 1 determines the first related word (or second related word) according to the word information of FIG.
  • the second phrase (or first phrase), which is a related phrase, is acquired and set.
  • the related words of “sweet” that are the first related words (or the second related words) are “apple” and “melon”, and “apple” is the first word (or second word). (Phrase) has been set. Therefore, here, “melon” is set as the second word (or first word).
  • the robot 1 sets the second related word (or first related word) that is a related word of the second word (or first word) according to the word information of FIG. ) Get and set.
  • the concept (usage status) of the second related phrase (or the first related phrase) is “How (how)”. This is because if the concept of the second related phrase (or first related phrase) is not “How”, it often does not become a natural sentence when combined with the second phrase (or first phrase). Because.
  • the related phrases of “melon” that is the second phrase (or the first phrase) are “sweet”, “round”, and “buy”.
  • the robot 1 sets “round” as the second related phrase (or first related phrase).
  • the robot 1 obtains the first word (or second word) and the second related word (or first related word) acquired by the processing so far as in S23 of FIG. Response message is generated from the phrase).
  • the phrase information in FIG. 2 also stores phrases such as particles that do not make sense only by the phrase. Specifically, the phrase information in FIG. 2 stores the phrase “ha”.
  • the combination of the concept of the word “ha” is (particle, ⁇ , ⁇ ), and the related words are “apple”, “melon”, “sweet”, and “round”. It is.
  • the first word (or second word) is “apple” and the second related word (or first related word) is “round”.
  • the second word (or first word) is “melon”.
  • the robot 1 uses “apple” and “round”, and further generates a response message “apple round” by combining with the particle “ha” having both “apple” and “melon” as related words. In this way, the robot 1 can generate a response message by inference based on the phrase included in the input content.
  • the configuration for generating the response message is not limited to the above content. Another example of a configuration for generating a response message will be described below.
  • the history information includes information on how the second word (or first word) and the second related word (or first related word) are related. For example, the relevance is expressed by word order or particle. According to the above-described example, history information whose “content” is “the melon is round” is acquired.
  • the robot 1 replaces the second word (or first word) included in the “content” of the history information with the first word (or second word) to thereby obtain a response message. Is generated. In other words, in the above example, “apple is round” is generated as a response message.
  • the robot 1 can appropriately determine the word order of each word / phrase in the response message and appropriately combine words / phrases such as particles that do not make sense.
  • storage part 13 may memorize
  • the robot 1 respondse message generator 26
  • the first related word or second related word
  • the response message may be generated from the template of FIG.
  • the template is “what else is there when saying [A]”
  • “sweet” combined with “sweet” which is the first related phrase (or second related phrase) “What else is there?” Is generated as a response message.
  • the input content of the user with respect to the response message includes the second word (or first word) associated with the first related word (or second related word).
  • the robot 1 registers the word / phrase acquired from the input content of the user with respect to the response message in the word / phrase information of FIG. 2 as the related word / phrase of the first related word / phrase (or the second related word / phrase) used in the response message. There may be. According to said structure, since it becomes possible to acquire a related phrase from the user's input content with respect to a response message, the convenience of the robot 1 can be improved.
  • the robot 1 when the input content of the user ignores the flow of dialogue, the robot 1 generates another response message from the first word and the template of FIG. 4 as in S26 of FIG. 7 described later. It may be. At this time, the robot 1 may have any configuration as long as it can compare the input content of the user and the flow of dialogue. For example, the robot 1 (concept determination unit 24) acquires the content output by the robot 1 immediately before the user's input from the input / output history of FIG. 3, and sets the word / phrase included in the history as the third word / phrase. At this time, the robot 1 determines whether or not the concept to which the third phrase belongs and the concept to which the first phrase (or the second phrase) belongs are the same using the phrase information shown in FIG.
  • the robot 1 may generate another response message in which, for example, the first word (or second word) is combined with the template related to the third word.
  • the robot 1 utters “Apple is sweet” and the user utters “Kyoto is Speaking of Kinkakuji”, the concept (classification) of “Kinkakuji”, which is the first (or second) phrase, is “ The third phrase “apple” is “fruit” and the concept (classification) is different, while “domestic region”.
  • the robot 1 uses the first word (or second word) “Kinkakuji” as the template “[A] related to“ fruit (food) ”which is the concept (classification) of the third word“ apple ”. Produces another response message combined with "I can't eat [A]". That is, according to the example sentence, either “Kinkakuji is not food” or “I can't eat Kinkakuji” is output as a response message.
  • the method of generating another response message is not limited to the above.
  • the robot 1 sets a first phrase (or second phrase) and a fourth phrase as a fourth phrase that is different from the first phrase (or second phrase) included in the input content of the user.
  • Another response message may be generated in combination with the template. If the user utters “Kinkakuji is sweet” after the robot 1 utters “Apple is sweet”, the concept (classification) of the first phrase (or second phrase) “Kinkakuji” and the third phrase “apple” Although they are different, the phrase “sweet” is common.
  • the robot 1 sets the phrase “sweet”, which is different from the first phrase (or the second phrase) “Kinkakuji”, included in the input content of the user as the fourth phrase. Then, the robot 1 uses the first phrase (or the second phrase) and the fourth phrase, the template “[A] is [B], what is it?”, “[A] is [B] Combine this with "Take me to eat this time” and "[A] is [B]” to generate another response message. That is, according to the example sentence, "Kinkakuji is sweet” Why is it, "Kinkakuji is sweet, please take me next time” and “Kinkakuji is sweet” is output .
  • the inference is not limited to the above-described method, and the second related phrase (or the first related phrase) necessary for generating the response message in combination with the first phrase (or the second phrase) and the template. Any method may be used as long as it can be acquired. Another example of inference will be described with reference to FIG. FIG. 6 shows another image in which the robot 1 according to the present embodiment generates a response message by inference.
  • (A) in FIG. 6 shows two sentences “apple is sweet” and “melon is round”.
  • One of the two sentences is input contents, and the other is past input / output contents stored in the storage unit 13.
  • the robot 1 accepts the user's input content “apple is sweet” and sets “apple” as the first phrase. Thereafter, the robot 1 has a combination of the first phrase “apple” and the concept (part of speech, classification, usage) (noun, fruit, What (what)), and the second phrase “melon” that is the same concept.
  • the history information including “melon is round” is searched from the storage unit, and the second word “melon” is acquired. Since the first phrase “apple” and the second phrase “melon” have the same concept, in this case, as shown in FIG. 6B, the robot 1 is a map in which two sentences are grouped with the same concept. Is generated. In the region where (noun, fruit, What (what)) is set, either “apple” or “melon” is set.
  • the above inference may be configured to be used in combination with generation of a response message based on a conversation scenario which is a conventional technique.
  • the storage unit 13 further stores scenario information.
  • the robot 1 uses the first word (or second word) included in the input content as in S25 of FIG. ) And a conversation scenario may be generated from the conversation scenario.
  • scenario information does not exist in the memory
  • FIG. 7 is a flowchart illustrating an example of a flow of processing executed by the robot 1 of the present embodiment.
  • the user's input content includes the first word and the first related word
  • the response message output by the robot 1 immediately before the user inputs includes the third word.
  • the voice recognition unit 21 performs voice recognition of the input content for the user input received by the voice input unit 11, and generates text data (S11).
  • the phrase / concept acquisition unit 22 receives the text data generated in S11, and from the text data and the phrase information stored in the storage unit 13, the first phrase, the first related phrase, and the first phrase A concept is acquired (S12).
  • the scenario unit 23 searches the storage unit 13 for scenario information corresponding to the first phrase or the concept of the first phrase acquired in S12 (S13), and determines whether the corresponding scenario information has been found. (S14).
  • the concept determination unit 24 determines whether or not the robot 1 has responded (spoken) immediately before the user input (S15). On the other hand, if it is determined that the corresponding scenario information has been found (YES in S14), the process proceeds to S25 described later.
  • the concept determination unit 24 determines whether the concepts of the first phrase and the third phrase are the same (S16).
  • the related phrase acquisition unit 25 searches for the second phrase from the related phrases of the first related phrase in the phrase information stored in the storage unit 13 ( S17: Search step).
  • the process directly proceeds to S17.
  • the related phrase acquisition unit 25 determines whether or not the second phrase has been found (whether or not the related phrase of the first related phrase is in the phrase information) (S18). When it is determined that the second phrase has not been found (NO in S18), the related phrase acquisition unit 25 searches the phrase information stored in the storage unit 13 for the second phrase having the same concept as the first phrase (S19). ). And the related phrase acquisition part 25 determines whether the 2nd phrase was found in the process of S19 (S20).
  • the related word / phrase acquisition unit 25 uses the second word / phrase information in the storage unit 13 as the second word / phrase information.
  • a second related phrase is retrieved from the related phrases of the phrase (S21).
  • the related phrase acquisition unit 25 determines whether or not the second related phrase is found in the process of S21 (whether or not the related phrase of the second phrase is in the phrase information) (S22). If it is determined that the second related phrase has been found (YES in S22), the response message generator 26 generates a response message from the first phrase and the second related phrase (S23: response message generation step), and the process proceeds to S27. move on. In this way, the response message generator 26 generates a response message having the content in which the first word / phrase and the second related word / phrase are associated with each other.
  • the response message generator 26 when the second word / phrase is not found in the process of S20 (NO in S20), or when the second related word / phrase is not found in the process of S22 (NO in S22), the response message generator 26 generates the first related word A response message is generated from the phrase and the template (S24). Thereafter, the process proceeds to S27. If it is determined in S14 that the corresponding scenario information has been found (YES in S14), the response message generator 26 generates a response message from the first phrase and scenario information (S25), and the process proceeds to S27. If it is determined in S16 that the concepts of the first phrase and the third phrase are not the same (NO in S16), the response message generator 26 generates a response message from the first phrase and the template (S26), and processing Advances to S27.
  • the voice synthesis unit 27 converts the response message into voice data, and outputs the voice via the voice output unit 12 (S27).
  • the robot 1 can generate a response message based on the words / phrases included in the input content even when no suitable conversation scenario is found.
  • the robot 1 according to the present embodiment differs from the previous embodiment in that a response message including three or more words / phrases is generated when the input content of the user includes three or more words / phrases.
  • members having the same functions as those described in the embodiment are given the same reference numerals, and descriptions thereof are omitted.
  • the first phrase (or second phrase) and the first related phrase (or second related phrase) in the same manner as in the above embodiment, regarding three or more words included in the input content of the user. ) Is set.
  • the phrase / concept acquisition unit 22 is a related phrase of the first related phrase (or second related phrase), and is different from the first phrase (or second phrase) with respect to the first accompanying phrase (or 2nd accompanying phrase).
  • any criteria may be used for setting the first accompanying phrase (or second accompanying phrase) from the input content. For example, after the first word / phrase is set in the same manner as in the above embodiment, among the remaining words / phrases, a word / phrase whose concept (part of speech) is “part of speech” and whose concept (classification) is other than “other” is first attached. You may set to a phrase (or 2nd accompanying phrase).
  • the order of setting the first related phrase (or second related phrase) and the first accompanying phrase (or second accompanying phrase) is not particularly limited, but the phrase whose concept (part of speech) is “noun” It is preferable to set in preference to one accompanying phrase (or second accompanying phrase). Then, a response message is generated in which the first phrase (or second phrase) and the first accompanying phrase (or second accompanying phrase) are associated with the second related phrase (or first related phrase) acquired by inference. .
  • FIG. 8 and FIG. 9 are diagrams illustrating an image in which the robot 1 according to the present embodiment generates a response message by inference.
  • FIG. 8 shows an image in which reasoning is applied to two sentences, “Osaka's famous okonomiyaki” and “Hiroshima is delicious oysters”.
  • “Hiroshima is delicious oysters” is stored in the storage unit 13 in advance as past input / output, and “Osaka specialty okonomiyaki” is input by the user.
  • the robot 1 the phrase / concept acquisition unit 22 sets “okonomiyaki” as the first phrase (or second phrase), “specialty” as the first related phrase (or second related phrase), and “Osaka”. Set to the first accompanying phrase (or second accompanying phrase).
  • the robot 1 (related phrase acquisition unit 25) includes the phrase “oyster” having the same combination of the first phrase (or second phrase) “okonomiyaki” and the concept (part of speech, classification, usage).
  • the history information “Hiroshima is oyster is delicious” including the phrase “Hiroshima” having the same concept as the accompanying phrase (or second accompanying phrase) “Osaka” is acquired from the input / output history of FIG.
  • “okonomiyaki” and “oyster” are different in terms of “food / okonomiyaki / takoyaki” and “food / seafood”, but the parts “food” are the same.
  • the first accompanying word (or second accompanying word) “Osaka” is set in combination with the first word (or second word) “okonomiyaki”, and similarly, “oyster” is set in combination with “Hiroshima”.
  • FIG. 9 shows an image in which reasoning is applied to two sentences, “Osaka's specialty is Okonomiyaki” and “Hiroshima souvenir is Momiji Wharf”.
  • “Hiroshima souvenir is Momiji Wharf” is previously stored in the storage unit 13 as past input / output, and “Osaka specialty okonomiyaki” is input by the user.
  • the robot 1 the phrase / concept acquisition unit 22
  • Related words) and “Osaka” are set as the first accompanying words (or second accompanying words).
  • the robot 1 (related phrase acquisition unit 25) includes the phrase “Momiji bun” having the same combination of the first phrase (or second phrase) “okonomiyaki” and the concept (part of speech, classification, usage).
  • the robot 1 (response message generator 26) generates a map in which two sentences are grouped together with words having the same concept as shown in FIG. 9B.
  • the robot 1 can acquire four sentences, “Osaka's specialty is okonomiyaki”, “Osaka's souvenir is okonomiyaki”, “Hiroshima's specialty is okonomiyaki”, and “Hiroshima's souvenir is okonomiyaki”.
  • the robot 1 can acquire “Osaka souvenir is okonomiyaki” and “Hiroshima specialties are oysters” as a result of inference, and can generate a response message.
  • FIG. 7 is a flowchart illustrating an example of a flow of processing executed by the robot 1 of the present embodiment.
  • the user's input content includes the first word and the first related word
  • the response message output by the robot 1 immediately before the user inputs includes the third word.
  • the input content of the user is “Osaka specialty is Okonomiyaki”
  • the first word and the first related phrase are “okonomiyaki” and “specialty”, respectively, and “Osaka” is the first accompanying phrase. is there.
  • the processing of S11 to S22 is the same as that of the first embodiment. If YES in S22, the response message generator 26 generates a response message from the first phrase, the first accompanying phrase, and the second related phrase (S23), and the process proceeds to S27. Specifically, the response message generator 26 generates a response message “Osaka souvenir is okonomiyaki” or “Hiroshima specialties are oysters” in which the first phrase, the first accompanying phrase, and the second related phrase are associated with each other. .
  • the response message generating unit 26 selects the first related word A response message is generated from the phrase and the template (S24). Thereafter, the process proceeds to S27. If it is determined in S14 that the corresponding scenario information has been found (YES in S14), the response message generating unit 26 generates a response message from the first word / phrase, the first accompanying word / phrase, and the scenario information (S25). Advances to S27.
  • the response message generator 26 If it is determined in S16 that the concepts of the first word and the third word are not the same (NO in S16), the response message generator 26 generates a response message from the first word, the first accompanying word, and the template. Then (S26), the process proceeds to S27.
  • the process of S27 is the same as that of the first embodiment.
  • the robot 1 is based on the words / phrases included in the input contents even when a conversation scenario that matches when the user's input contents include three or more words / phrases is not found.
  • a response message can be generated.
  • control unit 20 may be provided in a response device that interacts with a user using text. Further, voice data or text interaction between the user and the response device may be performed via a network.
  • the control unit 20 (particularly the response message generation unit 26 and the concept determination unit 24) of the robot 1 may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or a CPU (Central Processing). Unit) and may be realized by software.
  • a logic circuit hardware
  • IC chip integrated circuit
  • CPU Central Processing
  • the control unit 20 includes a CPU that executes instructions of a program that is software that realizes each function, a ROM (Read Only Memory) in which the program and various data are recorded so as to be readable by a computer (or CPU), or A storage device (these are referred to as “recording media”), a RAM (Random Access Memory) for expanding the program, and the like are provided.
  • a computer or CPU
  • the recording medium a “non-temporary tangible medium” such as a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used.
  • the program may be supplied to the computer via an arbitrary transmission medium (such as a communication network or a broadcast wave) that can transmit the program.
  • a transmission medium such as a communication network or a broadcast wave
  • the present invention can also be realized in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.
  • the response device (1) is a response device (1) that outputs a response message for the input content input by the user, and the first phrase and the first related phrase are related to each other. And a storage unit (13) that stores that the second word and the second related word are related to each other, and a response message generation unit (26) that generates a response message for the input content.
  • the first phrase or the second phrase is included in the input content, and the storage unit (13) has the second phrase and the first related phrase related to each other, or
  • the response message generator (26) has a content in which the first word and the second related word are associated with each other. The above response message Generated.
  • the response device includes a first related phrase or phrase that is associated with the first phrase or the second phrase included in the input content input by the user, and the first phrase stored in the storage unit, and Based on a second related phrase that is related to the second phrase, the second phrase and the first related phrase are related to each other, or the first phrase and the second phrase belong to the same concept Is stored, it is possible to generate a response message having a content in which the first phrase and the second related phrase are associated with each other. Therefore, even when a suitable conversation scenario is not found, there is an effect of realizing a response device that generates a response message based on a phrase included in input content.
  • the storage unit (13) determines how the second phrase and the second related phrase are related with each other.
  • the response message generator (26) stores the response message having the content in which the first word / phrase and the second related word / phrase are associated with each other.
  • a response apparatus produces
  • the storage unit (13) has the second phrase and the first related phrase related to each other.
  • the response message generator (26) generates the response message having a content in which the first word / phrase and the second related word / phrase are associated with each other.
  • the responding device when the storage unit stores that the second word and the first related word are related to each other, the responding device associates the first word and the second related word with each other. It becomes possible to generate a response message of contents. Therefore, even if a matching conversation scenario is not found, the content of the content that associates the first phrase and the second related phrase with each other based on the fact that the second phrase and the first related phrase are related to each other. There is an effect of realizing a response device that generates a response message.
  • the storage unit (13) stores that the first word and the second word belong to the same concept. If so, the response message generator (26) generates the response message having a content in which the first word / phrase and the second related word / phrase are associated with each other.
  • related the 1st phrase and the 2nd related phrase mutually A response message can be generated. Therefore, even when a matching conversation scenario is not found, a response message having contents in which the first word and the second related word are associated with each other based on the fact that the first word and the second word belong to the same concept. This produces an effect of realizing a response device that generates.
  • the storage unit (13) indicates that the second phrase and the plurality of related phrases are related to each other. If stored, the response message generator (26) preferentially extracts a related phrase belonging to the same concept as the first related phrase from the plurality of related phrases as the second related phrase. The response message having a content in which one word and the second related word are associated with each other is generated.
  • the response apparatus when the memory
  • the related word / phrase belonging to is preferentially extracted as the second related word / phrase, and the response message having the contents in which the first word / phrase and the second related word / phrase are associated with each other is generated.
  • the response device (1) according to Aspect 6 of the present invention is the response device (1) according to any one of Aspects 1 to 5, wherein the output unit (12) outputs the response message to the outside, and A concept determination unit (24) that determines whether the concept to which the third word included in the response message output from the unit (12) belongs and the concept to which the input word included in the input content belongs are the same.
  • the storage unit (13) stores a template of the response message
  • the response message generation unit (26) has the same concept to which the third phrase belongs and the concept to which the input phrase belongs. If there is, the response message having the contents associating the first phrase with the second related phrase is generated, and if the concept to which the third phrase belongs and the concept to which the input phrase belongs are different, It generates another response message using chromatography bets.
  • the responding device includes a concept to which the third phrase included in the response message output immediately before the input content is input by the response device, and a concept to which the input phrase included in the input content belongs. If the two words are the same, a response message is generated with the contents of the first word and the second related word associated with each other. When the concept to which the third word belongs and the concept to which the input word belongs are different, the response message is generated using a template. Can be generated. Therefore, even when a suitable conversation scenario is not found, there is an effect of realizing a response device that generates a response message based on the concept to which the third word / phrase belongs and the concept to which the input word / phrase belongs.
  • the response message generation unit (26) when the concept to which the third phrase belongs and the concept to which the input phrase belongs are different from each other, the response message generation unit (26) The phrase is combined with the template associated with the third phrase to generate the additional response message.
  • the response device when the concept to which the third word belongs and the concept to which the input word belongs are different from each other, the response device generates another response message by combining the input word with the template related to the third word. It becomes possible. Therefore, even when a suitable conversation scenario is not found, if the concept to which the third word belongs and the concept to which the input word belongs are different from each other, the input word is combined with a template related to the third word to obtain another response. There is an effect of realizing a response device that generates a message.
  • the response message generation unit (26) when the concept to which the third phrase belongs and the concept to which the input phrase belongs are different from each other, the response message generation unit (26) The other response message is generated by combining the phrase and the fourth phrase included in the input content with a template.
  • the response device when the concept to which the third word belongs and the concept to which the input word belongs are different from each other, the response device combines the input word and the fourth word included in the input content with the template to generate another A response message can be generated. Therefore, even if a matching conversation scenario is not found, if the concept to which the third word belongs and the concept to which the input word belongs are different, the input word and the fourth word included in the input content are used as a template. There is an effect of realizing a response device that generates another response message in combination.
  • the response message generation unit (26) includes the concept to which each phrase belongs and the phrases for the plurality of phrases included in the input content.
  • the input phrase is selected based on the context information including the phrases before and after.
  • the response device can select an input phrase based on the attribute to which each phrase belongs and the context information including the phrases before and after each phrase. Therefore, even when a matching conversation scenario is not found, the input phrase is selected from the context information including the attribute to which each phrase belongs and the phrases before and after each phrase, and combined with the template related to the third phrase, There is an effect of realizing a response device that generates a response message.
  • the control method of the response device (1) is a control method of the response device (1) that outputs a response message for the input content input by the user, and the response device (1)
  • a storage unit storing that the one word and the first related word are related to each other and the second word and the second related word are related to each other;
  • a search step (S17) for searching the storage unit for memory related to the first word or the second word, and a response message generation step (S23) for generating a response message for the input content.
  • the second word / phrase is included in the input content, and the storage unit has the second word / phrase and the first related word / phrase related to each other, or the first word / phrase and the second word / phrase.
  • DOO cases stores that belong to the same concept, in the response message generation step (S23), generates a response message content associated with each other and the first word and the second related phrases. According to said structure, there exists an effect similar to the aspect 1.
  • the response device (1) may be realized by a computer.
  • the response device is operated by causing the computer to operate as each unit (software element) included in the response device (1).
  • the control program of the response device (1) for realizing (1) by a computer and a computer-readable recording medium on which the control program is recorded also fall within the scope of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Multimedia (AREA)
  • Acoustics & Sound (AREA)
  • Artificial Intelligence (AREA)
  • Machine Translation (AREA)
  • Manipulator (AREA)
  • Toys (AREA)

