# US 10,770,068 B2 · Dialog agent, reply sentence generation method, and non-transitory computer-readable recording medium > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,770,068 B2 - Document ID: US10770068B2 - Status: Granted - Kind: B2 - Application number: 15/785,004 - Filed: 2017-10-16 - Published: 2018-06-07 - Granted: 2020-09-08 - Earliest claimed: 2016-12-06 - Assignees: Casio Computer Co., Ltd. - Applicants: Casio Computer Co., Ltd. - Inventors: Takahiro Tanaka - Web page: https://patentgrep.com/patent/US10770068B2 ## CPC - G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue - G06F 16/9535: Search customisation based on user profiles and personalisation - G10L 15/18: using natural language modelling - G10L 13/00: Speech synthesis; Text to speech systems - G10L 15/30: Distributed recognition, e.g. in client-server systems, for mobile phones or network applications - G10L 2015/223: {Execution procedure of a spoken command} - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract A text generating device includes an acquirer that acquires the utterance of the user relative to content provided in real time, a retriever that retrieves from a microblog server data relating to the content, a sentence generator that generates a sentence relating to the content and the acquired utterance of the user, based on the data retrieved by the retriever 140, and an voice outputter and/or a display for using the generated sentence to reply to the user. ## Claims ### Claim 1 A dialog device comprising: a processor; and a speaker or a display; the processor being configured to acquire an utterance of a user relative to content provided in real time, set a retrieving time period that is a time period of retrieval in response to a timing of the utterance of the user, continuously execute a retrieval operation to retrieve posted data from a microblog server during a period when the set retrieving time period elapses since the utterance of the user, the posted data relating to the content and being an utterance of another user who posts, during the period when the set retrieving time period elapses since the utterance of the user, an utterance related to the utterance of the user from a terminal other than the dialog device to the microblog server; generate, based on the posted data retrieved by the retrieval operation, a reply sentence that is a response to the acquired utterance of the user; and output the generated reply sentence to the user through the speaker or display, wherein the processor until a count of posted data related to the utterance of the user having given by other users since the utterance of the user is less than a threshold, measures a count of posting related to the utterance of the user given by each of the other users, updates a cumulative posting count of each of the other users by cumulating, each time the user makes an utterance, the count of posting of each of the other users having measured since start of the content, and selects, from among the other users, a user for which the posted data is to be retrieved from the microblog server, based on the updated cumulative posting count of each of the other users. ### Claim 2 (depends on 1) The dialog device according to claim 1, wherein the processor extracts a subject term of the utterance as a retrieval query from the acquired utterance of the user, and retrieves posted data including the retrieval query from the microblog server. ### Claim 3 (depends on 2) The dialog device according to claim 2, wherein the processor extracts as the retrieval query the subject term and a related term related to the subject term. ### Claim 4 (depends on 2) The dialog device according to claim 2, wherein the processor sets the retrieving time period based on an operation of the user. ### Claim 5 (depends on 4) The dialog device according to claim 4, wherein the processor when the processor is unable to retrieve the posted data including the retrieval query from among the microblog data until the retrieving period time elapses from the timing of the utterance of the user, generates a pro-forma sentence including a pro-forma expression for not imparting a sense of incongruity to the user, and outputs, instead of the reply sentence, the generated pro-forma sentence to the user through the speaker or the display. ### Claim 6 (depends on 5) The dialog device according to claim 5, wherein the processor acquires a trend term from the microblog server and generates the pro-forma sentence including the pro-forma expression and the trend term. ### Claim 7 (depends on 1) The dialog device according to claim 1, wherein the processor determines a degree of relevance between the user and the other user having posted the retrieved posted data, and selects, from among the retrieved posted data, posted data based on the determined degree of relevance and generates the reply sentence based on the selected posted data. ### Claim 8 (depends on 7) The dialog device according to claim 7, wherein the processor determines that the degree of relevance is higher when a count of postings, related to the utterance of the user, of the other user is higher. ### Claim 9 (depends on 7) The dialog device according to claim 7, wherein the processor selects the posted data based on a count of postings, related to the utterance of the user, of a plurality of the other users and the determined degree of relevance, and generates the reply sentence based on the selected posted data. ### Claim 10 (depends on 7) The dialog device according to claim 7, wherein the processor selects the posted data based on a number of comments or a number of retweets of the retrieved posted data and the determined degree of relevance, and generates the reply sentence based on the selected posted data. ### Claim 11 (depends on 1) The dialog device according to claim 1, wherein the processor selects a template for generating the reply sentence, wherein generates the reply sentence by fitting the retrieved posted data in the selected template. ### Claim 12 (depends on 11) The dialog device according to claim 11, wherein the processor sets a pseudo-personality of the dialog device, and selects the template in accordance with the pseudo-personality of the dialog device. ### Claim 13 (depends on 1) The dialog device according to claim 1, wherein the dialog device receives cooperative information from a broadcast source of the content provided in real time, and when the utterance of the user relates to the content, the processor generates the reply sentence based on the cooperative information and the posted data. ### Claim 14 (depends on 13) The dialog device according to claim 13, wherein the broadcast source and the dialog device are interconnected via a communication line, and the processor receives the cooperative information via the communication line. ### Claim 15 (depends on 13) The dialog device according to claim 13, wherein the broadcast source and the dialog device are not interconnected, the broadcasted content includes message information for the dialog device, the message information being unintelligible to the user, and the processor receives the cooperative information by detecting the message information. ### Claim 16 A reply sentence generation method comprising: acquiring an utterance of a user relative to content provided in real time; setting a retrieving time period that is a time period of retrieval in response to a timing of the utterance of the user; connecting to a microblog server; continuously executing a retrieval operation to retrieve posted data from the microblog server during a period when the set retrieving time period elapses since the utterance of the user, the posted data relating to the content and being an utterance of another user who posts, during the period when the set retrieving time period elapses since the utterance of the user, an utterance related to the utterance of the user from a terminal other than the dialog device to the microblog server; generating, based on the posted data retrieved by the retrieval operation, a reply sentence that is a response to the acquired utterance of the user; until a count of posted data related to the utterance of the user having given by other users since the utterance of the user is less than a threshold, measuring a count of posting related to the utterance of the user given by each of the other users; updating a cumulative posting count of each of the other users by cumulating, each time the user makes an utterance, the count of posting of each of the other users having measured since start of the content; and selecting, from among the other users, a user for which the posted data is to be retrieved from the microblog server, based on the updated cumulative posting count of each of the other users. ### Claim 17 A non-transitory computer-readable recording medium storing a program for causing a computer to function as: an acquirer that acquires the utterance of a user relative to content provided in real time; a setter that sets a retrieving time period that is a time period of retrieval in response to a timing of the utterance of the user; a retriever that continuously executes a retrieval operation to retrieve posted data from a microblog server during a period when the set retrieving time period elapses since the utterance of the user, the posted data relating to the content and being an utterance of another user who posts, during the period when the set retrieving time period elapses since the utterance of the user, an utterance related to the utterance of the user from a terminal other than the dialog device to the microblog server; a sentence generator that generates, based on the retrieved posted data retrieved by the retriever, a reply sentence that is a response to the acquired utterance of the user; and an evaluator that until a count of posted data related to the utterance of the user having given by other users since the utterance of the user is less than a threshold, measures a count of posting related to the utterance of the user given by each of the other users; updates a cumulative posting count of each of the other users by cumulating, each time the user makes an utterance, the count of posting of each of the other users having measured since start of the content; and selects, from among the other users, a user for which the posted data is to be retrieved from the microblog server, based on the updated cumulative posting count of each of the other users.