WO2016104736A1 - コミュニケーション提供システム及びコミュニケーション提供方法 - Google Patents
コミュニケーション提供システム及びコミュニケーション提供方法 Download PDFInfo
- Publication number
- WO2016104736A1 WO2016104736A1 PCT/JP2015/086309 JP2015086309W WO2016104736A1 WO 2016104736 A1 WO2016104736 A1 WO 2016104736A1 JP 2015086309 W JP2015086309 W JP 2015086309W WO 2016104736 A1 WO2016104736 A1 WO 2016104736A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- user
- answer
- question
- data
- database
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
-
- 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
Definitions
- the present invention relates to an interactive system for automatically communicating with a user who has accessed through a telecommunication line such as the Internet.
- Social networking services can overcome time and distance obstacles and share a large amount of information with each other in real time, but each user has private time. In reality, it is not always possible to exchange information in real time.
- users of social networking services can simultaneously send information to multiple people associated as friends, but it is difficult to communicate with each of multiple friends individually and simultaneously. is there.
- an automatic interactive system that stores document data of answers to expected questions on a recording medium and outputs answers in response to user questions (see Patent Documents 1 and 2). ).
- answer documents since answer documents are prepared in advance, the dialog is uniform and it is impossible to have a conversation according to the social situation that changes every moment.
- one embodiment of the present invention has an object to provide a communication system and a communication method capable of expressing individual personality and personality of a user.
- a first user's social data registered with one or more social networking services and a plurality of uses registered with multiple social networking services.
- a database in which information of a user's social data and a plurality of text data collected from a plurality of social networking services is stored, and a second user's question for the first user is stored in the database.
- a communication providing system including an information processing module having a function as an artificial intelligence that infers or learns an answer to a question and determines based on at least a part of the information being asked.
- the database stores a first database that stores social data of a first user, a second database that stores social data of a plurality of users, and a plurality of text data.
- the third database may be hierarchized.
- an information processing module having a function as an artificial intelligence sends an appropriate answer to a question from data recorded in the first database, the second database, and the third database. Inference may be made based on at least part of the information recorded in the upper database.
- an information processing module having a function as artificial intelligence performs trend analysis on answers to questions from social data of a plurality of users stored in the first database, and infers answers to the questions. Or may be determined by learning.
- the trend analysis may analyze temporal changes in answers of a plurality of users to a question.
- the first user's social data and the plurality of users' social data are text data and text data generated from one or more of audio data, photo data, and video data. Things are included.
- an information processing module including a similar database in which similar questions are grouped as similar questions and an answer corresponding to the similar question is associated, and having a function as an artificial intelligence is obtained from the similar database. An answer to the question may be selected.
- the similar database may include a similar question database that groups and records similar question contents and a similar answer database that stores answers corresponding to the similar questions.
- the similar database includes a first similar database in which data categorizing the contents of a question and an answer among social data of the first user, and social data of a plurality of users.
- a second similar database that stores data that classifies the contents of questions and answers from data
- a third similar database that stores data that classifies the contents of questions and answers from text data; It may be hierarchized.
- an evaluation module for acquiring information evaluated by the second user and attaching an evaluation value to the response may be included.
- the evaluation module may include a notification module that notifies the first user of the evaluation result of the answer.
- a first user who has received a notification from the notification module edits the content of the answer to the question, and the content of the answer to the question based on the content edited by the editing module. And an update module to be updated.
- the evaluation module assigns a priority level based on the evaluation result, and the information processing module having a function as artificial intelligence infers or learns an appropriate answer to the question based on the priority. May be determined.
- the database may have a function of updating and accumulating the first user's social data, the second user's social data, and text data over time.
- an image data generation module that generates three-dimensional image data of the first user may be included.
- the database includes the first user's voice data, and generates a phoneme data generation module that generates phoneme data from the voice data, and a voice generation module that generates voice of conversation using the phoneme data. And may be included.
- an information processing module having a function as artificial intelligence may have a function of analyzing a frequency of answers to a question and recording a question corresponding to a high-frequency answer in a similar database. Good.
- an information processing module having a function as artificial intelligence has a function of parsing text data included in personal social data and inferring or learning an answer to a question. It may be.
- a question that a virtual personal image of the first user is generated based on the registration information of the first user and is transmitted to the virtual individual by the second user.
- the answer to the question is determined by inferring or learning based on one type of social data of the first user, social data of a plurality of users or a plurality of text data registered in advance,
- a communication providing method for providing the determined answer to the second user is provided.
- the first user's social data is searched for an answer to the question, and when an appropriate answer cannot be obtained, the second is searched for a plurality of users' social data.
- a plurality of social data registered in advance may be searched.
- the answer to the question may be determined by trend analysis of the answer to the question from the social data of a plurality of users, and inferring or learning the answer to the question.
- the trend analysis may analyze temporal changes in answers of a plurality of users to a question.
- the first user's social data and the plurality of users' social data may be text data, text data generated from audio data, photo data, and video data.
- similar questions are grouped as similar questions, answers corresponding to similar questions are stored in a similar database in association with each other, and answers to the questions are selected from the similar database. May be.
- data similar to the contents of questions and answers is stored in the similar database from the social data of the first user, and the questions and answers are stored from the social data of a plurality of users. May be stored, and data in which the contents of questions and answers are categorized from text data may be stored.
- information evaluated by the second user may be acquired and an evaluation value may be attached to the answer.
- the evaluation result of the answer may be notified to the first user.
- the first user who receives the notification from the notification module receives the edited content of the answer to the question and updates the content of the answer to the question based on the edited content. It may be.
- a priority level may be assigned based on the evaluation result, and an appropriate answer to the question may be inferred or learned based on the priority order.
- the first user's social data, the second user's social data, and text data may be updated and accumulated over time.
- three-dimensional image data of the first user may be generated.
- the voice data of the first user may be stored, phoneme data may be generated from the voice data, and a conversational voice may be generated using the phoneme data.
- the frequency of answers to a question may be analyzed, and a question corresponding to a high-frequency answer may be recorded in a similar database.
- text data included in personal social data may be parsed and an answer to a question may be inferred or learned to be determined.
- AI artificial intelligence learns a user's thoughts by acquiring data from various social networking services used by the user and storing it in a database.