Abstract

The present invention makes it possible to create a response message on the basis of phrases included in input content even if a suitable conversation scenario cannot be found. This responding device (1) includes: a storage unit (13) that stores relationships between a first phrase and a first related phrase and between a second phrase and a second related phrase; and a response message creation unit (26) that, if the first phrase or the second phrase is included in the input content, creates the response message including the content mutually associated with the first phrase and the second related phrase.

Description

応答装置および応答装置の制御方法、制御プログラムResponse device, response device control method, and control program
　本発明は入力内容を認識し、認識した内容に基づいて応答する応答装置に関する。 The present invention relates to a response device that recognizes input contents and responds based on the recognized contents.
　ユーザの発話を受けると、該発話から入力内容を認識し、認識した結果に応じて応答する応答装置が従来技術として知られている。例えば、下記の特許文献１には、ユーザとの会話を成立させることが可能な装置である自動会話装置に用いられる会話シナリオを生成及び編集する会話シナリオ編集装置が開示されている。 2. Description of the Related Art A response device that recognizes input content from a utterance and responds according to the recognized result when a user utterance is received is known as a prior art. For example, Patent Document 1 below discloses a conversation scenario editing apparatus that generates and edits a conversation scenario used in an automatic conversation apparatus that is an apparatus that can establish a conversation with a user.
日本国公開特許公報「特開２０１０－７３１９２号」（２０１０年４月２日公開）Japanese Published Patent Publication “Japanese Patent Laid-Open No. 2010-73192” (published on April 2, 2010)
　しかしながら、上述のような従来技術は、認識した結果について、予め用意していた会話シナリオに基づいて応答する構成であるが、シナリオの数には限りがある。そのため、認識した結果に適合する会話シナリオが見つからない場合は、予め＜その他＞に規定された内容しか応答できない。さらに、シナリオの数が少ないと、応答内容が似通ってしまい、退屈な会話になるという問題がある。 However, the conventional technology as described above is configured to respond to the recognized result based on a conversation scenario prepared in advance, but the number of scenarios is limited. For this reason, when a conversation scenario that matches the recognized result cannot be found, only the contents defined in advance in <Others> can respond. Furthermore, if the number of scenarios is small, the response contents are similar and there is a problem that the conversation becomes boring.
　本発明は、前記の問題点に鑑みてなされたものであり、その目的は、入力内容に適合する会話シナリオが見つからないような場合であっても、入力内容に含まれる語句に基づいて応答メッセージを生成する応答装置を実現することにある。 The present invention has been made in view of the above problems, and its purpose is to respond to a response message based on a phrase included in the input content even when a conversation scenario that matches the input content cannot be found. It is to implement a response device that generates.
　上記の課題を解決するために、本発明の一態様に係る応答装置は、ユーザが入力した入力内容について応答メッセージを出力する応答装置であって、第１語句と第１関連語句とが互いに関連していること、および、第２語句と第２関連語句とが互いに関連していることを記憶している記憶部と、上記入力内容に対する応答メッセージを生成する応答メッセージ生成部とを備え、上記第１語句または上記第２語句が上記入力内容に含まれており、上記記憶部が、上記第２語句と上記第１関連語句とが互いに関連していること、または、上記第１語句と上記第２語句とが同じ概念に属することを記憶している場合、上記応答メッセージ生成部は、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成する。 In order to solve the above-described problem, a response device according to an aspect of the present invention is a response device that outputs a response message for an input content input by a user, and the first phrase and the first related phrase are related to each other. And a storage unit that stores that the second phrase and the second related phrase are associated with each other, and a response message generation unit that generates a response message for the input content, The first phrase or the second phrase is included in the input content, and the storage unit is configured such that the second phrase and the first related phrase are related to each other, or the first phrase and the second phrase When it is stored that the second word / phrase belongs to the same concept, the response message generating unit generates the response message having a content in which the first word / phrase and the second related word / phrase are associated with each other.
　また、上記の課題を解決するために、本発明の一態様に係る応答装置の制御方法は、ユーザが入力した入力内容について応答メッセージを出力する応答装置の制御方法であって、上記応答装置は、第１語句と第１関連語句とが互いに関連していること、および、第２語句と第２関連語句とが互いに関連していることを記憶している記憶部を備え、上記入力内容に含まれている上記第１語句または上記第２語句に関する記憶を上記記憶部から探す探索ステップと、上記入力内容に対する応答メッセージを生成する応答メッセージ生成ステップとを含み、上記第１語句または上記第２語句が上記入力内容に含まれており、上記記憶部が、上記第２語句と上記第１関連語句とが互いに関連していること、または、上記第１語句と上記第２語句とが同じ概念に属することを記憶している場合、上記応答メッセージ生成ステップでは、上記第１語句と上記第２関連語句とを互いに関連付けた内容の応答メッセージを生成する。 In order to solve the above problem, a control method for a response device according to an aspect of the present invention is a control method for a response device that outputs a response message with respect to input content input by a user, wherein the response device includes: A storage unit for storing that the first word and the first related word are related to each other and that the second word and the second related word are related to each other; A search step of searching the storage unit for storage relating to the first phrase or the second phrase included, and a response message generation step of generating a response message for the input content, the first phrase or the second A phrase is included in the input content, and the storage unit determines that the second phrase and the first related phrase are related to each other, or that the first phrase and the second phrase are If Flip stores that belong to the concept, in the response message generation step, generating a response message content associated with each other and the first word and the second related phrases.
　本発明の一態様によれば、入力内容に適合する会話シナリオが見つからないような場合であっても、入力内容に含まれる語句に基づいて応答メッセージを生成する応答装置を実現するという効果を奏する。 According to one aspect of the present invention, there is an effect of realizing a response device that generates a response message based on a phrase included in an input content even when a conversation scenario that matches the input content is not found. .
本発明の実施形態１に係るロボット１の要部構成の一例を示すブロック図である。It is a block diagram which shows an example of a principal part structure of the robot 1 which concerns on Embodiment 1 of this invention. 本発明の実施形態１に係るロボット１が記憶部に備える語句情報の一例である。It is an example of the phrase information with which the robot 1 which concerns on Embodiment 1 of this invention is provided in a memory | storage part. 本発明の実施形態１に係るロボット１が記憶部に備える入出力履歴の一例を示す図である。It is a figure which shows an example of the input / output log | history with which the robot which concerns on Embodiment 1 of this invention is provided in a memory | storage part. 本発明の実施形態１に係るロボット１が記憶部に備える応答メッセージのテンプレートの一例およびテンプレートを用いた例文を示す。4 shows an example of a response message template provided in the storage unit of the robot 1 according to the first embodiment of the present invention and an example sentence using the template. 本発明の実施形態１に係るロボット１が推論によって応答メッセージを生成するイメージを示す図である。It is a figure which shows the image which the robot 1 which concerns on Embodiment 1 of this invention produces | generates a response message by inference. 本発明の実施形態１に係るロボット１が推論によって応答メッセージを生成する別のイメージを示す図である。It is a figure which shows another image which the robot 1 which concerns on Embodiment 1 of this invention produces | generates a response message by inference. 本発明の実施形態１に係るロボット１が実行する処理の一例を示すフローチャートである。It is a flowchart which shows an example of the process which the robot 1 which concerns on Embodiment 1 of this invention performs. 本発明の実施形態２に係るロボット１が推論によって応答メッセージを生成する別のイメージを示す図である。It is a figure which shows another image which the robot 1 which concerns on Embodiment 2 of this invention produces | generates a response message by inference. 本発明の実施形態２に係るロボット１が推論によって応答メッセージを生成する別のイメージを示す図である。It is a figure which shows another image which the robot 1 which concerns on Embodiment 2 of this invention produces | generates a response message by inference.
　〔実施形態１〕
　以下、本発明の実施の形態について、図１～７を用いて詳細に説明する。 Embodiment 1
Hereinafter, embodiments of the present invention will be described in detail with reference to FIGS.
　＜ロボットの構成＞
　図１に基づいて本実施形態に係るロボット１の概要を説明する。なお、本実施形態に係るロボット１は、ユーザが発話によって入力した入力内容について応答メッセージを出力する応答装置である。図１は、本実施形態に係るロボット１の要部構成の一例を示すブロック図である。図１に示すように、ロボット１は、音声入力部１１、音声出力部１２、記憶部１３、および制御部２０を備えている。音声入力部１１は、ユーザの発話を検出するものである。音声入力部１１は具体的には、マイク等の集音装置であればよい。音声入力部１１は検出したユーザの発話を音声データとして後述する音声認識部２１に送る。なお、音声入力部１１は、ユーザの発話の間（音声を発していない時間）などから１回の発話（１まとまりの文または文章となる発話）を特定し、当該１回の発話毎の音声データを音声認識部２１に送信することが望ましい。音声出力部１２は、後述する音声合成部２７から受信した音声データを音声として外部に出力する出力部として機能する。音声出力部１２は具体的にはロボット１に備えられたスピーカ等で実現される。なお、図１の例では音声出力部１２はロボット１に内蔵されているが、音声出力部１２はロボット１に取付けられた外部装置であっても構わない。記憶部１３は、ロボット１にて扱われる各種データを記憶する。本実施形態に係る記憶部１３は、後述する語句情報、入出力履歴、応答メッセージのテンプレートを少なくとも記憶している。 <Robot configuration>
An outline of the robot 1 according to the present embodiment will be described based on FIG. Note that the robot 1 according to the present embodiment is a response device that outputs a response message regarding the input content input by the user by utterance. FIG. 1 is a block diagram illustrating an example of a main configuration of a robot 1 according to this embodiment. As shown in FIG. 1, the robot 1 includes a voice input unit 11, a voice output unit 12, a storage unit 13, and a control unit 20. The voice input unit 11 detects a user's utterance. Specifically, the sound input unit 11 may be a sound collecting device such as a microphone. The voice input unit 11 sends the detected user utterance as voice data to a voice recognition unit 21 described later. Note that the voice input unit 11 specifies one utterance (an utterance that is a single sentence or sentence) from the user's utterance (time during which no voice is uttered), and the voice for each utterance. It is desirable to transmit the data to the voice recognition unit 21. The audio output unit 12 functions as an output unit that outputs audio data received from the audio synthesis unit 27 described later as audio. Specifically, the audio output unit 12 is realized by a speaker or the like provided in the robot 1. In the example of FIG. 1, the audio output unit 12 is built in the robot 1, but the audio output unit 12 may be an external device attached to the robot 1. The storage unit 13 stores various data handled by the robot 1. The storage unit 13 according to the present embodiment stores at least word / phrase information, input / output history, and a response message template to be described later.
　＜記憶部が記憶するデータの例＞
　本実施形態において記憶部１３が記憶するデータの一例を、図２～４に基づいて説明する。図２は、本実施形態に係るロボット１が記憶部１３に記憶している語句情報の一例であり、図３は、本実施形態に係るロボット１が記憶部１３に記憶している入出力履歴の一例である。図４は、本発明の実施形態１に係るロボット１が記憶部１３に記憶している応答メッセージのテンプレートの一例およびテンプレートを用いた例文を示す。 <Example of data stored in the storage unit>
An example of data stored in the storage unit 13 in the present embodiment will be described with reference to FIGS. FIG. 2 is an example of word / phrase information stored in the storage unit 13 by the robot 1 according to the present embodiment, and FIG. 3 is an input / output history stored in the storage unit 13 by the robot 1 according to the present embodiment. It is an example. FIG. 4 shows an example of a response message template stored in the storage unit 13 by the robot 1 according to the first embodiment of the present invention and an example sentence using the template.
　図２の例において、記憶部１３が記憶している語句情報は語句が属する属性および他の語句との関連性を管理するものであり、語句、概念、関連語句の項目を備え、概念はさらに、品詞、分類、利用状況を備えている。図示の例において、語句は「リンゴ」、「大阪」等の名詞だけでなく、「甘い」、「丸い」等の形容詞や、「買う」という動詞を含む、単体で意味を備える言葉を示し、助詞等その語句のみでは意味をなさない語句は含まれない。品詞は、「名詞」、「形容詞」、「動詞」等の、各語句が属する品詞の種類を示す。分類は、「果物」、「国内地域」等の、各語句が属するより広いカテゴリーを示す。利用状況は、「Ｗｈａｔ（何が）」、「Ｗｈｅｒｅ（どこが）」等の、各語句が用いられるべき状況を示す。つまり、応答メッセージ生成部２６は、複数の語句について、概念（品詞、分類、利用状況）の組み合わせが同一である場合、それらの語句は同じ概念に属すると判断できる。関連語句は、各語句について、該語句と互いに関連しており、該語句および応答メッセージのテンプレートの少なくともいずれかと組み合わせて応答メッセージを構成することができる別の語句を示す。すなわち、記憶部１３は、それぞれの語句は、関連語句を介して、語句同士が互いに関連していることを記憶している。図示の例では、「リンゴ」の関連語句は「甘い」であるため、これらを組み合わせて応答メッセージ「リンゴは甘い」を構成することができる。 In the example of FIG. 2, the phrase information stored in the storage unit 13 manages the attribute to which the phrase belongs and the relationship with other phrases, and includes the phrase, concept, and related phrase items. , Part of speech, classification, usage status. In the example shown in the figure, the phrase indicates not only nouns such as “apple” and “Osaka”, but also adjectives such as “sweet” and “round”, and a verb having the meaning of “buy”. Words such as particles that do not make sense only with the words are not included. The part of speech indicates the type of part of speech to which each phrase belongs, such as “noun”, “adjective”, “verb”, and the like. The classification indicates a broader category to which each word belongs, such as “fruit” and “domestic region”. The usage status indicates a status in which each word / phrase such as “What” or “Where” should be used. That is, the response message generation unit 26 can determine that a plurality of words / phrases belong to the same concept when the combinations of the concepts (parts of speech, classification, usage status) are the same. A related phrase indicates, for each phrase, another phrase that is associated with the phrase and that can be combined with the phrase and / or a template of the response message to form a response message. That is, the memory | storage part 13 has memorize | stored that each phrase is mutually related through the related phrase. In the illustrated example, since the related phrase of “apple” is “sweet”, these can be combined to form the response message “apple is sweet”.
　図３の例において、記憶部１３が備える入出力履歴はロボット１の入出力に関する履歴情報を管理するものであり、番号、日時、入出力、内容、使用語句の項目を備えている。番号はユーザの入力内容およびロボット１の出力内容を時系列に沿って管理するものであり、日時はユーザまたはロボット１による入出力がなされた日時を記録する。入出力は、該履歴情報がユーザまたはロボット１のいずれによるものかを示し、図示の例では「入力」がユーザの入力を、「出力」がロボットの出力をそれぞれ示す。内容は、ユーザまたはロボット１による入出力内容を示す。使用語句は、ユーザまたはロボット１による入出力内容に含まれていた語句を示す。なお、図示の例では使用語句１～３の３つまで語句を記録する構成となっているが、これに限定される必要はない。 3, the input / output history provided in the storage unit 13 manages history information related to input / output of the robot 1 and includes items of number, date / time, input / output, contents, and words used. The number manages the input contents of the user and the output contents of the robot 1 in chronological order, and the date and time records the date and time when the user or the robot 1 inputs / outputs. The input / output indicates whether the history information is from the user or the robot 1. In the example shown in the figure, “input” indicates a user input, and “output” indicates a robot output. The contents indicate input / output contents by the user or the robot 1. The used phrase indicates a phrase included in the input / output contents by the user or the robot 1. In the illustrated example, up to three words 1 to 3 are used, but the present invention is not limited to this.
　図４は、記憶部１３が備える応答メッセージのテンプレートとその概要、そしてテンプレートを用いた例文を示す。なお、図示の例において、記憶部１３は少なくともテンプレートの項目のみを記憶すればよく、概要および例文の項目は記憶しなくてもよい。テンプレートは、ロボット１が選択し、語句と組み合わせて応答メッセージを構成するための文章の雛形であり、図示の例において、語句を組み込む箇所は[A]および[B]として記載されている。例えば、１つ目のテンプレートである「[A]は[B]よね　ってなんでやねん」では、[A]および[B]がそれぞれ語句に置換されることによって、応答メッセージが構成される。概要は、テンプレートが対応する、応答メッセージの内容を示す。図示の例では、テンプレート「[A]は[B]よね　ってなんでやねん」の概要は「ジョーク」である。例文は、ユーザの入力内容から取得した語句を、それぞれのテンプレートに組み合わせた場合の応答メッセージの例を示す。なお、図示の例において、[A]に置換される語句は「金閣寺」であり、[B]に置換される語句は「甘い」である。このとき、図示の例からわかるように、テンプレートは、[A]に置換される語句と[B]に置換される語句とがどのような関連性で関連しているかという情報を含む。例えば、関連性は、語順または助詞等によって表される。また、複数のテンプレートの中に当てはまる定型文がなければ、ロボット１は図の末尾に記載されている「ボクの聞き方が悪かったのかな」という文言を応答メッセージとして生成してもよい。 FIG. 4 shows a response message template provided in the storage unit 13, an outline thereof, and an example sentence using the template. In the example shown in the figure, the storage unit 13 only needs to store at least the template item, and may not store the outline and example sentence items. The template is a template of a sentence that is selected by the robot 1 and composes a response message in combination with a phrase. In the example shown in the figure, positions where the phrase is incorporated are described as [A] and [B]. For example, in the first template, “[A] is [B] is what?”, A response message is formed by replacing [A] and [B] with words. The outline indicates the content of the response message corresponding to the template. In the example shown in the drawing, the outline of the template “[A] is [B] is why?” Is “jokes”. The example sentence shows an example of a response message when words / phrases acquired from user input contents are combined with each template. In the illustrated example, the phrase replaced with [A] is “Kinkakuji”, and the phrase replaced with [B] is “sweet”. At this time, as can be seen from the example shown in the figure, the template includes information on the relationship between the word / phrase replaced by [A] and the word / phrase replaced by [B]. For example, the relevance is expressed by word order or particle. Further, if there is no fixed sentence that fits in a plurality of templates, the robot 1 may generate a message “I wonder how I heard” described at the end of the figure as a response message.
　制御部２０は、ロボット１の各部を統括して制御するものであり、音声認識部２１、語句・概念取得部２２、シナリオ部２３、概念判定部２４、関連語句取得部２５、応答メッセージ生成部２６、および音声合成部２７を備えている。 The control unit 20 controls and controls each part of the robot 1. The speech recognition unit 21, phrase / concept acquisition unit 22, scenario unit 23, concept determination unit 24, related phrase acquisition unit 25, response message generation unit 26 and a speech synthesizer 27.
　音声認識部２１は、音声入力部１１から受信した、１回の発話の音声データについて音声認識を行う。なお、本発明において「音声認識」とは、発話の音声データから発話内容（入力内容）を示すテキストデータを得る処理を示す。音声認識部２１の音声認識の方法は特に限定されず、従来あるいずれの方法を用いて音声認識を行ってもよい。 The voice recognition unit 21 performs voice recognition on the voice data of one utterance received from the voice input unit 11. In the present invention, “speech recognition” refers to a process of obtaining text data indicating speech content (input content) from speech data of speech. The speech recognition method of the speech recognition unit 21 is not particularly limited, and speech recognition may be performed using any conventional method.
　語句・概念取得部２２は、音声認識部２１が音声認識により得たテキストデータを受け付けると、該テキストデータを語句に区切り、区切られた各語句の概念を記憶部１３より取得する。また、語句・概念取得部２２は、区切られた語句の中から後述する第１語句（または第２語句）と、第１関連語句（または第２関連語句）を設定する。さらに、語句・概念取得部２２は、区切られた各語句が、図２の語句情報の中に存在しない場合は、該語句を新たに語句情報に登録してもよい。そして、区切られた語句の１つが区切られた別の語句の関連語句として登録されていない場合は、該語句を別の語句の関連語句として登録してもよい。 When the speech recognition unit 21 receives the text data obtained by speech recognition, the phrase / concept acquisition unit 22 divides the text data into words and acquires the concept of each divided phrase from the storage unit 13. The phrase / concept acquisition unit 22 sets a first phrase (or second phrase) and a first related phrase (or second related phrase), which will be described later, from the divided phrases. Furthermore, the phrase / concept acquisition unit 22 may newly register the phrase in the phrase information when each segmented phrase does not exist in the phrase information of FIG. Then, when one of the separated words is not registered as a related word of another word, the word may be registered as a related word of another word.
　シナリオ部２３は、記憶部１３に記憶されたシナリオ情報のうち、語句・概念取得部２２が取得した語句および語句の概念に対応するシナリオ情報を取得する。 The scenario unit 23 acquires scenario information corresponding to the phrase and the phrase concept acquired by the phrase / concept acquisition unit 22 among the scenario information stored in the storage unit 13.
　概念判定部２４は、語句・概念取得部２２が取得した語句の概念について、記憶部１３より取得した、ロボット１の直前の出力内容に含まれていた語句の概念と同じであるか否かを判定する。すなわち、概念判定部２４は、ユーザによる入力内容が、入力される直前に出力された応答メッセージに応じた内容であるか否かを判定する。 The concept determination unit 24 determines whether the phrase concept acquired by the phrase / concept acquisition unit 22 is the same as the concept of the phrase included in the output content immediately before the robot 1 acquired from the storage unit 13. judge. That is, the concept determination unit 24 determines whether or not the content input by the user is content corresponding to the response message output immediately before being input.
　関連語句取得部２５は、語句・概念取得部２２が取得した語句および語句の概念に基づいて、該語句と関連性があり、該語句とは別の語句である関連語句を図２の語句情報から取得する。また、関連語句取得部２５は、取得した関連語句を、後述する推論に用いるための語句に設定する。ここで、関連語句取得部２５が関連語句を取得するための条件はどのようなものであってもよい。例えば、特定の語句について、図２の語句情報において関連語句として登録されている語句をすべて取得してもよいし、入力内容に含まれていない語句のみを取得してもよいし、語句の概念（品詞）が「名詞」である語句を優先して取得してもよい。 The related phrase acquisition unit 25 determines the related phrase that is related to the phrase and is different from the phrase based on the phrase and the concept of the phrase acquired by the phrase / concept acquisition unit 22. Get from. Moreover, the related phrase acquisition part 25 sets the acquired related phrase as a phrase to be used for inference to be described later. Here, the condition for the related phrase acquisition unit 25 to acquire the related phrase may be any. For example, for a specific word / phrase, all the words / phrases registered as related words / phrases in the word / phrase information of FIG. 2 may be acquired, or only words / phrases that are not included in the input content may be acquired. You may preferentially acquire a phrase whose “part of speech” is “noun”.
　応答メッセージ生成部２６は、各部からの指示内容に基づいて、ユーザの入力内容に応じてロボット１が出力する応答メッセージとなるテキストデータを生成する。具体的には、語句・概念取得部２２が取得した語句、および関連語句取得部２５が取得した関連語句の少なくともいずれかと、記憶部１３が記憶するテンプレート、履歴情報、およびシナリオ部２３が取得したシナリオ情報のいずれかとを必要に応じて組み合わせて、テキストデータを生成する。ここで、応答メッセージ生成部２６は、関連語句取得部２５が取得した関連語句が複数であった場合は、所定の条件にしたがって特定の関連語句を抽出する構成であってもよい。たとえば、応答メッセージ生成部２６は、語句・概念取得部２２が取得した語句と同じ概念に属する関連語句を優先的に抽出し、これらの語句を互いに関連付けた内容の応答メッセージを生成する構成であってもよい。なお、応答メッセージ生成部２６は応答メッセージとなるテキストデータを生成する構成であればどのようなものであってもよい。例えば、語句・概念取得部２２のように、上記応答メッセージ生成部２６は、ユーザの入力内容に含まれる複数の語句について、各語句が属する概念および各語句の前後の語句を含む文脈情報に基づいて上記入力語句を選択する構成をさらに備えてもよい。ここで、文脈情報とは、語句が入力内容における主語、述語、その他のいずれに分類されるかを示す情報である。例えば、応答メッセージ生成部２６は、概念（分類）が固有名詞に属する語句、および入力内容において述語である語句を入力語句として優先的に選択する構成であってもよい。 The response message generation unit 26 generates text data that is a response message output by the robot 1 according to the input content of the user, based on the instruction content from each unit. Specifically, at least one of the phrase acquired by the phrase / concept acquisition unit 22 and the related phrase acquired by the related phrase acquisition unit 25, the template stored in the storage unit 13, history information, and the scenario unit 23 are acquired. Text data is generated by combining any of the scenario information as necessary. Here, when there are a plurality of related phrases acquired by the related phrase acquisition unit 25, the response message generation unit 26 may be configured to extract a specific related phrase according to a predetermined condition. For example, the response message generation unit 26 is configured to preferentially extract related words belonging to the same concept as the word acquired by the word / concept acquisition unit 22 and generate a response message having contents associated with these words. May be. Note that the response message generator 26 may have any configuration as long as it generates text data to be a response message. For example, like the phrase / concept acquisition unit 22, the response message generation unit 26 is based on context information including a concept to which each phrase belongs and a phrase before and after each phrase for a plurality of phrases included in the input content of the user. And a configuration for selecting the input words. Here, the context information is information indicating whether a phrase is classified as a subject, predicate, or other in the input content. For example, the response message generation unit 26 may be configured to preferentially select a phrase whose concept (classification) belongs to a proper noun and a phrase that is a predicate in the input content as an input phrase.
　音声合成部２７は、応答メッセージ生成部２６から受信した応答メッセージのテキストデータを音声データに変換する。音声合成部２７は、変換した音声データを音声出力部１２に出力する。 The voice synthesizer 27 converts the text data of the response message received from the response message generator 26 into voice data. The voice synthesizer 27 outputs the converted voice data to the voice output unit 12.
　＜ロボットの動作＞
　図５に基づいて、本実施形態に係るロボット１の動作の概要について説明する。図５は、本実施形態に係るロボット１が推論によって応答メッセージを生成するイメージを示す。なお、以下の説明において、ユーザはロボット１に対して「リンゴは甘い」と発話による入力を行ったものとする。 <Robot motion>
Based on FIG. 5, the outline | summary of operation | movement of the robot 1 which concerns on this embodiment is demonstrated. FIG. 5 shows an image in which the robot 1 according to the present embodiment generates a response message by inference. In the following description, it is assumed that the user inputs an utterance to the robot 1 that “apple is sweet”.
　まず、ユーザがロボット１に対して「リンゴは甘い」と発話すると、ロボット１は該発話を音声認識によってテキストデータに変換し、該テキストデータを語句に区切る。すなわち、「リンゴは甘い」であれば、「リンゴ」、「は」、および「甘い」の３つの語句に区切られる。このとき、ロボット１は、区切られた各語句について、図２の語句情報に登録または更新してもよい。 First, when the user utters “Apple is sweet” to the robot 1, the robot 1 converts the utterance into text data by voice recognition, and divides the text data into words. That is, if “apple is sweet”, it is divided into three phrases “apple”, “ha”, and “sweet”. At this time, the robot 1 may register or update the divided words in the word information in FIG.
　次に、ロボット１は、区切られた語句の中から、推論の起点となる語句を選択し、第１語句（または第２語句）に設定する。ここで、第１語句（または第２語句）は、ユーザの入力内容を特徴づける語句である、入力語句である。なお、第１語句（または第２語句）は、図２で示した語句情報と同じように、単体で意味を備える言葉であれば、どの語句を設定してもよく、所定の優先度にしたがって設定することが好適である。例えば、述語である語句が名詞を含む場合は、これを第１語句（または第２語句）に設定し、含まない場合は主語である名詞を第１語句（または第２語句）に設定する構成であってもよい。また、主語にも述語にも名詞が含まれない場合は、適当な語句を第１語句（または第２語句）に設定する構成であってもよい。上述の例によれば、「リンゴ」、「は」、および「甘い」のうち、述語である「甘い」は形容詞であり、主語である「リンゴ」は名詞であるため、この場合、「リンゴ」を第１語句（または第２語句）に設定する。さらに、第１語句（または第２語句）に設定されなかった語句の中から、第１語句（または第２語句）と互いに関連していると認められる語句について、第１関連語句（または第２関連語句）に設定する。上述の例によれば、「甘い」を第１関連語句（または第２関連語句）に設定する。 Next, the robot 1 selects a word / phrase as a starting point of inference from the divided words and sets it as the first word / phrase (or second word / phrase). Here, the first phrase (or second phrase) is an input phrase that is a phrase that characterizes the user's input content. Note that the first word (or second word) may be set to any word as long as it is a word having meaning in the same manner as the word information shown in FIG. It is preferable to set. For example, when a phrase that is a predicate includes a noun, this is set as a first word (or second word), and when it does not, a noun that is a subject is set as a first word (or second word) It may be. Further, when nouns are included in neither the subject nor the predicate, an appropriate word / phrase may be set as the first word / phrase (or the second word / phrase). According to the above example, among the “apple”, “ha”, and “sweet”, the predicate “sweet” is an adjective, and the subject “apple” is a noun. "Is set to the first word (or second word). Further, among the phrases that are not set as the first phrase (or the second phrase), the first related phrase (or the second phrase) that is recognized as being related to the first phrase (or the second phrase) is used. Related words). According to the above example, “sweet” is set as the first related phrase (or second related phrase).
　第１語句（または第２語句）および第１関連語句（または第２関連語句）を設定した後、ロボット１は、図２の語句情報にしたがって、第１関連語句（または第２関連語句）の関連語句である、第２語句（または第１語句）を取得し、設定する。上述の例によれば、第１関連語句（または第２関連語句）である「甘い」の関連語句は、「リンゴ」、および「メロン」であり、「リンゴ」は第１語句（または第２語句）に設定済みである。したがって、ここでは「メロン」を第２語句（または第１語句）に設定する。 After setting the first word (or second word) and the first related word (or second related word), the robot 1 determines the first related word (or second related word) according to the word information of FIG. The second phrase (or first phrase), which is a related phrase, is acquired and set. According to the above-described example, the related words of “sweet” that are the first related words (or the second related words) are “apple” and “melon”, and “apple” is the first word (or second word). (Phrase) has been set. Therefore, here, “melon” is set as the second word (or first word).
　第２語句（または第１語句）を設定すると、ロボット１は、図２の語句情報にしたがって、第２語句（または第１語句）の関連語句である、第２関連語句（または第１関連語句）を取得し、設定する。なお、第２関連語句（または第１関連語句）は、その概念（利用状況）が「Ｈｏｗ（どのように）」であることが好適である。なぜならば、第２関連語句（または第１関連語句）の概念が「Ｈｏｗ（どのように）」でない場合、第２語句（または第１語句）と組み合わせたときに自然な文章とならないことが多いためである。上述の例によれば、第２語句（または第１語句）である「メロン」の関連語句は、「甘い」、「丸い」、および「買う」である。ここで、「甘い」は第１関連語句（または第２関連語句）に設定済みであるため、残る「丸い」、または「買う」が第２関連語句（または第１関連語句）として設定可能である。このとき、図２の語句情報を用いて「丸い」および「買う」を比較すると、「丸い」の概念（利用状況）が「Ｈｏｗ（どのように）」であり、「買うの」の概念（利用状況）は「－」である。この場合、ロボット１は、「丸い」を第２関連語句（または第１関連語句）に設定する。 When the second word (or first word) is set, the robot 1 sets the second related word (or first related word) that is a related word of the second word (or first word) according to the word information of FIG. ) Get and set. In addition, it is preferable that the concept (usage status) of the second related phrase (or the first related phrase) is “How (how)”. This is because if the concept of the second related phrase (or first related phrase) is not “How”, it often does not become a natural sentence when combined with the second phrase (or first phrase). Because. According to the above-described example, the related phrases of “melon” that is the second phrase (or the first phrase) are “sweet”, “round”, and “buy”. Here, since “sweet” has been set as the first related phrase (or second related phrase), the remaining “round” or “buy” can be set as the second related phrase (or first related phrase). is there. At this time, when “round” and “buy” are compared using the phrase information of FIG. 2, the concept of “round” (usage status) is “How (how)” and the concept of “buy” ( Usage status) is “-”. In this case, the robot 1 sets “round” as the second related phrase (or first related phrase).
　そして、ロボット１（応答メッセージ生成部２６）は、後述する図７のＳ２３のように、これまでの処理によって取得した、第１語句（または第２語句）および第２関連語句（または第１関連語句）から応答メッセージを生成する。 Then, the robot 1 (response message generator 26) obtains the first word (or second word) and the second related word (or first related word) acquired by the processing so far as in S23 of FIG. Response message is generated from the phrase).
　応答メッセージを生成する方法の一例について以下に説明する。以下の例において、図２の語句情報は、助詞等その語句のみでは意味をなさない語句も記憶する。具体的には、図２の語句情報は、「は」という語句を記憶している。このとき、「は」という語句の概念（品詞、分類、利用状況）の組み合わせは（助詞、－、－）であり、関連語句は「リンゴ」、「メロン」、「甘い」、および「丸い」である。上述の例において、第１語句（または第２語句）は「リンゴ」であり、第２関連語句（または第１関連語句）は「丸い」である。また、第２語句（または第１語句）は「メロン」である。したがって、ロボット１は、「リンゴ」および「丸い」を用い、さらに「リンゴ」と「メロン」の両方を関連語句に持つ助詞「は」と組み合わせて「リンゴは丸い」を応答メッセージとして生成する。このようにして、ロボット１は、入力内容に含まれる語句に基づいて、推論によって応答メッセージを生成することができる。 An example of a method for generating a response message is described below. In the following example, the phrase information in FIG. 2 also stores phrases such as particles that do not make sense only by the phrase. Specifically, the phrase information in FIG. 2 stores the phrase “ha”. At this time, the combination of the concept of the word “ha” (part of speech, classification, usage) is (particle, −, −), and the related words are “apple”, “melon”, “sweet”, and “round”. It is. In the above example, the first word (or second word) is “apple” and the second related word (or first related word) is “round”. The second word (or first word) is “melon”. Therefore, the robot 1 uses “apple” and “round”, and further generates a response message “apple round” by combining with the particle “ha” having both “apple” and “melon” as related words. In this way, the robot 1 can generate a response message by inference based on the phrase included in the input content.
　なお、応答メッセージを生成する構成は、上記の内容に限定されることはない。応答メッセージを生成する構成の別の例について以下に説明する。まず、ロボット１は、第２語句（または第１語句）および第２関連語句（または第１関連語句）を取得すると、図３で示した入出力履歴の中から、これらの語句を含む履歴情報を取得する。履歴情報は、第２語句（または第１語句）と第２関連語句（または第１関連語句）とがどのような関連性で関連しているかという情報を含む。例えば、関連性は、語順または助詞等によって表される。上述の例によれば、「内容」が「メロンは丸いです」である履歴情報を取得する。そして、ロボット１（応答メッセージ生成部２６）は、履歴情報の「内容」に含まれる第２語句（または第１語句）を、第１語句（または第２語句）に置換することにより、応答メッセージを生成する。すなわち、上述の例では、「リンゴは丸いです」を応答メッセージとして生成する。入出力履歴の「内容」を用いることにより、ロボット１は、応答メッセージにおける各語句の語順を適切に決定し、助詞等その語句のみでは意味をなさない語句を適切に組み合わせることができる。なお、記憶部１３は、複数の語句の間の関連性（複数の語句の間の語順および／または助詞等）の情報を履歴情報とは別に記憶していてもよい。 Note that the configuration for generating the response message is not limited to the above content. Another example of a configuration for generating a response message will be described below. First, when the robot 1 acquires the second word (or first word) and the second related word (or first related word), the history information including these words from the input / output history shown in FIG. To get. The history information includes information on how the second word (or first word) and the second related word (or first related word) are related. For example, the relevance is expressed by word order or particle. According to the above-described example, history information whose “content” is “the melon is round” is acquired. Then, the robot 1 (response message generator 26) replaces the second word (or first word) included in the “content” of the history information with the first word (or second word) to thereby obtain a response message. Is generated. In other words, in the above example, “apple is round” is generated as a response message. By using the “content” of the input / output history, the robot 1 can appropriately determine the word order of each word / phrase in the response message and appropriately combine words / phrases such as particles that do not make sense. In addition, the memory | storage part 13 may memorize | store the information (relative word order between several phrases, and / or a particle, etc.) information separately from historical information between several phrases.