- a system that provides appropriate answers to questions can be provided.
- FIG. 1 It is a figure which shows the relationship between the communication provision system which concerns on one Embodiment of this invention, a social networking service, and the user who utilizes this communication provision system.
- the communication provision system which concerns on one Embodiment of this invention it is an example figure of the screen for asking a question.
- the communication provision system which concerns on one Embodiment of this invention it is an example figure of the screen for asking a question.
- the communication provision system which concerns on one Embodiment of this invention it is an example figure of the screen by which an answer is provided.
- a plurality of second when a virtual first user answers a question asked by a plurality of second users at a terminal of the first user, a plurality of second It is a figure explaining the function to set automatic answer propriety for every user.
- the communication provision system which concerns on one Embodiment of this invention it is a flowchart explaining the flow of the process which judges whether a reply is possible for every use which asked a question, and answers. It is a figure which shows the functional structure of the communication provision system which concerns on this embodiment. It is a figure explaining the structure of the communication provision system which concerns on one Embodiment of this invention. It is a figure explaining the structure of the communication provision system which concerns on one Embodiment of this invention.
- the communication provision system which concerns on one Embodiment of this invention WHEREIN: The aspect which integrates a context by grouping the question and answer in conversation is shown.
- the communication provision system which concerns on one Embodiment of this invention it is a figure which illustrates what kind of keyword order is important when seeking a match.
- a question and an answer show the aspect clustered by the similarity of the answer, and it is a figure which shows the aspect that each cluster comprises one group.
- it is a figure explaining the vector (keyword, concept, type) which comprises each question has the prediction factor which estimates an equivalent answer vector. It is a figure explaining the structure of the communication provision system which concerns on one Embodiment of this invention.
- FIG. 1A shows a relationship between a communication providing system according to an embodiment of the present invention, a social networking service (hereinafter also referred to as “SNS”), and a user who uses this communication providing system.
- the communication providing system 100 according to the present embodiment has a function of acquiring data from various SNS used by a user and storing it in a database.
- the communication providing system 100 further has a function (function as an artificial intelligence) for inferring and learning a user's thought based on information stored in the database, and can provide an appropriate answer to the question. Details will be described below.
- the communication providing system 100 is placed in a state where it can be accessed with a plurality of users 202 through a telecommunication line.
- a plurality of users 202 can communicate with each other directly and indirectly by a service provided by the communication providing system 100.
- FIG. 1A shows that a plurality of users 202 are placed in a state where each user terminal 200 can communicate with the communication providing system 100 and both through an electric communication line.
- the plurality of users 202 is preferably a set of users registered in advance to receive services provided via the communication providing system 100.
- the safety and reliability of communication can be improved.
- the plurality of users 202 may be registered in advance in order to receive a service provided via the communication providing system 100.
- at least one of the plurality of users 202 may be able to use the communication providing system 100 as an anonymous user.
- the communication providing system 100 is placed in a state where it can communicate with one or a plurality of SNSs 300 provided to the public.
- SNSs 300 include Twitter (registered trademark), Facebook (registered trademark), and various communication services provided on the Internet.
- communication modes provided by the SNS are various and are not limited to the services exemplified above.
- the SNS 300 that cooperates with the communication providing system 100 only needs to be provided with social data of the user, and preferably has a mode in which information can be shared among a plurality of users.
- social data refers to characters (including emoticons) that a user sends to a specific person (for example, another user registered as a friend) or an unspecified person through some kind of SNS. ), Information such as symbols, sounds, still images, and moving images.
- personal social data refers to social data of a specific individual among a plurality of users.
- the communication providing system 100 acquires social data stored in each from at least one, preferably a plurality of SNSs 300.
- the acquisition of social data performed by the communication providing system 100 may be executed in real time, or may be performed in a timely manner or at regular intervals.
- the communication providing system 100 accumulates the acquired social data in a database.
- the communication providing system 100 preferably acquires social data in cooperation with a plurality of SNSs 300.
- the first user 204 registers the attributes of the first user 204 in the communication providing system 100 using a personal computer, a smartphone, or the like.
- the first user 204 displays the screen shown in “Welcome personalized AI” using a browser or an application, and clicks the “Register” button 1.
- Push a screen for inputting the attributes of the first user 204 is displayed.
- information SNS specifying information
- Authentication information may be included.
- SNS identification information is input in field 2
- a user name and a password are input in fields 3 and 4 as authentication information.
- the “Register” button 5 is pressed.
- the attributes of the first user 204 are stored in a storage device of the communication providing system 100, for example, a secondary storage that stores a database or the like.
- the communication providing system 100 uses the SNS specifying information and the authentication information of the SNS to access the SNS specified by the SNS specifying information, and logs in to the SNS using the authentication information of the first user 204. Thereafter, the communication providing system 100 acquires information on the first user 204 stored in the SNS, that is, social data.
- Information related to the first user 204 includes messages exchanged by the first user 204 with a specific user in the SNS, articles, photos, audio information, and videos posted by the first user 204 on the SNS. (Which may include audio information).
- the communication providing system 100 stores information on the first user 204 acquired from the SNS in a database.
- the communication providing system 100 can directly use the information regarding the first user 204 acquired from the SNS as a database.
- the communication provision system 100 may convert the information regarding the 1st user 204 into another information according to the attribute and content of the acquired information, and may accumulate
- a question from another user and the first user 204 corresponding thereto. can be stored in the database in association with each other.
- the database is composed of a plurality of areas (such as sub-databases or tables), and can include (1) an area for storing questions and (2) an area for storing answers. Therefore, a question from another user is accumulated in the area where the question is accumulated, and the answer of the first user 204 is accumulated in the area where the answer is accumulated. Further, a question from another user and the answer of the first user 204 are associated with each other, and for example, information for association is stored in the third area. For example, a set of a primary key for a question and a primary key for an answer is stored in the third area.
- the communication providing system 100 acquires the first information stored in the database when all the information related to the first user 204 is acquired from the SNS specified by the SNS specifying information and is not stored in the database.
- the ratio of all information related to the user 204 may be presented to the first user 204.
- the first user 204 can know the degree of information accumulation.