　なお、第２語句（または第１語句）が見つからない場合、ロボット１（応答メッセージ生成部２６）は、後述する図７のＳ２４のように、第１関連語句（または第２関連語句）、および図４のテンプレートから応答メッセージを生成する構成であってもよい。上述の例において、例えば、テンプレートが「[A]と言えば他に何があるのかな」である場合、第１関連語句（または第２関連語句）である「甘い」と組み合わせた「甘いと言えば他に何があるのかな」を応答メッセージとして生成する。このとき、応答メッセージに対するユーザの入力内容には、第１関連語句（または第２関連語句）に関連付けられた第２語句（または第１語句）が含まれることが期待される。したがって、ロボット１は、応答メッセージに対するユーザの入力内容から取得した語句について、応答メッセージに用いた第１関連語句（または第２関連語句）の関連語句として、図２の語句情報に登録する構成であってもよい。上記の構成によれば、応答メッセージに対するユーザの入力内容から関連語句を取得することが可能となるため、ロボット１の利便性を向上させることができる。 If the second word (or first word) is not found, the robot 1 (response message generator 26), as shown in S24 of FIG. 7 described later, and the first related word (or second related word), and The response message may be generated from the template of FIG. In the above-described example, for example, when the template is “what else is there when saying [A]”, “sweet” combined with “sweet” which is the first related phrase (or second related phrase) “What else is there?” Is generated as a response message. At this time, it is expected that the input content of the user with respect to the response message includes the second word (or first word) associated with the first related word (or second related word). Therefore, the robot 1 registers the word / phrase acquired from the input content of the user with respect to the response message in the word / phrase information of FIG. 2 as the related word / phrase of the first related word / phrase (or the second related word / phrase) used in the response message. There may be. According to said structure, since it becomes possible to acquire a related phrase from the user's input content with respect to a response message, the convenience of the robot 1 can be improved.
　さらに、ロボット１は、ユーザの入力内容が対話の流れを無視したものである場合は、後述する図７のＳ２６のように、第１語句および図４のテンプレートから別の応答メッセージを生成する構成であってもよい。このとき、ロボット１は、ユーザの入力内容と対話の流れとを比較できるのであればどのような構成を備えてもよい。例えば、ロボット１（概念判定部２４）は、ユーザの入力の直前にロボット１が出力した内容を図３の入出力履歴より取得し、該履歴に含まれる語句を第３語句とする。このとき、ロボット１は、図２で示した語句情報を用いて第３語句が属する概念と第１語句（または第２語句）が属する概念とが同じであるか否かを判定し、異なる場合は対話の流れを無視した入力が行われたと判定してもよい。このとき、ロボット１は、例えば第１語句（または第２語句）を、第３語句に関連するテンプレートに組み合わせた別の応答メッセージを生成してもよい。ロボット１が「リンゴは甘いです」と発話した後にユーザが「京都と言えば金閣寺だね」と発話した場合、第１語句（または第２語句）である「金閣寺」の概念（分類）は「国内地域」である一方、第３語句である「リンゴ」は「果物」であり、概念（分類）が異なる。 Furthermore, when the input content of the user ignores the flow of dialogue, the robot 1 generates another response message from the first word and the template of FIG. 4 as in S26 of FIG. 7 described later. It may be. At this time, the robot 1 may have any configuration as long as it can compare the input content of the user and the flow of dialogue. For example, the robot 1 (concept determination unit 24) acquires the content output by the robot 1 immediately before the user's input from the input / output history of FIG. 3, and sets the word / phrase included in the history as the third word / phrase. At this time, the robot 1 determines whether or not the concept to which the third phrase belongs and the concept to which the first phrase (or the second phrase) belongs are the same using the phrase information shown in FIG. May determine that input has been made ignoring the flow of dialogue. At this time, the robot 1 may generate another response message in which, for example, the first word (or second word) is combined with the template related to the third word. When the robot 1 utters “Apple is sweet” and the user utters “Kyoto is Speaking of Kinkakuji”, the concept (classification) of “Kinkakuji”, which is the first (or second) phrase, is “ The third phrase “apple” is “fruit” and the concept (classification) is different, while “domestic region”.
　このとき、ロボット１は、第１語句（または第２語句）「金閣寺」を第３語句「リンゴ」概念（分類）である「果物（食べ物）」に関連するテンプレート「[A]は食べ物じゃないでしょ」または「[A]は食べられないよ」に組み合わせた別の応答メッセージを生成する。すなわち、例文によれば、「金閣寺は食べ物じゃないでしょ」、「金閣寺は食べられないよ」のいずれかを応答メッセージとして出力する。 At this time, the robot 1 uses the first word (or second word) “Kinkakuji” as the template “[A] related to“ fruit (food) ”which is the concept (classification) of the third word“ apple ”. Produces another response message combined with "I can't eat [A]". That is, according to the example sentence, either “Kinkakuji is not food” or “I can't eat Kinkakuji” is output as a response message.
　なお、別の応答メッセージを生成する方法は上記に限定されることはない。例えば、ロボット１は、ユーザの入力内容に含まれる、第１語句（または第２語句）とは別の語句を第４語句として、第１語句（または第２語句）と、第４語句とを、テンプレートに組み合わせて別の応答メッセージを生成してもよい。ロボット１が「リンゴは甘いです」と発話した後にユーザが「金閣寺も甘いね」と発話した場合、第１語句（または第２語句）「金閣寺」と第３語句「リンゴ」の概念（分類）が異なるものの、「甘い」の語句が共通している。このとき、ロボット１は、ユーザの入力内容に含まれる、第１語句（または第２語句）「金閣寺」とは別の語句「甘い」を第４語句とする。そして、ロボット１は、第１語句（または第２語句）と、第４語句とを、テンプレート「[A]は[B]よね　ってなんでやねん」、「[A]は[B]なんだ　今度食べに連れて行ってよね」、および「[A]は[B]よね」のいずれかに組み合わせて別の応答メッセージを生成する。すなわち、例文によれば、「金閣寺は甘いよね　ってなんでやねん」、「金閣寺って甘いんだね　今度食べに連れて行ってよね」、および「金閣寺は甘いよね」のいずれかを出力する。 Note that the method of generating another response message is not limited to the above. For example, the robot 1 sets a first phrase (or second phrase) and a fourth phrase as a fourth phrase that is different from the first phrase (or second phrase) included in the input content of the user. Another response message may be generated in combination with the template. If the user utters “Kinkakuji is sweet” after the robot 1 utters “Apple is sweet”, the concept (classification) of the first phrase (or second phrase) “Kinkakuji” and the third phrase “apple” Although they are different, the phrase “sweet” is common. At this time, the robot 1 sets the phrase “sweet”, which is different from the first phrase (or the second phrase) “Kinkakuji”, included in the input content of the user as the fourth phrase. Then, the robot 1 uses the first phrase (or the second phrase) and the fourth phrase, the template “[A] is [B], what is it?”, “[A] is [B] Combine this with "Take me to eat this time" and "[A] is [B]" to generate another response message. That is, according to the example sentence, "Kinkakuji is sweet" Why is it, "Kinkakuji is sweet, please take me next time" and "Kinkakuji is sweet" is output .
　また、推論は上述の方法に限定されることはなく、第１語句（または第２語句）、およびテンプレートと組み合わせて応答メッセージを生成するために必要な第２関連語句（または第１関連語句）を取得できるのであれば、どのような方法であってもよい。推論の別の例について、図６を用いて説明する。図６は、本実施形態に係るロボット１が推論によって応答メッセージを生成する別のイメージを示す。 In addition, the inference is not limited to the above-described method, and the second related phrase (or the first related phrase) necessary for generating the response message in combination with the first phrase (or the second phrase) and the template. Any method may be used as long as it can be acquired. Another example of inference will be described with reference to FIG. FIG. 6 shows another image in which the robot 1 according to the present embodiment generates a response message by inference.
　図６の（ａ）は、「リンゴは甘い」と、「メロンは丸い」という２つの文章を示している。２つの文章のうち一方が入力内容であり、他方が記憶部１３に記憶されている過去の入出力内容である。例えば、「リンゴは甘い」が入力内容であり、「メロンは丸い」が過去の入出力内容であるものとする。具体的には、ロボット１はユーザの入力内容「リンゴは甘い」を受け付け、「リンゴ」を第１語句に設定する。その後、ロボット１は、第１語句「リンゴ」と概念（品詞、分類、利用状況）の組み合わせが（名詞、果物、Ｗｈａｔ（何が））であり、同一概念である第２語句「メロン」を含む履歴情報「メロンは丸い」を記憶部から検索し、第２語句「メロン」を取得する。第１語句「リンゴ」と第２語句「メロン」は同一概念であるため、この場合、ロボット１は、図６の（ｂ）のように、概念が同一の語句で２つの文章をまとめた地図を生成する。そして、（名詞、果物、Ｗｈａｔ（何が））が設定されている領域には、「リンゴ」および「メロン」のいずれかを設定する。 (A) in FIG. 6 shows two sentences “apple is sweet” and “melon is round”. One of the two sentences is input contents, and the other is past input / output contents stored in the storage unit 13. For example, it is assumed that “apple is sweet” is the input content and “melon is round” is the past input / output content. Specifically, the robot 1 accepts the user's input content “apple is sweet” and sets “apple” as the first phrase. Thereafter, the robot 1 has a combination of the first phrase “apple” and the concept (part of speech, classification, usage) (noun, fruit, What (what)), and the second phrase “melon” that is the same concept. The history information including “melon is round” is searched from the storage unit, and the second word “melon” is acquired. Since the first phrase “apple” and the second phrase “melon” have the same concept, in this case, as shown in FIG. 6B, the robot 1 is a map in which two sentences are grouped with the same concept. Is generated. In the region where (noun, fruit, What (what)) is set, either “apple” or “melon” is set.
　このとき、図６の（ｂ）からは、「リンゴは甘い」、「リンゴは丸い」、「メロンは甘い」、および「メロンは丸い」の４つの文章を取得することができる。結果として、ロボット１は、記憶部１３において互いに関連付けられていない２つの語句の組み合わせを含む文章「リンゴは丸い」、および「メロンは甘い」を推論の結果として取得し、応答メッセージを生成することができる。 At this time, from (b) of FIG. 6, it is possible to obtain four sentences of “apple is sweet”, “apple is round”, “melon is sweet”, and “melon is round”. As a result, the robot 1 acquires a sentence “apple is round” and “melon is sweet” including a combination of two words and phrases that are not associated with each other in the storage unit 13 as a result of inference, and generates a response message Can do.
　そして、上述の推論は、従来技術である会話シナリオに基づく応答メッセージの生成と組み合わせて用いる構成であってもよい。このとき、記憶部１３はシナリオ情報をさらに記憶していることが好適である。この場合、ロボット１は、ユーザの入力内容に関連付けられたシナリオ情報が記憶部１３に存在する場合は、後述する図７のＳ２５のように、入力内容に含まれる第１語句（または第２語句）と会話シナリオから応答メッセージを生成する構成であってもよい。また、シナリオ情報が記憶部１３に存在しない場合は推論による応答メッセージの生成を行う構成であってもよい。 Further, the above inference may be configured to be used in combination with generation of a response message based on a conversation scenario which is a conventional technique. At this time, it is preferable that the storage unit 13 further stores scenario information. In this case, when the scenario information associated with the user input content exists in the storage unit 13, the robot 1 uses the first word (or second word) included in the input content as in S25 of FIG. ) And a conversation scenario may be generated from the conversation scenario. Moreover, when scenario information does not exist in the memory | storage part 13, the structure which produces | generates the response message by inference may be sufficient.
　＜処理の流れ＞
　次に、本実施形態のロボット１が実行する処理の流れについて、図７に基づいて説明する。図７は、本実施形態のロボット１が実行する処理の流れの一例を示すフローチャートである。なお、以下の説明において、ユーザの入力内容には第１語句および第１関連語句が含まれ、ユーザが入力する直前にロボット１が出力した応答メッセージには第３語句が含まれるものとする。 <Process flow>
Next, the flow of processing executed by the robot 1 of this embodiment will be described with reference to FIG. FIG. 7 is a flowchart illustrating an example of a flow of processing executed by the robot 1 of the present embodiment. In the following description, it is assumed that the user's input content includes the first word and the first related word, and the response message output by the robot 1 immediately before the user inputs includes the third word.
　まず、音声認識部２１は、音声入力部１１が受け付けたユーザの入力について、入力内容の音声認識を行い、テキストデータを生成する（Ｓ１１）。次に、語句・概念取得部２２は、Ｓ１１で生成されたテキストデータを受け付け、該テキストデータおよび記憶部１３が記憶している語句情報から第１語句、第１関連語句、および第１語句の概念を取得する（Ｓ１２）。そして、シナリオ部２３は、Ｓ１２で取得した第１語句または第１語句の概念に対応するシナリオ情報を記憶部１３の中から検索し（Ｓ１３）、対応するシナリオ情報が見つかったか否かを判定する（Ｓ１４）。対応するシナリオ情報が見つからなかったと判定した場合（Ｓ１４でＮＯ）、概念判定部２４は、ユーザの入力直前にロボット１が応答（発話）したか否かを判定する（Ｓ１５）。一方、対応するシナリオ情報が見つかったと判定した場合（Ｓ１４でＹＥＳ）、処理は後述のＳ２５へ進む。 First, the voice recognition unit 21 performs voice recognition of the input content for the user input received by the voice input unit 11, and generates text data (S11). Next, the phrase / concept acquisition unit 22 receives the text data generated in S11, and from the text data and the phrase information stored in the storage unit 13, the first phrase, the first related phrase, and the first phrase A concept is acquired (S12). Then, the scenario unit 23 searches the storage unit 13 for scenario information corresponding to the first phrase or the concept of the first phrase acquired in S12 (S13), and determines whether the corresponding scenario information has been found. (S14). When it is determined that the corresponding scenario information has not been found (NO in S14), the concept determination unit 24 determines whether or not the robot 1 has responded (spoken) immediately before the user input (S15). On the other hand, if it is determined that the corresponding scenario information has been found (YES in S14), the process proceeds to S25 described later.
　Ｓ１５の処理において、ロボット１が応答したと判定した場合（Ｓ１５でＹＥＳ）、概念判定部２４はさらに、第１語句と第３語句の概念は同一であるか否かを判定する（Ｓ１６）。そして、概念は同一であると判定した場合（Ｓ１６でＹＥＳ）、関連語句取得部２５は、記憶部１３が記憶している語句情報の第１関連語句の関連語句から第２語句を検索する（Ｓ１７：探索ステップ）。一方、ロボット１が応答していなかったと判定した場合（Ｓ１５でＮＯ）、処理はＳ１７へ直接進む。 In the process of S15, when it is determined that the robot 1 has responded (YES in S15), the concept determination unit 24 further determines whether the concepts of the first phrase and the third phrase are the same (S16). When it is determined that the concepts are the same (YES in S16), the related phrase acquisition unit 25 searches for the second phrase from the related phrases of the first related phrase in the phrase information stored in the storage unit 13 ( S17: Search step). On the other hand, if it is determined that the robot 1 has not responded (NO in S15), the process directly proceeds to S17.
　Ｓ１７の後、関連語句取得部２５は、第２語句が見つかったか否か（第１関連語句の関連語句が語句情報にあるか否か）を判定する（Ｓ１８）。第２語句が見つからなかったと判定した場合（Ｓ１８でＮＯ）、関連語句取得部２５は、第１語句と同一概念である第２語句を記憶部１３が記憶している語句情報から検索する（Ｓ１９）。そして、関連語句取得部２５は、Ｓ１９の処理において第２語句が見つかったか否かを判定する（Ｓ２０）。第２語句が見つかったと判定した場合（Ｓ２０でＹＥＳ）、またはＳ１８において第２語句が見つかったと判定した場合（Ｓ１８でＹＥＳ）、関連語句取得部２５は、記憶部１３の語句情報において、第２語句の関連語句から第２関連語句を検索する（Ｓ２１）。 After S17, the related phrase acquisition unit 25 determines whether or not the second phrase has been found (whether or not the related phrase of the first related phrase is in the phrase information) (S18). When it is determined that the second phrase has not been found (NO in S18), the related phrase acquisition unit 25 searches the phrase information stored in the storage unit 13 for the second phrase having the same concept as the first phrase (S19). ). And the related phrase acquisition part 25 determines whether the 2nd phrase was found in the process of S19 (S20). When it is determined that the second word / phrase is found (YES in S20), or when it is determined that the second word / phrase is found in S18 (YES in S18), the related word / phrase acquisition unit 25 uses the second word / phrase information in the storage unit 13 as the second word / phrase information. A second related phrase is retrieved from the related phrases of the phrase (S21).
　その後、関連語句取得部２５は、Ｓ２１の処理において第２関連語句が見つかったか否か（第２語句の関連語句が語句情報にあるか否か）を判定する（Ｓ２２）。第２関連語句が見つかったと判定した場合（Ｓ２２でＹＥＳ）、応答メッセージ生成部２６は、第１語句および第２関連語句から応答メッセージを生成し（Ｓ２３：応答メッセージ生成ステップ）、処理はＳ２７へ進む。このようにして、応答メッセージ生成部２６は、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成する。一方、Ｓ２０の処理で第２語句が見つからなかった場合（Ｓ２０でＮＯ）、またはＳ２２の処理で第２関連語句が見つからなかった場合（Ｓ２２でＮＯ）、応答メッセージ生成部２６は、第１関連語句、およびテンプレートから応答メッセージを生成する（Ｓ２４）。その後、処理はＳ２７へ進む。また、Ｓ１４において対応するシナリオ情報が見つかったと判定した場合（Ｓ１４でＹＥＳ）、応答メッセージ生成部２６は、第１語句とシナリオ情報から応答メッセージを生成し（Ｓ２５）、処理はＳ２７へ進む。そして、Ｓ１６において第１語句と第３語句の概念は同一ではないと判定した場合（Ｓ１６でＮＯ）、応答メッセージ生成部２６は、第１語句およびテンプレートから応答メッセージを生成し（Ｓ２６）、処理はＳ２７へ進む。 Thereafter, the related phrase acquisition unit 25 determines whether or not the second related phrase is found in the process of S21 (whether or not the related phrase of the second phrase is in the phrase information) (S22). If it is determined that the second related phrase has been found (YES in S22), the response message generator 26 generates a response message from the first phrase and the second related phrase (S23: response message generation step), and the process proceeds to S27. move on. In this way, the response message generator 26 generates a response message having the content in which the first word / phrase and the second related word / phrase are associated with each other. On the other hand, when the second word / phrase is not found in the process of S20 (NO in S20), or when the second related word / phrase is not found in the process of S22 (NO in S22), the response message generator 26 generates the first related word A response message is generated from the phrase and the template (S24). Thereafter, the process proceeds to S27. If it is determined in S14 that the corresponding scenario information has been found (YES in S14), the response message generator 26 generates a response message from the first phrase and scenario information (S25), and the process proceeds to S27. If it is determined in S16 that the concepts of the first phrase and the third phrase are not the same (NO in S16), the response message generator 26 generates a response message from the first phrase and the template (S26), and processing Advances to S27.
　Ｓ２７において、音声合成部２７は応答メッセージを音声データに変換し、音声出力部１２を介して音声出力する（Ｓ２７）。 In S27, the voice synthesis unit 27 converts the response message into voice data, and outputs the voice via the voice output unit 12 (S27).
　以上の処理により、本実施形態に係るロボット１は、適合する会話シナリオが見つからない場合であっても、入力内容に含まれる語句に基づいて応答メッセージを生成することが可能となる。 Through the above processing, the robot 1 according to the present embodiment can generate a response message based on the words / phrases included in the input content even when no suitable conversation scenario is found.
　〔実施形態２〕
　本発明の他の実施形態について、図７～図９に基づいて説明する。本実施形態のロボット１は、ユーザの入力内容が３つ以上の語句を含む場合、３つ以上の語句を含む応答メッセージを生成する点で前記実施形態と異なる。なお、説明の便宜上、前記実施形態にて説明した部材と同じ機能を有する部材については、同じ符号を付記し、その説明を省略する。本実施形態に係るロボット１は、ユーザの入力内容に含まれる３つ以上の語句について、前記実施形態と同様にして第１語句（または第２語句）および第１関連語句（または第２関連語句）を設定する。また、語句・概念取得部２２は、第１関連語句（または第２関連語句）の関連語句であり、第１語句（または第２語句）とは異なる別の語句について、第１付随語句（または第２付随語句）に設定する。ここで、入力内容から第１付随語句（または第２付随語句）を設定する基準はどのようなものであってもよい。例えば、前記実施形態と同様にして第１語句を設定した後、残る語句のうち、概念（品詞）が「品詞」であり、かつ概念（分類）が「その他」以外である語句を第１付随語句（または第２付随語句）に設定してもよい。また、第１関連語句（または第２関連語句）と第１付随語句（または第２付随語句）を設定する順番は特に限定されないが、語句の概念（品詞）が「名詞」である語句を第１付随語句（または第２付随語句）に優先して設定することが好適である。そして、第１語句（または第２語句）および第１付随語句（または第２付随語句）を、推論によって取得された第２関連語句（または第１関連語句）と互いに関連付けた応答メッセージを生成する。 [Embodiment 2]
Another embodiment of the present invention will be described with reference to FIGS. The robot 1 according to the present embodiment differs from the previous embodiment in that a response message including three or more words / phrases is generated when the input content of the user includes three or more words / phrases. For convenience of explanation, members having the same functions as those described in the embodiment are given the same reference numerals, and descriptions thereof are omitted. In the robot 1 according to the present embodiment, the first phrase (or second phrase) and the first related phrase (or second related phrase) in the same manner as in the above embodiment, regarding three or more words included in the input content of the user. ) Is set. Further, the phrase / concept acquisition unit 22 is a related phrase of the first related phrase (or second related phrase), and is different from the first phrase (or second phrase) with respect to the first accompanying phrase (or 2nd accompanying phrase). Here, any criteria may be used for setting the first accompanying phrase (or second accompanying phrase) from the input content. For example, after the first word / phrase is set in the same manner as in the above embodiment, among the remaining words / phrases, a word / phrase whose concept (part of speech) is “part of speech” and whose concept (classification) is other than “other” is first attached. You may set to a phrase (or 2nd accompanying phrase). The order of setting the first related phrase (or second related phrase) and the first accompanying phrase (or second accompanying phrase) is not particularly limited, but the phrase whose concept (part of speech) is “noun” It is preferable to set in preference to one accompanying phrase (or second accompanying phrase). Then, a response message is generated in which the first phrase (or second phrase) and the first accompanying phrase (or second accompanying phrase) are associated with the second related phrase (or first related phrase) acquired by inference. .
　＜ロボットの動作＞
　図８および図９に基づいて、本実施形態に係るロボット１の動作の概要について説明する。図８および図９は、本実施形態に係るロボット１が推論によって応答メッセージを生成するイメージを示す図である。 <Robot motion>
Based on FIG. 8 and FIG. 9, an outline of the operation of the robot 1 according to the present embodiment will be described. 8 and 9 are diagrams illustrating an image in which the robot 1 according to the present embodiment generates a response message by inference.
　図８の（ａ）は、「大阪の名物はお好み焼き」、および「広島は牡蠣がおいしい」の２つの文章について推論を適用するイメージを示す。例えば、「広島は牡蠣がおいしい」は過去の入出力として予め記憶部１３に記憶されており、「大阪の名物はお好み焼き」がユーザによって入力されたとする。このとき、ロボット１（語句・概念取得部２２）は、「お好み焼き」を第１語句（または第２語句）、「名物」を第１関連語句（または第２関連語句）、および「大阪」を第１付随語句（または第２付随語句）に設定する。そして、ロボット１（関連語句取得部２５）は、第１語句（または第２語句）「お好み焼き」と概念（品詞、分類、利用状況）の組み合わせが同一である語句「牡蠣」を含み、第１付随語句（または第２付随語句）「大阪」と同一概念である語句「広島」を含む履歴情報「広島は牡蠣がおいしい」を図３の入出力履歴から取得する。ここで、「お好み焼き」と「牡蠣」は、分類が「食べ物／お好み焼き／たこ焼き」と「食べ物／海産物」という点で異なるが、「食べ物」という部分は一致している。そのため、「お好み焼き」と「牡蠣」は、同じように用いることができる語句（下位概念は異なるが上位概念が一致する語句）であると判定している。このとき、ロボット１（応答メッセージ生成部２６）は、図６の（ｂ）と同様にして、図８の（ｂ）のように、概念が同一の語句で２つの文章をまとめた地図を生成する。そして、（名詞、国内地域、Ｗｈｅｒｅ（どこが））の領域には「大阪」および「広島」のいずれかを設定し、（名詞、食べ物、Ｗｈａｔ（何が））の領域には「お好み焼き」および「牡蠣」のいずれかを設定する。ただし、設定される語句の組み合わせには意味があることが多いため、同一の発話で用いられた語句の組み合わせを維持しつつ、設定する必要がある。例えば、第１付随語句（または第２付随語句）「大阪」は第１語句（または第２語句）「お好み焼き」と組み合わせて設定し、同様に、「牡蠣」は「広島」と組み合わせて設定する必要がある。 (A) in FIG. 8 shows an image in which reasoning is applied to two sentences, “Osaka's famous okonomiyaki” and “Hiroshima is delicious oysters”. For example, “Hiroshima is delicious oysters” is stored in the storage unit 13 in advance as past input / output, and “Osaka specialty okonomiyaki” is input by the user. At this time, the robot 1 (the phrase / concept acquisition unit 22) sets “okonomiyaki” as the first phrase (or second phrase), “specialty” as the first related phrase (or second related phrase), and “Osaka”. Set to the first accompanying phrase (or second accompanying phrase). The robot 1 (related phrase acquisition unit 25) includes the phrase “oyster” having the same combination of the first phrase (or second phrase) “okonomiyaki” and the concept (part of speech, classification, usage). The history information “Hiroshima is oyster is delicious” including the phrase “Hiroshima” having the same concept as the accompanying phrase (or second accompanying phrase) “Osaka” is acquired from the input / output history of FIG. Here, “okonomiyaki” and “oyster” are different in terms of “food / okonomiyaki / takoyaki” and “food / seafood”, but the parts “food” are the same. For this reason, “okonomiyaki” and “oyster” are determined to be terms that can be used in the same manner (words that have different concepts but match the higher concepts). At this time, the robot 1 (response message generator 26) generates a map in which two sentences are grouped together with words having the same concept, as shown in FIG. 8B, in the same manner as FIG. 6B. To do. Then, set either “Osaka” or “Hiroshima” in the (noun, domestic, where) area, and “okonomiyaki” and (noun, food, what) in the area. Set one of “Oysters”. However, since combinations of words and phrases that are set are often meaningful, it is necessary to set them while maintaining a combination of words and phrases used in the same utterance. For example, the first accompanying word (or second accompanying word) “Osaka” is set in combination with the first word (or second word) “okonomiyaki”, and similarly, “oyster” is set in combination with “Hiroshima”. There is a need.
　したがって、図８の（ｂ）からは、「大阪の名物はお好み焼き」、「大阪はお好み焼きがおいしい」、「広島の名物は牡蠣」、および「広島は牡蠣がおいしい」の４つの文章を取得することができる。結果として、ロボット１は、「大阪はお好み焼きがおいしい」、または「広島の名物は牡蠣」を推論の結果として取得し、応答メッセージを生成することができる。 Therefore, from FIG. 8B, four sentences are obtained: “Osaka's specialty is Okonomiyaki”, “Osaka is Okonomiyaki”, “Hiroshima's specialty is oyster” and “Hiroshima is oyster” be able to. As a result, the robot 1 can acquire “Osaka okonomiyaki is delicious” or “Hiroshima's specialty is oysters” as a result of inference and generate a response message.
　図９の（ａ）は、「大阪の名物はお好み焼き」、および「広島のおみやげはもみじ饅頭」の２つの文章について推論を適用するイメージを示す。例えば、「広島のおみやげはもみじ饅頭」は過去の入出力として予め記憶部１３に記憶されており、「大阪の名物はお好み焼き」がユーザによって入力されたとする。このとき、ロボット１（語句・概念取得部２２）は、図８の場合と同様にして、「お好み焼き」を第１語句（または第２語句）、「名物」を第１関連語句（または第２関連語句）、および「大阪」を第１付随語句（または第２付随語句）に設定する。そして、ロボット１（関連語句取得部２５）は、第１語句（または第２語句）「お好み焼き」と概念（品詞、分類、利用状況）の組み合わせが同一である語句「もみじ饅頭」を含み、第１付随語句（または第２付随語句）「大阪」と同一概念である語句「広島」を含む履歴情報「広島のおみやげはもみじ饅頭」を図３の入出力履歴から取得する。このとき、ロボット１（応答メッセージ生成部２６）は、図９の（ｂ）のように概念が同一の語句で２つの文章をまとめた地図を生成する。ここで、「名物」と「おみやげ」は、概念（品詞、分類）の組み合わせが（名詞、その他）で一致している。しかし、概念（品詞、分類）の組み合わせが（名詞、その他）である語句は、該語句が示す特徴が他の語句に対して少ないため、まとめずに用いる。この場合、図９の（ｂ）からは、図８の（ｂ）と同様にして、（名詞、国内地域、Ｗｈｅｒｅ（どこが））の領域には「大阪」および「広島」のいずれかを設定する。そして、（名詞、食べ物、Ｗｈａｔ（何が））の領域には「お好み焼き」および「もみじ饅頭」のいずれかを設定する。 (A) in FIG. 9 shows an image in which reasoning is applied to two sentences, “Osaka's specialty is Okonomiyaki” and “Hiroshima souvenir is Momiji Wharf”. For example, it is assumed that “Hiroshima souvenir is Momiji Wharf” is previously stored in the storage unit 13 as past input / output, and “Osaka specialty okonomiyaki” is input by the user. At this time, the robot 1 (the phrase / concept acquisition unit 22), as in the case of FIG. 8, sets “okonomiyaki” as the first phrase (or the second phrase) and “specialties” as the first related phrase (or the second phrase). Related words) and “Osaka” are set as the first accompanying words (or second accompanying words). Then, the robot 1 (related phrase acquisition unit 25) includes the phrase “Momiji bun” having the same combination of the first phrase (or second phrase) “okonomiyaki” and the concept (part of speech, classification, usage). The history information “Hiroshima souvenir is Momiji Wharf” including the phrase “Hiroshima”, which is the same concept as the 1 accompanying phrase (or second accompanying phrase) “Osaka”, is acquired from the input / output history of FIG. At this time, the robot 1 (response message generator 26) generates a map in which two sentences are grouped together with words having the same concept as shown in FIG. 9B. Here, “specialties” and “souvenirs” match in terms of the combination of concepts (parts of speech, classification) (nouns, etc.). However, a phrase having a combination of concepts (parts of speech, classification) (noun, other) is used without being summarized because it has fewer features than other phrases. In this case, from (b) of FIG. 9, as in (b) of FIG. 8, either “Osaka” or “Hiroshima” is set in the area of (noun, domestic region, where). To do. In the area of (noun, food, What), either “okonomiyaki” or “maple bun” is set.
　これにより、ロボット１は、「大阪の名物はお好み焼き」、「大阪のおみやげはお好み焼き」、「広島の名物は牡蠣」、および「広島のおみやげは牡蠣」の４つの文章を取得することができる。結果として、ロボット１は、「大阪のおみやげはお好み焼き」、および「広島の名物は牡蠣」のを推論の結果として取得し、応答メッセージを生成することができる。 Thus, the robot 1 can acquire four sentences, “Osaka's specialty is okonomiyaki”, “Osaka's souvenir is okonomiyaki”, “Hiroshima's specialty is okonomiyaki”, and “Hiroshima's souvenir is okonomiyaki”. As a result, the robot 1 can acquire “Osaka souvenir is okonomiyaki” and “Hiroshima specialties are oysters” as a result of inference, and can generate a response message.
　＜処理の流れ＞
　次に、本実施形態のロボット１が実行する処理の流れについて、図７に基づいて説明する。図７は、本実施形態のロボット１が実行する処理の流れの一例を示すフローチャートである。なお、以下の説明において、ユーザの入力内容には第１語句および第１関連語句が含まれ、ユーザが入力する直前にロボット１が出力した応答メッセージには第３語句が含まれるものとする。ここで、ユーザの入力内容が「大阪の名物はお好み焼き」である場合、第１語句および第１関連語句は、それぞれ「お好み焼き」および「名物」であり、「大阪」は、第１付随語句である。 <Process flow>
Next, the flow of processing executed by the robot 1 of this embodiment will be described with reference to FIG. FIG. 7 is a flowchart illustrating an example of a flow of processing executed by the robot 1 of the present embodiment. In the following description, it is assumed that the user's input content includes the first word and the first related word, and the response message output by the robot 1 immediately before the user inputs includes the third word. Here, when the input content of the user is “Osaka specialty is Okonomiyaki”, the first word and the first related phrase are “okonomiyaki” and “specialty”, respectively, and “Osaka” is the first accompanying phrase. is there.
　Ｓ１１～Ｓ２２の処理は、実施形態１と同一である。Ｓ２２でＹＥＳの場合、応答メッセージ生成部２６は、第１語句、第１付随語句および第２関連語句から応答メッセージを生成し（Ｓ２３）、処理はＳ２７へ進む。具体的には、応答メッセージ生成部２６は、第１語句、第１付随語句、および第２関連語句を互いに関連付けた応答メッセージ「大阪のおみやげはお好み焼き」または「広島の名物は牡蠣」を生成する。また、Ｓ２０の処理で第２語句が見つからなかった場合（Ｓ２０でＮＯ）、またはＳ２２の処理で第２関連語句が見つからなかった場合（Ｓ２２でＮＯ）、応答メッセージ生成部２６は、第１関連語句、およびテンプレートから応答メッセージを生成する（Ｓ２４）。その後、処理はＳ２７へ進む。また、Ｓ１４において対応するシナリオ情報が見つかったと判定した場合（Ｓ１４でＹＥＳ）、応答メッセージ生成部２６は、第１語句、第１付随語句、およびシナリオ情報から応答メッセージを生成し（Ｓ２５）、処理はＳ２７へ進む。そして、Ｓ１６において第１語句と第３語句の概念は同一ではないと判定した場合（Ｓ１６でＮＯ）、応答メッセージ生成部２６は、第１語句、第１付随語句、およびテンプレートから応答メッセージを生成し（Ｓ２６）、処理はＳ２７へ進む。Ｓ２７の処理は、実施形態１と同一である。 The processing of S11 to S22 is the same as that of the first embodiment. If YES in S22, the response message generator 26 generates a response message from the first phrase, the first accompanying phrase, and the second related phrase (S23), and the process proceeds to S27. Specifically, the response message generator 26 generates a response message “Osaka souvenir is okonomiyaki” or “Hiroshima specialties are oysters” in which the first phrase, the first accompanying phrase, and the second related phrase are associated with each other. . When the second word / phrase is not found in the process of S20 (NO in S20) or the second related word / phrase is not found in the process of S22 (NO in S22), the response message generating unit 26 selects the first related word A response message is generated from the phrase and the template (S24). Thereafter, the process proceeds to S27. If it is determined in S14 that the corresponding scenario information has been found (YES in S14), the response message generating unit 26 generates a response message from the first word / phrase, the first accompanying word / phrase, and the scenario information (S25). Advances to S27. If it is determined in S16 that the concepts of the first word and the third word are not the same (NO in S16), the response message generator 26 generates a response message from the first word, the first accompanying word, and the template. Then (S26), the process proceeds to S27. The process of S27 is the same as that of the first embodiment.
　以上の処理により、本実施形態に係るロボット１は、ユーザの入力内容が３語以上の語句を含む場合に適合する会話シナリオが見つからない場合であっても、入力内容に含まれる語句に基づいて応答メッセージを生成することが可能となる。 Through the above processing, the robot 1 according to the present embodiment is based on the words / phrases included in the input contents even when a conversation scenario that matches when the user's input contents include three or more words / phrases is not found. A response message can be generated.
　〔変形例〕
　前記各実施形態では、ロボット１がユーザに対して音声で対話する構成について説明したがこれに限らない。例えば制御部２０は、ユーザに対してテキストを用いて対話する応答装置に備えられていてもよい。また、ユーザおよび応答装置間の音声データまたはテキストによる対話は、ネットワークを介して行われてもよい。 [Modification]
In each of the above-described embodiments, the configuration in which the robot 1 interacts with the user by voice has been described. For example, the control unit 20 may be provided in a response device that interacts with a user using text. Further, voice data or text interaction between the user and the response device may be performed via a network.
　〔ソフトウェアによる実現例〕
　ロボット１の制御部２０（特に応答メッセージ生成部２６および概念判定部２４）は、集積回路（ＩＣチップ）等に形成された論理回路（ハードウェア）によって実現してもよいし、ＣＰＵ（Central Processing Unit）を用いてソフトウェアによって実現してもよい。 [Example of software implementation]
The control unit 20 (particularly the response message generation unit 26 and the concept determination unit 24) of the robot 1 may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or a CPU (Central Processing). Unit) and may be realized by software.
　後者の場合、制御部２０は、各機能を実現するソフトウェアであるプログラムの命令を実行するＣＰＵ、上記プログラムおよび各種データがコンピュータ（またはＣＰＵ）で読み取り可能に記録されたＲＯＭ（Read Only Memory）または記憶装置（これらを「記録媒体」と称する）、上記プログラムを展開するＲＡＭ（Random Access Memory）などを備えている。そして、コンピュータ（またはＣＰＵ）が上記プログラムを上記記録媒体から読み取って実行することにより、本発明の目的が達成される。上記記録媒体としては、「一時的でない有形の媒体」、例えば、テープ、ディスク、カード、半導体メモリ、プログラマブルな論理回路などを用いることができる。また、上記プログラムは、該プログラムを伝送可能な任意の伝送媒体（通信ネットワークや放送波等）を介して上記コンピュータに供給されてもよい。なお、本発明は、上記プログラムが電子的な伝送によって具現化された、搬送波に埋め込まれたデータ信号の形態でも実現され得る。 In the latter case, the control unit 20 includes a CPU that executes instructions of a program that is software that realizes each function, a ROM (Read Only Memory) in which the program and various data are recorded so as to be readable by a computer (or CPU), or A storage device (these are referred to as “recording media”), a RAM (Random Access Memory) for expanding the program, and the like are provided. And the objective of this invention is achieved when a computer (or CPU) reads the said program from the said recording medium and runs it. As the recording medium, a “non-temporary tangible medium” such as a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. The program may be supplied to the computer via an arbitrary transmission medium (such as a communication network or a broadcast wave) that can transmit the program. The present invention can also be realized in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.
　〔まとめ〕
　本発明の態様１に係る応答装置（１）は、ユーザが入力した入力内容について応答メッセージを出力する応答装置（１）であって、第１語句と第１関連語句とが互いに関連していること、および、第２語句と第２関連語句とが互いに関連していることを記憶している記憶部（１３）と、上記入力内容に対する応答メッセージを生成する応答メッセージ生成部（２６）とを備え、上記第１語句または上記第２語句が上記入力内容に含まれており、上記記憶部（１３）が、上記第２語句と上記第１関連語句とが互いに関連していること、または、上記第１語句と上記第２語句とが同じ概念に属することを記憶している場合、上記応答メッセージ生成部（２６）は、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成する。 [Summary]
The response device (1) according to the first aspect of the present invention is a response device (1) that outputs a response message for the input content input by the user, and the first phrase and the first related phrase are related to each other. And a storage unit (13) that stores that the second word and the second related word are related to each other, and a response message generation unit (26) that generates a response message for the input content. The first phrase or the second phrase is included in the input content, and the storage unit (13) has the second phrase and the first related phrase related to each other, or When it is stored that the first word and the second word belong to the same concept, the response message generator (26) has a content in which the first word and the second related word are associated with each other. The above response message Generated.