- the communication providing system 100 responds to a question from one of the plurality of users 202 by the first user 204. It is possible to decide whether or not to reply on behalf of In addition, it is possible to predict the appropriateness of an answer to a question by presenting the degree of accumulation of such information to a plurality of users 202.
- the communication providing system 100 answers questions from other users using the information stored in the database.
- the first user 204 may be able to ask the communication providing system 100 some questions. After seeing the answers to the questions of some attempts, the “Publish” buttons 7 and 8 may be pressed.
- the communication providing system 100 can reflect information individually set by the user in each SNS 300. For example, when a user who uses a certain SNS 300 sets his / her personal profile private and sets a message posted to the SNS to be public, a range of disclosure or private / public is set for each message. In this case, the communication providing system 100 can take over the setting.
- the disclosure range of the information is also stored in the database. For example, as shown in FIG. 1F, if the one-to-one message exchange 301-1 with the user 207 in the SNS 6 (301), the question and answer by the message exchange 301-1 are sent to the user 207. Only disclosed. In other words, if the question made by the user A other than the user 207 using the communication providing system 100 appears only when the first user 204 exchanges a one-to-one message with the user 207. The answer to the question is not provided to the user A.
- the answer is not provided.
- the answer is sent to the user 208 belonging to the group 3010-2.
- the first user 204 makes an answer to a question in a state where anyone can see it in the SNS, no matter which user makes the question, the answer is sent to the communication providing system 100.
- the first user 204 answers a question under a state (301-3) that can be viewed by any user belonging to the SNS 6 (301), the first user 204 If a similar question is received from the user 209 belonging to the same SNS 6, the answer is provided.
- the plurality of users 202 to whom the service is provided from the communication providing system 100 are users of one or a plurality of SNSs 300 at the same time. This is because the communication providing system 100 acquires social data of each of the plurality of users 202 from the SNS 300 provided to the public. That is, a plurality of users 202 can receive service provision via the communication providing system 100 by participating in the community of the SNS 300.
- FIG. 1A shows a mode in which the second user 206 interacts with the first user 204 via the communication providing system 100.
- the first user 204 and the second user 206 do not directly exchange messages, but the virtual user generated by the communication providing system 100.
- a typical first user 204b and a real second user 206 interact with each other.
- the communication providing system 100 answers as the virtual first user 204b on behalf of the first user 204. It can be carried out. Thereby, the effort of the 1st user 204 can be reduced.
- the first user 204 when the first user 204 is a celebrity, even if many users ask the same question, the first user 204 can be prevented from becoming a burden. Even if a special question is asked, if the first user 204 has answered such a question in the past, the first user 204 does not look for a past answer, The communication providing system 100 can make an answer as the virtual first user 204b. This can also reduce the effort of the first user 204.
- An example of dialogue in this case is as follows.
- the second user 206 accesses the communication providing system 100, logs in if necessary, and then specifies the other party (first user 204b) who wants to interact.
- the other party first user 204b
- FIG. 1G a screen for asking a question is displayed, and a “login” button 9 is pressed to log in.
- the “anonymous” button 10 is pressed.
- a list of acquaintances in the communication providing system 100 of the second user 206 can be displayed to specify a partner with whom a conversation is desired.
- the user terminal 200 transmits a question to the other party (first user 204b).
- the question is received by the communication providing system 100 by inputting the question in the field 12 and pressing the enter key or the like.
- the communication providing system 100 that has received the question transmits an appropriate answer to the question to the user terminal 200. For example, if a question “Who are you?” Is input in the field 12, an answer “I am the president” is displayed in the field 14, as shown in FIG. 1J.
- the face photograph data 13 of the first user 204 may be displayed.
- the face photograph data 13 may be data generated by three-dimensional data, as will be described later.
- the response transmitted by the communication providing system 100 is generated based on the personal social data of the first user 204 as a first example.
- the communication providing system 100 cannot find an appropriate answer from personal social data, the communication providing system 100 generates an appropriate answer by referring to social data of a plurality of users and responds.
- the response of the virtual first user 204b is extremely fast with respect to the time required for the second user 206 to ask a question (time for operating the user terminal 200). It becomes. Therefore, a plurality of users 202 can interact with the virtual first user 204b at the same time as described above.
- the virtual first user 204b generated based on the personal social data of the first user 204 is embodied via the communication providing system 100. From another point of view, it can be considered that the virtual first user 204b exists on the computer program or application program executed by the communication providing system 100. Alternatively, the virtual first user 204b can be regarded as an entity embodied by hardware resources constituting the communication providing system 100 and software resources executed on the hardware resources.
- FIG. 2 shows a functional configuration of the communication providing system 100 according to the present embodiment.
- the communication providing system 100 includes an information processing module 104 having a function as artificial intelligence and a database 102.
- the communication providing system 100 shown in FIG. 2 has a database 102 realized by a memory module or storage device such as a hard disk, a semiconductor memory, and a magnetic memory, and a central processing unit (CPU) or an arithmetic processing circuit having an equivalent function.
- the information processing module 104 can be realized by a device realized by the above.
- the information processing module can be regarded as a functional block realized by hardware resources or hardware resources and software resources, and is sometimes called an information processing unit or information processing means.
- the artificial intelligence means that intelligent functions such as inference and judgment are realized by using hardware resources and software resources, and may be recognized as a concept including a database for storing data as knowledge.
- Artificial intelligence also has a learning function, which may have the ability to predict the future from past information (data).
- the information processing module 104 having a function as artificial intelligence includes at least one having a function as normal artificial intelligence as described above.
- the database 102 has at least an area for storing social data.
- FIG. 2 shows an aspect in which the database 102 includes a first database 102 a and a second database 102 b that store social data acquired from the SNS 300.
- the third database 102c may include social data, or words, vocabularies, and fixed phrases used in communication may be stored in advance.
- the first database 102a, the second database 102b, and the third database 102c are stored in association with the contents of the dialogue included in the social data, such as questions and answers (or questions and responses). Moreover, social data transmitted unilaterally like a personal tweet (tweet) may be included.
- tweets are often not answers to questions.
- the tweet is parsed and decomposed into adverb phrases representing the subject, object, place, time, and form, and adverb phrases representing the subject, object, place, time, and form, etc.