　上記の構成によれば、応答装置は、ユーザが入力した入力内容に含まれる第１語句または第２語句、および記憶部が記憶している第１語句と互いに関連している第１関連語句および第２語句と互いに関連している第２関連語句に基づいて、第２語句と第１関連語句とが互いに関連していること、または、第１語句と第２語句とが同じ概念に属することを記憶している場合、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成することが可能となる。したがって、適合する会話シナリオが見つからない場合であっても、入力内容に含まれる語句に基づいて応答メッセージを生成する応答装置を実現するという効果を奏する。 According to the above configuration, the response device includes a first related phrase or phrase that is associated with the first phrase or the second phrase included in the input content input by the user, and the first phrase stored in the storage unit, and Based on a second related phrase that is related to the second phrase, the second phrase and the first related phrase are related to each other, or the first phrase and the second phrase belong to the same concept Is stored, it is possible to generate a response message having a content in which the first phrase and the second related phrase are associated with each other. Therefore, even when a suitable conversation scenario is not found, there is an effect of realizing a response device that generates a response message based on a phrase included in input content.
　本発明の態様２に係る応答装置（１）は、上記態様１において、上記記憶部（１３）は、上記第２語句と上記第２関連語句とがどのような関連性で関連しているかを記憶しており、上記応答メッセージ生成部（２６）は、上記第１語句と上記第２関連語句とを上記関連性で関連付けた内容の上記応答メッセージを生成する。 In the response device (1) according to aspect 2 of the present invention, in the aspect 1, the storage unit (13) determines how the second phrase and the second related phrase are related with each other. The response message generator (26) stores the response message having the content in which the first word / phrase and the second related word / phrase are associated with each other.
　上記の構成によれば、応答装置は、第２語句と第２関連語句との間の関連性に基づいて、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成することが可能となる。したがって、適合する会話シナリオが見つからない場合であっても、入力内容に含まれる語句および第２語句と第２関連語句との間の関連性に基づいて応答メッセージを生成する応答装置を実現するという効果を奏する。 According to said structure, a response apparatus produces | generates the response message of the content which linked | related the 1st phrase and the 2nd related phrase based on the relationship between a 2nd phrase and a 2nd related phrase Is possible. Therefore, even when a suitable conversation scenario is not found, a response device that generates a response message based on the phrase included in the input content and the relationship between the second phrase and the second related phrase is realized. There is an effect.
　本発明の態様３に係る応答装置（１）は、上記態様１または２のいずれかにおいて、上記記憶部（１３）が、上記第２語句と上記第１関連語句とが互いに関連していることを記憶している場合、上記応答メッセージ生成部（２６）は、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成する。 In the response device (1) according to aspect 3 of the present invention, in the aspect 1 or 2, the storage unit (13) has the second phrase and the first related phrase related to each other. When the message is stored, the response message generator (26) generates the response message having a content in which the first word / phrase and the second related word / phrase are associated with each other.
　上記の構成によれば、応答装置は、記憶部が第２語句と第１関連語句とが互いに関連していることを記憶している場合、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成することが可能となる。したがって、適合する会話シナリオが見つからない場合であっても、第２語句と第１関連語句とが互いに関連していることに基づいて、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成する応答装置を実現するという効果を奏する。 According to the above configuration, when the storage unit stores that the second word and the first related word are related to each other, the responding device associates the first word and the second related word with each other. It becomes possible to generate a response message of contents. Therefore, even if a matching conversation scenario is not found, the content of the content that associates the first phrase and the second related phrase with each other based on the fact that the second phrase and the first related phrase are related to each other. There is an effect of realizing a response device that generates a response message.
　本発明の態様４に係る応答装置（１）は、上記態様１または２のいずれかにおいて、上記記憶部（１３）が、上記第１語句と上記第２語句とが同じ概念に属することを記憶している場合、上記応答メッセージ生成部（２６）は、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成する。 In the response device (1) according to aspect 4 of the present invention, in any of the above aspects 1 or 2, the storage unit (13) stores that the first word and the second word belong to the same concept. If so, the response message generator (26) generates the response message having a content in which the first word / phrase and the second related word / phrase are associated with each other.
　上記の構成によれば、応答装置は、記憶部が第１語句と第２語句とが同じ概念に属することを記憶している場合、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成することが可能となる。したがって、適合する会話シナリオが見つからない場合であっても、第１語句と第２語句とが同じ概念に属することに基づいて、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成する応答装置を実現するという効果を奏する。 According to said structure, when the memory | storage part has memorize | stored that the 1st phrase and the 2nd phrase belong to the same concept, the response apparatus of the content which linked | related the 1st phrase and the 2nd related phrase mutually A response message can be generated. Therefore, even when a matching conversation scenario is not found, a response message having contents in which the first word and the second related word are associated with each other based on the fact that the first word and the second word belong to the same concept. This produces an effect of realizing a response device that generates.
　本発明の態様５に係る応答装置（１）は、上記態様１～４のいずれかにおいて、上記記憶部（１３）が、上記第２語句と複数の関連語句とが互いに関連していることを記憶している場合、上記応答メッセージ生成部（２６）は、上記複数の関連語句のうち上記第１関連語句と同じ概念に属する関連語句を優先的に上記第２関連語句として抽出し、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成する。 In the response device (1) according to aspect 5 of the present invention, in any one of the aspects 1 to 4, the storage unit (13) indicates that the second phrase and the plurality of related phrases are related to each other. If stored, the response message generator (26) preferentially extracts a related phrase belonging to the same concept as the first related phrase from the plurality of related phrases as the second related phrase. The response message having a content in which one word and the second related word are associated with each other is generated.
　上記の構成によれば、応答装置は、記憶部が第２語句と複数の関連語句とが互いに関連していることを記憶している場合、複数の関連語句のうち第１関連語句と同じ概念に属する関連語句を優先的に第２関連語句として抽出し、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成するという効果を奏する。したがって、適合する会話シナリオが見つからない場合であっても、第２語句と複数の関連語句とが互いに関連していることを記憶している場合、複数の関連語句のうち第１関連語句と同じ概念に属する関連語句を優先的に第２関連語句として抽出し、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成する応答装置を実現するという効果を奏する。 According to said structure, when the memory | storage part has memorize | stored that the 2nd phrase and several related phrases are mutually related, the response apparatus has the same concept as the 1st related phrase among several related phrases. The related word / phrase belonging to is preferentially extracted as the second related word / phrase, and the response message having the contents in which the first word / phrase and the second related word / phrase are associated with each other is generated. Therefore, even when a suitable conversation scenario is not found, if the second word and the plurality of related words are stored in association with each other, the same as the first related word among the plurality of related words There is an effect of realizing a response device that extracts a related phrase belonging to a concept preferentially as a second related phrase and generates a response message having a content in which the first related phrase and the second related phrase are associated with each other.
　本発明の態様６に係る応答装置（１）は、上記態様１～５のいずれかにおいて、上記応答メッセージを外部に出力する出力部（１２）と、上記入力内容が入力される直前に上記出力部（１２）から出力された応答メッセージに含まれる第３語句が属する概念と、該入力内容に含まれる入力語句が属する概念とが同じであるか否かを判定する概念判定部（２４）とを備え、上記記憶部（１３）は、上記応答メッセージのテンプレートを記憶しており、上記応答メッセージ生成部（２６）は、上記第３語句が属する概念と上記入力語句が属する概念とが同じである場合、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成し、上記第３語句が属する概念と上記入力語句が属する概念とが異なる場合、テンプレートを用いて別の応答メッセージを生成する。 The response device (1) according to Aspect 6 of the present invention is the response device (1) according to any one of Aspects 1 to 5, wherein the output unit (12) outputs the response message to the outside, and A concept determination unit (24) that determines whether the concept to which the third word included in the response message output from the unit (12) belongs and the concept to which the input word included in the input content belongs are the same. The storage unit (13) stores a template of the response message, and the response message generation unit (26) has the same concept to which the third phrase belongs and the concept to which the input phrase belongs. If there is, the response message having the contents associating the first phrase with the second related phrase is generated, and if the concept to which the third phrase belongs and the concept to which the input phrase belongs are different, It generates another response message using chromatography bets.
　上記の構成によれば、応答装置は、該応答装置が入力内容が入力される直前に出力した応答メッセージに含まれる第３語句が属する概念と、該入力内容に含まれる入力語句が属する概念とが同じである場合、第１語句と第２関連語句とを互いに関連付けた内容の応答メッセージを生成し、第３語句が属する概念と入力語句が属する概念とが異なる場合、テンプレートを用いて応答メッセージを生成することが可能となる。したがって、適合する会話シナリオが見つからない場合であっても、第３語句が属する概念および入力語句が属する概念に基づいて応答メッセージを生成する応答装置を実現するという効果を奏する。 According to the above configuration, the responding device includes a concept to which the third phrase included in the response message output immediately before the input content is input by the response device, and a concept to which the input phrase included in the input content belongs. If the two words are the same, a response message is generated with the contents of the first word and the second related word associated with each other. When the concept to which the third word belongs and the concept to which the input word belongs are different, the response message is generated using a template. Can be generated. Therefore, even when a suitable conversation scenario is not found, there is an effect of realizing a response device that generates a response message based on the concept to which the third word / phrase belongs and the concept to which the input word / phrase belongs.
　本発明の態様７に係る応答装置（１）は、上記態様６において、上記第３語句が属する概念と上記入力語句が属する概念とが異なる場合、上記応答メッセージ生成部（２６）は、上記入力語句を、上記第３語句に関連するテンプレートに組み合わせて上記別の応答メッセージを生成する。 In the response device (1) according to aspect 7 of the present invention, in the aspect 6, when the concept to which the third phrase belongs and the concept to which the input phrase belongs are different from each other, the response message generation unit (26) The phrase is combined with the template associated with the third phrase to generate the additional response message.
　上記の構成によれば、応答装置は、第３語句が属する概念と入力語句が属する概念とが異なる場合、該入力語句を、第３語句に関連するテンプレートに組み合わせて別の応答メッセージを生成することが可能となる。したがって、適合する会話シナリオが見つからない場合であっても、第３語句が属する概念と入力語句が属する概念とが異なる場合、該入力語句を、第３語句に関連するテンプレートに組み合わせて別の応答メッセージを生成する応答装置を実現するという効果を奏する。 According to the above configuration, when the concept to which the third word belongs and the concept to which the input word belongs are different from each other, the response device generates another response message by combining the input word with the template related to the third word. It becomes possible. Therefore, even when a suitable conversation scenario is not found, if the concept to which the third word belongs and the concept to which the input word belongs are different from each other, the input word is combined with a template related to the third word to obtain another response. There is an effect of realizing a response device that generates a message.
　本発明の態様８に係る応答装置（１）は、上記態様６において、上記第３語句が属する概念と上記入力語句が属する概念とが異なる場合、上記応答メッセージ生成部（２６）は、上記入力語句と、上記入力内容に含まれる第４語句とを、テンプレートに組み合わせて上記別の応答メッセージを生成する。 In the response device (1) according to the aspect 8 of the present invention, in the aspect 6, when the concept to which the third phrase belongs and the concept to which the input phrase belongs are different from each other, the response message generation unit (26) The other response message is generated by combining the phrase and the fourth phrase included in the input content with a template.
　上記の構成によれば、応答装置は、第３語句が属する概念と入力語句が属する概念とが異なる場合、該入力語句と、入力内容に含まれる第４語句とを、テンプレートに組み合わせて別の応答メッセージを生成することが可能となる。したがって、適合する会話シナリオが見つからない場合であっても、第３語句が属する概念と入力語句が属する概念とが異なる場合、該入力語句と、入力内容に含まれる第４語句とを、テンプレートに組み合わせて別の応答メッセージを生成する応答装置を実現するという効果を奏する。 According to the above configuration, when the concept to which the third word belongs and the concept to which the input word belongs are different from each other, the response device combines the input word and the fourth word included in the input content with the template to generate another A response message can be generated. Therefore, even if a matching conversation scenario is not found, if the concept to which the third word belongs and the concept to which the input word belongs are different, the input word and the fourth word included in the input content are used as a template. There is an effect of realizing a response device that generates another response message in combination.
　本発明の態様９に係る応答装置（１）は、上記態様７または８において、上記応答メッセージ生成部（２６）は、上記入力内容に含まれる複数の語句について、各語句が属する概念および各語句の前後の語句を含む文脈情報に基づいて上記入力語句を選択する。 In the response device (1) according to the aspect 9 of the present invention, in the aspect 7 or 8, the response message generation unit (26) includes the concept to which each phrase belongs and the phrases for the plurality of phrases included in the input content. The input phrase is selected based on the context information including the phrases before and after.
　上記の構成によれば、応答装置は、各語句が属する属性および各語句の前後の語句を含む文脈情報に基づいて入力語句を選択することが可能となる。したがって、適合する会話シナリオが見つからない場合であっても、各語句が属する属性および各語句の前後の語句を含む文脈情報から入力語句を選択し、第３語句に関連するテンプレートに組み合わせて別の応答メッセージを生成する応答装置を実現するという効果を奏する。 According to the above configuration, the response device can select an input phrase based on the attribute to which each phrase belongs and the context information including the phrases before and after each phrase. Therefore, even when a matching conversation scenario is not found, the input phrase is selected from the context information including the attribute to which each phrase belongs and the phrases before and after each phrase, and combined with the template related to the third phrase, There is an effect of realizing a response device that generates a response message.
　本発明の態様１０に係る応答装置（１）の制御方法は、ユーザが入力した入力内容について応答メッセージを出力する応答装置（１）の制御方法であって、上記応答装置（１）は、第１語句と第１関連語句とが互いに関連していること、および、第２語句と第２関連語句とが互いに関連していることを記憶している記憶部を備え、上記入力内容に含まれている上記第１語句または上記第２語句に関する記憶を上記記憶部から探す探索ステップ（Ｓ１７）と、上記入力内容に対する応答メッセージを生成する応答メッセージ生成ステップ（Ｓ２３）とを含み、上記第１語句または上記第２語句が上記入力内容に含まれており、上記記憶部が、上記第２語句と上記第１関連語句とが互いに関連していること、または、上記第１語句と上記第２語句とが同じ概念に属することを記憶している場合、上記応答メッセージ生成ステップ（Ｓ２３）では、上記第１語句と上記第２関連語句とを互いに関連付けた内容の応答メッセージを生成する。上記の構成によれば、態様１と同様の作用効果を奏する。 The control method of the response device (1) according to the tenth aspect of the present invention is a control method of the response device (1) that outputs a response message for the input content input by the user, and the response device (1) A storage unit storing that the one word and the first related word are related to each other and the second word and the second related word are related to each other; A search step (S17) for searching the storage unit for memory related to the first word or the second word, and a response message generation step (S23) for generating a response message for the input content. Alternatively, the second word / phrase is included in the input content, and the storage unit has the second word / phrase and the first related word / phrase related to each other, or the first word / phrase and the second word / phrase. DOO cases stores that belong to the same concept, in the response message generation step (S23), generates a response message content associated with each other and the first word and the second related phrases. According to said structure, there exists an effect similar to the aspect 1. FIG.
　本発明の各態様に係る応答装置（１）は、コンピュータによって実現してもよく、この場合には、コンピュータを上記応答装置（１）が備える各部（ソフトウェア要素）として動作させることにより上記応答装置（１）をコンピュータにて実現させる応答装置（１）の制御プログラム、およびそれを記録したコンピュータ読み取り可能な記録媒体も、本発明の範疇に入る。 The response device (1) according to each aspect of the present invention may be realized by a computer. In this case, the response device is operated by causing the computer to operate as each unit (software element) included in the response device (1). The control program of the response device (1) for realizing (1) by a computer and a computer-readable recording medium on which the control program is recorded also fall within the scope of the present invention.
　本発明は上述した各実施形態に限定されるものではなく、請求項に示した範囲で種々の変更が可能であり、異なる実施形態にそれぞれ開示された技術的手段を適宜組み合わせて得られる実施形態についても本発明の技術的範囲に含まれる。さらに、各実施形態にそれぞれ開示された技術的手段を組み合わせることにより、新しい技術的特徴を形成することができる。 The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention. Furthermore, a new technical feature can be formed by combining the technical means disclosed in each embodiment.
　１　ロボット（応答装置）　　１１　音声入力部　　１２　音声出力部（出力部）
　１３　記憶部　　２０　制御部　　２１　音声認識部　　２２　語句・概念取得部
　２３　シナリオ部　　２４　概念判定部
　２５　関連語句取得部　２６　応答メッセージ生成部　　２７　音声合成部 DESCRIPTION OF SYMBOLS 1 Robot (response apparatus) 11 Voice input part 12 Voice output part (output part)
DESCRIPTION OF SYMBOLS 13 Memory | storage part 20 Control part 21 Speech recognition part 22 Phrase and concept acquisition part 23 Scenario part 24 Concept determination part 25 Related phrase acquisition part 26 Response message generation part 27 Speech synthesis part

Claims (11)

  1. 　ユーザが入力した入力内容について応答メッセージを出力する応答装置であって、
    　第１語句と第１関連語句とが互いに関連していること、および、第２語句と第２関連語句とが互いに関連していることを記憶している記憶部と、
    　上記入力内容に対する応答メッセージを生成する応答メッセージ生成部とを備え、
    　上記第１語句または上記第２語句が上記入力内容に含まれており、上記記憶部が、上記第２語句と上記第１関連語句とが互いに関連していること、または、上記第１語句と上記第２語句とが同じ概念に属することを記憶している場合、
    　　上記応答メッセージ生成部は、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成することを特徴とする応答装置。     A response device that outputs a response message for the input content entered by the user,
    A storage unit storing that the first phrase and the first related phrase are related to each other; and that the second phrase and the second related phrase are related to each other;
    A response message generator for generating a response message for the input content,
    The first phrase or the second phrase is included in the input content, and the storage unit is configured such that the second phrase and the first related phrase are related to each other, or the first phrase If you remember that the second word belongs to the same concept,
    The response message generating unit generates the response message having a content in which the first phrase and the second related phrase are associated with each other.
  2. 　上記記憶部は、上記第２語句と上記第２関連語句とがどのような関連性で関連しているかを記憶しており、
    　上記応答メッセージ生成部は、上記第１語句と上記第２関連語句とを上記関連性で関連付けた内容の上記応答メッセージを生成することを特徴とする請求項１に記載の応答装置。     The storage unit stores in what relationship the second phrase and the second related phrase are related,
    The response device according to claim 1, wherein the response message generation unit generates the response message having a content in which the first word and the second related word are associated with each other by the relevance.
  3. 　上記記憶部が、上記第２語句と上記第１関連語句とが互いに関連していることを記憶している場合、
    　　上記応答メッセージ生成部は、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成することを特徴とする請求項１または２に記載の応答装置。     When the storage unit stores that the second phrase and the first related phrase are related to each other,
    3. The response device according to claim 1, wherein the response message generation unit generates the response message having a content in which the first phrase and the second related phrase are associated with each other.
  4. 　上記記憶部が、上記第１語句と上記第２語句とが同じ概念に属することを記憶している場合、
    　　上記応答メッセージ生成部は、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成することを特徴とする請求項１または２に記載の応答装置。     When the storage unit stores that the first word and the second word belong to the same concept,
    3. The response device according to claim 1, wherein the response message generation unit generates the response message having a content in which the first phrase and the second related phrase are associated with each other.
  5. 　上記記憶部が、上記第２語句と複数の関連語句とが互いに関連していることを記憶している場合、
    　　上記応答メッセージ生成部は、上記複数の関連語句のうち上記第１関連語句と同じ概念に属する関連語句を優先的に上記第２関連語句として抽出し、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成することを特徴とする請求項１から４のいずれか一項に記載の応答装置。     When the storage unit stores that the second word and the plurality of related words are related to each other,
    The response message generator preferentially extracts a related phrase belonging to the same concept as the first related phrase from the plurality of related phrases as the second related phrase, and the first related phrase and the second related phrase The response device according to any one of claims 1 to 4, wherein the response message having a content associated with each other is generated.
  6. 　上記応答メッセージを外部に出力する出力部と、
    　上記入力内容が入力される直前に上記出力部から出力された応答メッセージに含まれる第３語句が属する概念と、該入力内容に含まれる入力語句が属する概念とが同じであるか否かを判定する概念判定部とを備え、
    　上記記憶部は、上記応答メッセージのテンプレートを記憶しており、
    　上記応答メッセージ生成部は、
    　　上記第３語句が属する概念と上記入力語句が属する概念とが同じである場合、上記第１語句と上記第２関連語句とを互いに関連付けた内容の上記応答メッセージを生成し、
    　　上記第３語句が属する概念と上記入力語句が属する概念とが異なる場合、テンプレートを用いて別の応答メッセージを生成することを特徴とする請求項１から５のいずれか一項に記載の応答装置。     An output unit for outputting the response message to the outside;
    It is determined whether the concept to which the third word included in the response message output from the output unit immediately before the input content is input is the same as the concept to which the input word included in the input content belongs. A concept determination unit that
    The storage unit stores a template of the response message,
    The response message generator
    If the concept to which the third word belongs and the concept to which the input word belongs are the same, generate the response message with the content associating the first word and the second related word with each other,
    6. The response device according to claim 1, wherein when a concept to which the third word / phrase belongs and a concept to which the input word / phrase belong are different, another response message is generated using a template. .
  7. 　上記第３語句が属する概念と上記入力語句が属する概念とが異なる場合、上記応答メッセージ生成部は、上記入力語句を、上記第３語句に関連するテンプレートに組み合わせて上記別の応答メッセージを生成することを特徴とする請求項６に記載の応答装置。 When the concept to which the third phrase belongs and the concept to which the input phrase belongs are different, the response message generation unit generates the another response message by combining the input phrase with a template related to the third phrase. The response device according to claim 6.
  8. 　上記第３語句が属する概念と上記入力語句が属する概念とが異なる場合、上記応答メッセージ生成部は、上記入力語句と、上記入力内容に含まれる第４語句とを、テンプレートに組み合わせて上記別の応答メッセージを生成することを特徴とする請求項６に記載の応答装置。 When the concept to which the third word belongs and the concept to which the input word belongs are different from each other, the response message generating unit combines the input word and the fourth word included in the input content with a template to The response device according to claim 6, wherein a response message is generated.
  9. 　上記応答メッセージ生成部は、上記入力内容に含まれる複数の語句について、各語句が属する概念および各語句の前後の語句を含む文脈情報に基づいて上記入力語句を選択することを特徴とする請求項７または８に記載の応答装置。 The response message generation unit selects, for a plurality of words included in the input content, the input words based on a concept to which each word belongs and context information including words before and after each word. 9. The response device according to 7 or 8.
  10. 　ユーザが入力した入力内容について応答メッセージを出力する応答装置の制御方法であって、
    　上記応答装置は、第１語句と第１関連語句とが互いに関連していること、および、第２語句と第２関連語句とが互いに関連していることを記憶している記憶部を備え、
    　上記入力内容に含まれている上記第１語句または上記第２語句に関する記憶を上記記憶部から探す探索ステップと、
    　上記入力内容に対する応答メッセージを生成する応答メッセージ生成ステップとを含み、
    　上記第１語句または上記第２語句が上記入力内容に含まれており、上記記憶部が、上記第２語句と上記第１関連語句とが互いに関連していること、または、上記第１語句と上記第２語句とが同じ概念に属することを記憶している場合、
    　上記応答メッセージ生成ステップでは、上記第１語句と上記第２関連語句とを互いに関連付けた内容の応答メッセージを生成することを特徴とする応答装置の制御方法。     A control method of a response device that outputs a response message for input content input by a user,
    The response device includes a storage unit that stores that the first phrase and the first related phrase are related to each other, and that the second phrase and the second related phrase are related to each other,
    A search step for searching the storage unit for memory related to the first word or the second word included in the input content;
    A response message generation step for generating a response message for the input content,
    The first phrase or the second phrase is included in the input content, and the storage unit is configured such that the second phrase and the first related phrase are related to each other, or the first phrase If you remember that the second word belongs to the same concept,
    In the response message generation step, a response message having a content in which the first phrase and the second related phrase are associated with each other is generated.
  11. 　請求項１に記載の応答装置としてコンピュータを機能させるための制御プログラムであって、上記応答メッセージ生成部としてコンピュータを機能させるための制御プログラム。 A control program for causing a computer to function as the response device according to claim 1, wherein the control program causes the computer to function as the response message generator.
PCT/JP2017/006333 2016-06-08 2017-02-21 Responding device, method for controlling responding device, and control program WO2017212689A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
JP2018522313A JP6644141B2 (en) 2016-06-08 2017-02-21 Response device, control method of response device, and control program
CN201780035585.2A CN109313899A (en) 2016-06-08 2017-02-21 The control method of answering device and answering device, control program