- Information about the first user 204 may be accumulated in the database. For example, suppose the first user 204 tweeted, “My sister signed a contract to purchase a car yesterday.” At this time, the subject is “my sister”, the object is “car” and “contract”, and the adverb phrase representing time is “yesterday”.
- the similar database 106 stores similar questions as one group, and answers to the group are associated with each other.
- the similar database 106 may be divided into a first similar database 106a, a second similar database 106b, and a third similar database 106c corresponding to the first to third databases described above.
- the information processing module 104 having a function as an artificial intelligence performs a function as an artificial intelligence in cooperation with the database 102 and the similar database 106, and also functions for editing a message (answer information, etc.) and a message (answer information, etc.). It includes an evaluation function, an evaluation result notification function, a phoneme generation function for reproducing the user's voice, a question generation function for generating a new question, and a 3D imaging function for generating a 3D video of the user.
- the communication providing system 100 is placed in a state capable of bidirectional communication with a plurality of users (user terminals 202a and 202b illustrated in FIG. 2).
- the user terminals 202a and 202b can communicate with the communication providing system 100 and the SNS 300.
- the user terminals 202a and 202b can communicate with a plurality of user terminals 202a and 202b. It is in an available state.
- the communication providing system 100 can generate a virtual first user 204b for the first user 204 on the system, and can have a conversation with a plurality of other users sequentially or simultaneously. is there. In this case, the virtual first user 204b can have an equal conversation with all accessing users.
- the first user 204 among the plurality of second users 202, depending on closeness (family, friendship), organizational relationship (whether they belong to the same corporation or group), etc. There is a request to limit the scope of responses. That is, the first user 204 may want to reserve an answer to the question in relation to the second user 206.
- FIG. 1K shows an example of a screen display of a user terminal operated by the first user 204.
- the screen display shown in FIG. 1K indicates whether or not the first user 204 may make an automatic answer for each registered user to receive the service provided via the communication providing system 100.
- the stage to set up is shown.
- a display 210 for identifying a user based on a user name, a thumbnail image of the user, and the like, and an answer availability selection switch 211 are shown.
- the first user 204 can set automatic answer permission for each user while viewing such a screen display. Note that such settings can be changed in a timely manner, and the first user 204 can change the settings by operating the selection switch 211 for timely answer availability.
- FIG. 1L shows the operation of the communication providing system 100 in the case where the answer permission / inhibition is individually set for the plurality of second users 202.
- the answer permission / inhibition is individually set for the plurality of second users 202.
- S31 a question is made by the second user 206
- the user is identified and an automatic answer is judged (S32).
- S32 an automatic answer is judged
- the virtual first user 204b generates an answer (S33)
- An answer is made to the user 206 (S34).
- the virtual first user 204b holds the answer. Become.
- the communication providing system 100 can provide a function in which the first user 204 individually sets whether or not to answer a plurality of second users 202.
- the communication providing system 100 has a function of setting whether or not to generate an answer in the information processing module 104 for the second usage question.
- the range of communication by the virtual first user 204b can be set.
- communication with a specific user can be restricted for the first user 204.
- the virtual first user 204b is a plurality of second users. Since it is not necessary to answer all the questions 202 and all the questions, the load on the system can be reduced.
- FIG. 3A is a diagram illustrating the configuration of the database 102 in the communication providing system 100 according to an embodiment of the present invention.
- the database 102 may be constructed in a plurality of layers.
- the first database 102a is a database corresponding to each user, and stores personal social data.
- the personal social data stored in the first database 102a is also used as basic data when generating virtual user data in the communication providing system 100. Therefore, personal social data and information generated therefrom are stored in the database for each user.
- the information generated from personal social data includes vectors obtained by analyzing personal social data using keywords, vectors obtained by analyzing personal social data using concepts, and personal social data, depending on the type. One or more of the analyzed vector and the vector obtained by analyzing the personal social data by the group are included.
- the data stored in the first database 102a includes personal social data acquired from a plurality of SNSs 300. For example, information such as a comment posted by a certain user on the SNS, communication contents exchanged on the SNS between the user and other users, and tweets (tweets) of the user are stored. Yes. Also, as described above, questions and answers can be generated and stored.
- the second database 102b stores social data of a plurality of users.
- the data stored in the second database 102b includes social data of each user acquired from the plurality of SNSs 300.
- the second database 102b stores a larger amount of data than the first database 102a.
- the second database 102b is hierarchically positioned below the first database 102a.
- the data stored in the second database 102b can be changed based on the change in the social data stored in the first database 102a. For example, when it is detected that a large number of users change the length, style, and contents of answers to similar questions, the data stored in the second database 102b can be changed according to the change. Is possible. That is, as shown in FIG. 3B, it is possible to perform a trend analysis on the first database 102a and update the second database 102b. As an example of such a case, in response to the question "What kind of person are you?" In contrast to responding, recently, when changes such as “I am a company employee” are detected that responds to occupations, the second database is responded to this change. The data stored in 102b may be changed so that the occupation is answered in response to the question "What kind of person are you?" Such a change may be referred to as “analyzing answers to questions”.
- the third database 102c stores a plurality of text data.
- the data stored in the third database 102c may be any text data set in advance, or may be a collection of texts that appear frequently among texts that appear on the SNS. .
- texts having a high frequency of appearance may be collected from real-time communication called chat that is exchanged in SNS.
- the first user's social data stored in the first database 102a includes real-time communication data called chat, tweets uploaded to Twitter (registered trademark), Facebook, etc.
- the uploaded comments and comments of other users with respect to the comments are stored.
- the social data of a plurality of users stored in the second database 102b may include data similar to the above, and further may store a question and text data of an answer to the question in association with each other.
- a plurality of text data may be stored in association with each other in the form of a question and an answer to the question.
- the social data stored in the first database 102a and the second database 102b is preferably text data.
- this text data may include not only text data created by the SNS user but also text data generated from audio data, photo data, and video data.
- audio data included in audio data and video data can be converted into text and stored in a database as text data.
- a comment may be attached at the same time, and another user's comment may be attached to the comment.