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2016114838 2016-06-08
JP2016-114838 2016-06-08

Publications (1)

Publication Number Publication Date
WO2017212689A1 true WO2017212689A1 (en) 2017-12-14

Family

ID=60578456

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2017/006333 WO2017212689A1 (en) 2016-06-08 2017-02-21 Responding device, method for controlling responding device, and control program

Country Status (3)

Country Link
JP (1) JP6644141B2 (en)
CN (1) CN109313899A (en)
WO (1) WO2017212689A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020054451A1 (en) * 2018-09-13 2020-03-19 株式会社Ｎｔｔドコモ Conversation device

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI725577B (en) * 2019-10-17 2021-04-21 中國信託商業銀行股份有限公司 Intelligent voice information quality inspection method and system

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007328283A (en) * 2006-06-09 2007-12-20 Kenwood Corp Interaction system, program and interactive method
JP2013072887A (en) * 2011-09-26 2013-04-22 Toshiba Corp Interactive device

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8914885B2 (en) * 2006-11-03 2014-12-16 Alcatel Lucent Methods and apparatus for delivering control messages during a malicious attack in one or more packet networks
JP5087590B2 (en) * 2009-05-22 2012-12-05 京セラドキュメントソリューションズ株式会社 Image reading apparatus and image forming apparatus
JP2011018178A (en) * 2009-07-08 2011-01-27 Sony Corp Apparatus and method for processing information and program
JP6150282B2 (en) * 2013-06-27 2017-06-21 国立研究開発法人情報通信研究機構 Non-factoid question answering system and computer program
CN104834704A (en) * 2015-04-29 2015-08-12 深圳市梦网科技股份有限公司 Response method and response system
CN112527975B (en) * 2019-09-02 2024-07-26 科沃斯商用机器人有限公司 Man-machine interaction method and device, intelligent robot and storage medium

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007328283A (en) * 2006-06-09 2007-12-20 Kenwood Corp Interaction system, program and interactive method
JP2013072887A (en) * 2011-09-26 2013-04-22 Toshiba Corp Interactive device

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020054451A1 (en) * 2018-09-13 2020-03-19 株式会社Ｎｔｔドコモ Conversation device
JPWO2020054451A1 (en) * 2018-09-13 2021-08-30 株式会社Ｎｔｔドコモ Dialogue device
JP7166350B2 (en) 2018-09-13 2022-11-07 株式会社Ｎｔｔドコモ dialogue device

Also Published As

Publication number Publication date
JPWO2017212689A1 (en) 2019-02-14
JP6644141B2 (en) 2020-02-12
CN109313899A (en) 2019-02-05

Similar Documents

Publication Publication Date Title
US11264030B2 (en) Indicator for voice-based communications
US11740863B2 (en) Search and knowledge base question answering for a voice user interface
US10726833B2 (en) System and method for rapid customization of speech recognition models
US10056078B1 (en) Output of content based on speech-based searching and browsing requests
US11823678B2 (en) Proactive command framework
US10074369B2 (en) Voice-based communications
US10453449B2 (en) Indicator for voice-based communications
US10134388B1 (en) Word generation for speech recognition
US20200120396A1 (en) Speech recognition for localized content
US10811013B1 (en) Intent-specific automatic speech recognition result generation
US9154629B2 (en) System and method for generating personalized tag recommendations for tagging audio content
US10366690B1 (en) Speech recognition entity resolution
US9922650B1 (en) Intent-specific automatic speech recognition result generation
CN109976702A (en) A kind of audio recognition method, device and terminal
US10593319B1 (en) Parallelization of instruction steps
JP6715943B2 (en) Interactive device, interactive device control method, and control program
JP6559417B2 (en) Information processing apparatus, information processing method, dialogue system, and control program
WO2017212689A1 (en) Responding device, method for controlling responding device, and control program
WO2018045154A1 (en) Voice-based communications
WO2015040751A1 (en) Voice selection assistance device, voice selection method, and program
JP2017211610A (en) Output controller, electronic apparatus, control method of output controller, and control program of output controller
JP6260208B2 (en) Text summarization device
US11823671B1 (en) Architecture for context-augmented word embedding
JP5381211B2 (en) Spoken dialogue apparatus and program
JP2014110002A (en) Category determination device and category determination method

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2018522313

Country of ref document: JP

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17809873

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17809873

Country of ref document: EP

Kind code of ref document: A1