- a photo location, a content, or a content is added to a photo, a moving image, or a sound as a tag. Therefore, in generating text data from photo data and video data, data such as tags and comments attached thereto can be stored as text data in a database.
- position information country name, place name, etc.
- date information from which photo data and video data are acquired can be stored in a database as text data.
- text data can be generated and stored in a database based on one or more of audio data, photo data, and video data.
- Such a database 102 operates in cooperation with or in cooperation with the information processing module 104 having a function as artificial intelligence, as described with reference to FIG. Since the user's personal social data is stored in the first database 102a, what uses this hierarchy may be expressed as “private AI” or “personalized AI” as artificial intelligence reflecting the individual. it can. Further, since the second database stores social data of the entire user, what is realized using this hierarchy can also be expressed as “everyone AI” or “common sense AI”.
- the information processing module 104 having a function as artificial intelligence generates a virtual individual using social data stored in the database 102 and communicates with a real user.
- the information processing module 104 having a function as artificial intelligence has, for example, a function of recognizing text data stored in the database 102 and creating or generating an answer to a question.
- the information processing module 104 having a function as artificial intelligence has a function of inferring or learning an appropriate answer to a question from text data stored in the database 102 and determining it.
- the information processing module 104 having a function as artificial intelligence receives a question of the second user 206 with respect to the virtual first user 204b, and creates or generates an answer to the question. At this time, the information processing module 104 having a function as artificial intelligence performs processing for obtaining answers in order from the upper hierarchy of the database 102.
- FIG. 4 is a flowchart showing an example of a process in which the information processing module 104 having a function as artificial intelligence searches the database 102 and creates an answer to a question.
- the flow of processing will be described with reference to FIG.
- an information processing module 104 having a function as artificial intelligence creates or generates an answer to a question
- an appropriate answer to the question is searched from the first database 102a (S01). If an answer is obtained from the first database 102a, it will most accurately reflect the thoughts, thoughts, feelings, etc. of the first user 204 (S06).
- the question stored in the first database 102a is searched for a question that matches the question as a character string.
- the answers associated with the asked questions can be retrieved.
- a search is performed by determining whether or not they match in consideration of a certain degree of notation fluctuation. You can also. It is also possible to search for a question having the largest number of words included in the question among the questions stored in the first database 102a. In this case, in addition to the number of words, the search can be performed in consideration of the order in which the words are arranged.
- Semantic analysis can include, for example, deriving a logical outcome from an answer as a conclusion or deriving a condition assumed by the answer based on a predetermined logic system.
- Semantic analysis can include, for example, deriving a logical outcome from an answer as a conclusion or deriving a condition assumed by the answer based on a predetermined logic system.
- an answer is detected, if it includes a case where the character string is not completely matched, there is a case where the answer is inferred and determined.
- there is a complete match as a character string there is a case of “learning and determining an answer”.
- the information processing module 104 having the function as artificial intelligence searches the second database 102b in the lower hierarchy (S03).
- the answers obtained from the second database 102b can know the tendency of how the majority of users are answering a specific question, and can obtain an average and appropriate answer. For example, "who are you” In response to the question “I am XX (name).”
- the information processing module 104 having a function as artificial intelligence can determine a similar answer as an answer to the question.
- the above is an example of a simple question. However, based on the second database 102b, the response tendency of a large number of users is reflected, so that it is possible to obtain an answer in time (S06).
- the information processing module 104 searches the third database 102c (S05). Since an enormous number of text data is stored in the third database 102c, an answer can be selected from the stored text data. Then, the selected one is set as an answer (S06).
- the database 102 stores a question and a response to the question in association with each other.
- questions (questions) and responses (responses) to communication in human society are usually not uniform. For example, when trying to ask the name of a person you meet for the first time, you may ask “Who are you?” Or “Please tell me your name”.
- a mechanism for selecting a response to the corresponding question is provided. For example, in response to the above-mentioned questions “Who are you?” And “Tell me your name”, there is no sense of incongruity even if you reply “I am Annie.”
- This mechanism may be constructed as the similar database 106 as described in FIG.
- FIG. 5 shows an example of the similar database 106.
- the similar database 106 includes a question database 107 that stores questions and a response database 108 that stores responses. Alternatively, an area for storing a question and an area for storing a response may be provided in the similar database 106.
- the question database 107 stores a plurality of questions as data.
- the plurality of questions are stored as one group in association with similar questions. For example, the above-mentioned “Who are you?” And “Tell me your name” are associated and stored as similar questions.
- the response database 108 stores the response content to the question, that is, a plurality of responses as data. For example, the above-mentioned reply “I am Annie” is stored.
- the response data is associated with a specific question. According to the example above, the questions “Who are you?” And “Tell me your name” are grouped together and the response is associated with the response “I am Annie.” ing.
- FIG. 6A shows another aspect of the similar database 106.
- FIG. 6A shows an aspect in which the similar database 106 is hierarchized.
- the similar database 106 includes a first similar database 106a corresponding to the first database 102a and a second similar database corresponding to the second database 102b.
- the database 106b and the third similar database 106c corresponding to the third database 102c may be hierarchized.
- the first similar database 106a is created for each specific user based on the personal social data of the user.
- Text data corresponding to a question sentence and text data corresponding to a reply sentence are stored from text data by chat or the like included in the personal social data or the contents of the textized conversation.
- similar question sentence data is grouped and stored in the question database, and the corresponding reply sentence data is stored in the response database and associated. It has been.
- the second similar database 106b is created based on social data of a plurality of users. Text data corresponding to a question sentence and text data corresponding to a reply sentence are stored from text data by chat or the like included in social data of a plurality of users, or from the contents of a textized conversation. In the second similar database 106b, as described with reference to FIG. 5, similar text data corresponding to the question text is grouped and stored in the question database, and response text data corresponding thereto is stored in the response database. Stored and associated.
- the data stored in the second similar database 106b can be changed based on the change in the social data stored in the first similar database 106a. That is, as shown in FIG. 6B, the trend analysis of the first similar database 106a can be performed, and the second similar database 106b can be updated using the result. For example, suppose that there are cases where a large number of users respond to nominations for each of the questions “Who are you?” And “Tell me your name”.
- the second similar database 106b may be updated so that both the question “Who are you?” And “Tell me your name” are similar questions. it can.
- the third similar database 106c stores a plurality of text data.
- the plurality of text data may be prepared as a question text and text data of a response text to the question text. These text data may be collected text data having a high appearance frequency among text data appearing on the SNS.
- similar text data corresponding to the question text is grouped and stored in the question database, and response text data corresponding thereto is stored in the response database. Stored and associated.
- the similar database 106 includes the first similar database in which the data categorizing the contents of the question and the answer among the social data of the first user and the social data of a plurality of users are stored. Is divided into a second similar database that stores data that classifies the contents of questions and answers, and a third similar database that stores data that classifies the contents of questions and answers from text data It is preferable that
- the information processing module 104 having a function as artificial intelligence refers to the second database 102b and there are many identical contents in answers from many users.
- a question item for the answer may be automatically added to the similar database 106 and updated. In this way, by updating the similar database with reference to the social data of multiple users, the combination of questions and answers that are determined to be similar can be more accurately and accurately, and the accuracy of communication can be improved. it can.
- the second user 206 recognizes the voice by speaking with the voice, and the voice of the first user 204 (or the pseudo user) A function of responding with the voice of the person himself / herself.
- the second user 206 makes a non-voice question using text information or the like, it has a function of responding with the voice of the first user 204 (or a pseudo person's voice). It may be.
- a voice response method a method can be used in which a large number of voices of the first user 204 are recorded in advance and a response sentence is created as voice data using the voice.
- such a method forces the user himself / herself to input a large amount of audio data in advance, which is not preferable for all users and impairs the user's convenience.
- communication data by voice of the user performed via the SNS is recorded as needed, and phoneme data is created based on the data. For example, a word spoken by a user in voice communication is taken out and a set of a plurality of words is created.
- this word set is stored as data for creating phonemes.
- a feature (waveform) of voice data prepared in advance and an actual user's voice waveform are compared in accordance with user attributes (for example, sex, age, etc.). Then, the waveform of the voice data prepared in advance is adjusted so as to approach the waveform of the user's voice. The adjusted voice data is used as the phoneme data of the user.
- a method in which a large number of voices of the first user 204 are recorded in advance and a response sentence is created as voice data using the voices
- the following may be possible. That is, when a large number of voices of the first user 204 are recorded in advance, it is assumed that voice data is acquired by causing the first user 204 to read a predetermined sentence. At this time, communication data by voice of the user performed via the SNS is recorded at any time, and voice generated by the first user 204 is recognized. When a predetermined sentence or a part of a word constituting a part thereof is recognized, the voice data of the recognized part is acquired and stored.
- the voice data of the entire sentence determined in advance is obtained by repeating this process and appropriately connecting the acquired and stored voice data, the connected voice data is registered.
- the first user 204 can be used without burdening the first user 204. It is possible to obtain voice data in which many voices of the person 204 are recorded.
- the creation of such phoneme data may be performed by the information processing module 104 having a function as an artificial intelligence and the database 102, or may be executed in cooperation with an external server that creates phonemes.
- the text data of the reply sentence is converted into voice data by using the phoneme data created as described above. Create Then, the created voice data is transmitted to the second user 206 as a response sentence.
- the communication providing system 100 can acquire the voice data of the conversation performed via the SNS and create phoneme data that meets each user without giving a special load to the user. Even if it is pseudo, voice data close to the user can be created and voice communication can be established.
- the communication providing system can include an evaluation module in which the second user 206 evaluates the content of the answer transmitted from the communication providing system 100.
- the answer can be evaluated by operating the user terminal 200 of the second user 206.
- the user terminal 200 can display the answer contents on the screen and display an icon (or button) for performing the evaluation, and the second user 206 can execute the evaluation by performing an operation according to the screen display.
- the evaluation screen can be constructed by displaying icons (or buttons) indicating “Like” and “Dislike” (buttons 15 and 16 in FIG. 1J) together with the reply message. Further, the evaluation may be in a form in which the user gives a score instead of the two-choice set.
- FIG. 7 is a diagram for explaining a mode in which an answer answered from the communication providing system 100 is evaluated with respect to a question made by the second user 206 to the virtual first user 204b.
- the second user 206 (any of Mr. A, Mr. B, and Mr. C) has a question to communicate with the virtual first user 204 b generated on the communication providing system 100.
- the case is shown (left side in the figure).
- the second user 206 (any one of Mr. A, Mr. B, and Mr. C) evaluates the answer provided from the communication providing system 100. For example, in the case of Mr. A, it is evaluated as “Dislike”, and in the case of Mr. B or Mr. C, it is evaluated as “Like” (approximately the center in the figure).
- the information processing module 104 having a function as an artificial intelligence can total the evaluation information input to the evaluation module for each response sentence and accumulate it as evaluation data.
- the evaluation data may be stored in a point system in correspondence with the response text data.
- the information processing module 104 having a function as an artificial intelligence can rank the response data having a high evaluation value so that the response data with a high evaluation value is preferentially selected.
- the information may be notified to the first user 204.
- the first user 204 can edit the response content and update the response text.
- the communication providing system includes an editing module that allows the first user 204 to edit the reply sentence created by the communication providing system 100 by itself. For example, "Who are you” or "Tell me your name” In response to the question “I am XX (name).” It is assumed that the information processing module 104 having a function as artificial intelligence is learned so as to respond. However, some users may think that the response content thus derived is not appropriate. For example, to the above question: “I am the CEO of XX Company.” May be considered appropriate.
- the first user 204 may receive a low evaluation notification for the response content (left side in the figure). In that case, the first user 204 may want to modify and edit the response content.
- the communication providing system 100 individually sets a response sentence when the user wants to set a response sentence that is different from the response sentence created by the information processing module 104 having a function as an artificial intelligence. It can be edited by the function of the editing module.
- the user may input the text data of the response content from the user terminal 200 and update the content by the update module.
- the user may learn by interacting with the information processing module 104 having a function as artificial intelligence.
- the first user 204 accesses the communication providing system 100 using the user terminal 200, and gives a pseudo first user 204 created on the communication providing system 100 to, for example, "Tell me your name” Ask.
- the information processing module 104 having the function as artificial intelligence has a prepared answer as follows: “I am XX (name).” Responds. The first user 204 can determine that there is no problem with the response content. In this case, the first user 204 need not use the editing function. Next, another question is "Who are you?" In response to the question, the information processing module 104 having a function as artificial intelligence, “I am XX (name).” When the response is made, the first user 204 determines that this response is not appropriate, and uses the editing function.
- the icon is operated according to the screen display of the user terminal 200 to set the editing mode. Then, the response content “I am XX (name)” is set to the response content “I am the representative director of XX company”. This allows you to answer the question “Who are you?” To answer “I am the representative director of XX company”.
- FIG. 8A shows a flowchart explaining the evaluation of questions and answers by the second user 206 and the manner in which the first user 204 performs editing.
- the second user 206 accesses the virtual first user 204b in which the communication providing system 100 exists and asks a question (question) (S11).
- the communication providing system 100 generates an answer to the question (S12) and transmits the answer to the second user 206 (S13). Then, the second user 206 evaluates the obtained answer (S14).
- the communication providing system 100 aggregates the evaluation values (S15). If the evaluation value is normal or high, or the evaluation value is higher than a certain level, the answer may be weighted (S16, S17), assuming that the answer is a preferable answer. In this case, the weighted answer may have a high probability of being selected at the next search.
- the first user 204 is notified to that effect (S16, S18).
- the first user 204 that has received the notification (S19) edits the content of the response so that a favorable response is obtained (S20).
- the communication providing system 100 receives the edited answer, the communication providing system 100 corrects and updates the corresponding answer recorded in the database (S21). If the first user 204 determines that there is no need to correct the content of the answer, the first user 204 can leave it as it is.
- the second user 206 asks a question and the first user 204 edits the answer to the question.
- the first user 204 has a conversation with the virtual first user 204b appearing in the communication providing system 100, asks himself a question, and evaluates himself. And editing can be performed.
- the first user 204 can know the content of the answer to the question, and can evaluate the answer and edit the answer so that a preferable answer can be obtained. That is, the information processing module 104 having a function as artificial intelligence can be learned.
- the communication providing system 100 can create a new response sentence by the user editing data associated with a question and a response.
- the user can learn the artificial intelligence function to respond appropriately to the question while interacting with the pseudo self generated on the communication providing system 100.
- the first user 204 edits the answer to the question while interacting with the virtual first user 204b generated on the communication providing system 100, and functions as an artificial intelligence.
- the information processing module 104 can be learned.
- FIG. 8B shows an aspect in which the communication providing system 100 generates a candidate for an answer to a question and the user selects it.
- the communication providing system 100 generates candidate answers to the question based on the first database 102a in which personal social data is stored.
- the first information processing module 104a having a function as artificial intelligence has a function of generating a plurality of answer candidates based on different logics.
- the first information processing module 104a having a function as artificial intelligence generates answer candidates by weighting the contents of recent conversations from the personal social data of the first user, and generates another logic.
- answer candidates are created based on the content of the conversation with a high appearance frequency from all the data. For example, in response to a question about favorite food, there is logic for answering food that has been eaten recently and logic for answering food that the first user has long liked.
- the information processing module 104b having a function as artificial intelligence is stored in the second database 102b when the first information processing module 104a cannot find an appropriate answer in the first database 102a. Create candidate answers based on the data.
- the second database 102b in addition to storing social data of the entire user, for example, various information on the Internet is collected and stored by crawling.
- the answer to the question generated by the information processing module 104b having the function as artificial intelligence is not an answer reflecting the individual personality of the user, but a generalized answer for social wisdom. For example, an answer that is pleasing is generated for a question about a player in the home country that won the international match.
- the answer candidates generated by the information processing module 104 having a function as artificial intelligence are output to the first user's terminal 204a.
- the first information processing module 104a having a function as artificial intelligence generates seven types of answer candidates and outputs the answer candidates to the first user terminal 204a.
- FIG. 8B shows a mode in which the answers A1 to A7 are displayed on the terminal 204a of the first user.
- the answer candidates generated by the information processing module 104 having the function as artificial intelligence are in order of highest priority (the order of answers judged to be most accurate). Answer candidates A1 to A7 are displayed.
- the first user can change the order of answers by his / her own judgment with reference to the answers A1 to A7.
- the order of the answer candidates A1 and A3 is switched.
- the order of answer candidates changed by the first user is reflected in the information processing module 104 having a function as artificial intelligence.
- answer candidates are generated in the order of A3, A2, and A1.
- This result is used as a response of the pseudo first user 204b generated by the information processing module 104 having a function as artificial intelligence.
- the information processing module 104 having a function as artificial intelligence generates a plurality of answer candidates for a question, and the first user selects the answer candidate, thereby making a pseudo first user.
- the answer of 204b can be made more appropriate.
- Such an answer candidate editing function can edit an answer candidate by the function of the editing module described above, and can update an answer candidate for the question based on the contents edited by the editing module by the function of the update module.
- the communication providing system 100 can acquire a user's appearance as image data, convert it into three-dimensional image data, and display it on a user terminal. For example, in the configuration shown in FIG. 1A, when the second user 206 interacts with the virtual first user generated on the communication providing system 100, the video of the first user is displayed on the user terminal 200. Can be displayed.
- User image data can be acquired using the imaging function of the user terminal 200.
- the user takes a picture of himself / herself using the user terminal 200 and uploads the image data to the communication providing system 100.
- image data may be extracted using the photographed information. it can. Thereby, the latest image data of the user can be acquired, and the three-dimensionalization of the user photo can be realized based on the latest image data of the user.
- the communication providing system 100 can generate 3D image data by estimating and calculating the 3D shape of a person shown in the 2D photo data.
- the method for example, three-dimensional image data is created by a method for obtaining depth information from the density distribution of the texture of the object or a method for calculating a normal direction from the intensity of light reflection on the object surface.
- the user's 3D image data may be appropriately updated as described above.
- image data is retrieved from the personal social data of a specific user recorded in the first database 102a and the user is captured using the face recognition function. Is identified. And the structure which selects the newest data from the image data in which a user is reflected, and updates a user's three-dimensional image data based on the data can also be provided.
- the other party's video is displayed on the user terminal as a three-dimensional image, thereby giving a sense of reality to the communication.
- the 3D image of the virtual individual is updated to the most recent data, so that the appearance of the other party becomes more specific, and the feeling of intimacy can be further enhanced through pseudo communication.
- the communication providing system 100 has a function of automatically generating and registering a question sentence using social data stored in the database 102.
- many personal “tweets” may be recorded as personal social data of the first user 204.
- personal “tweets” include, for example, “I went shopping in Tokyo today” and “I'm waiting in Shibuya”. Such personal social data does not have a form of response to the inquiry, but “tweet” includes at least one piece of information.
- the information processing module 104 having a function as an artificial intelligence creates and registers question data from information obtained by decomposing text data of such personal social data into nouns, verbs, adjectives, etc., and syntactic interpretation. It has a function to keep.
- the information processing module 104 having a function as artificial intelligence refers to social data of a plurality of users registered in the second database 102b and learns what kind of conversation is being made. For example, in response to the question “Where did you go?”, If there are many conversations such as “The function went to XX (place name).” ? "Is generated and registered in the first database 102. If a plurality of question sentences can be created, they can be registered in the similar database 106.
- the information processing module 104 having a function as artificial intelligence automatically creates a question sentence from personal social data, so that the dialogue ability can be enhanced.
- This parameter may be assigned to the user according to a personality analysis of the user, such as a person who has a high probability of occurrence of fluctuation or a person who is not easily influenced by other users. Thereby, the personality of each user or the userness can be produced depending on how the fluctuation is reflected.
- Such parameters can be stored in the communication providing system 100 as user attributes at the time of registration in the communication providing system 100 and read out as necessary.
- the result of detecting the temporal change of the user's answer accumulated in the first database 102a is stored in the communication providing system 100, and if necessary, Can be read.
- the result of detecting temporal changes in the information related to the user accumulated in the first similar database 106a can be stored in the communication providing system 100 and read out as necessary.
- the information processing module 104 having a function as artificial intelligence can perform a response to the inquiry as follows.
- ⁇ Users may give answers without receiving questions, such as tweets.
- tweets For example, the following tweet:
- Table 3 summarizes such tweets from user 1.
- Question and answer combinations can be obtained from past conversations.
- the communication providing system 100 can learn past conversations.
- the communication providing system 100 according to the present embodiment can generate new questions and infer and predict responses by learning past conversation data (social data accumulated in a database).
- FIG. 9 shows a configuration for retrieving an answer from a past conversation when a question is received and predicting an appropriate answer.
- An important aspect for predicting responses is where artificial intelligence must imitate human behavior.
- the response to a question depends on the person asking the question. Therefore, in order to predict a response to a question made by a certain user, the artificial intelligence needs to refer to the user's past conversation data and predict a response content with a high priority obtained from the data.
- the artificial intelligence finds an appropriate response sentence from the conversation that the first user exchanges with another user. There is a need. When doing this, it is necessary to be careful not to leak confidential matters such as personal information of other users by searching for conversations with other users.
- the artificial intelligence estimates the confidentiality score of each answer in the content of the conversation exchanged by other users.
- an answer sentence that has already been shared with another user has a low confidentiality score.
- the confidentiality score can be low.
- FIG. 10 shows a process flow for selecting an answer to a question from user 2 to user 1.
- the artificial intelligence first searches for a suitable answer from conversations between the first user and the second user exchanged in the past (S21). As a result, when a suitable answer is found, it is selected as an answer (S22, S26). On the other hand, when a suitable answer is not found, one or a plurality of answers learned from the conversations of all users are selected (S22, S23). The confidentiality score is evaluated for the answer obtained here (S24). Then, one with a low confidentiality score is selected and selected as an answer (S25, S26).
- the answers to the questions may be ranked as follows. For example, as shown in FIG. 11, when an answer is found in a conversation that the first user and the second user have exchanged in the past, the answer can be shared between the two at the “personal level”. Can be an answer. Further, when finding an appropriate response sentence from the conversation in which the first user exchanges with other users, the confidentiality score is evaluated for one or a plurality of obtained answers. An answer with a low confidentiality score can be set as a “public level” answer. Furthermore, as shown in FIG. 12, the social data of a plurality of users is learned, the confidentiality score of one or a plurality of answers obtained therefrom is evaluated, and an answer with a low score is assigned to a “shared level”. Can be an answer.
- I a submodel representing a conversation between user i and user j.
- Q (i, j) represents a question made from user j to user i
- a (i, j) represents an answer made from user i to user j
- t corresponds. Indicates the time to do.
- q is a question (or query) of user j with respect to user i
- A is a predicted answer.
- the predicted answer when the confidentiality score of the answer is higher than the first threshold, the predicted answer is an individual level answer. Further, when the confidentiality score is equal to or lower than the first threshold value and the confidentiality score is lower than the second threshold value, the predicted answer is a public level answer. If it does not correspond to these, it can be treated as a shared level answer.
- Such an aspect of the answer level can be expressed by the following equation.
- Cprivate gives a confidentiality score for the personal level answer
- C public gives a publicity score.
- sim (q, Qx) is a function indicating the similarity or matching between q and tuple (Qx, Ax, tx).
- FIG. 13 illustrates how contexts are integrated by grouping questions and answers in a conversation. It is shown that the matching of the query with the question and the answer is a weighted sum.
- the similarity when integrating contexts can be expressed as:
- a i is a weighting factor
- W indicates the number of tuples included in the context window.
- Each query for a question and an answer is composed of a set of vectors of keyword (W), concept (C), type (T), and group (G) as shown in the following equation.
- Keywords Keywords are obtained by building words from all messages exchanged between users. Vector similarity between two keywords is weighted (TF-IDF: Robertson, SE, Walker, S., Beaulieu, MM, Gatford, M., & Payne, A. Okapi at TREC-4. In Proceedings of the 4th Text REtrieval Conference (TREC-4), pp. 73-96, 1995]). Or, OkapiBM25 (Harman, D. Ranking algorithms. In WB Frakes & R. Baeza-Yates (Eds.), Information retrieval: Data structures and algorithms, pp. 363-392. Englewood Cliffs, New Jersey, USA: Prentice Hall, 1992]), or any other ranking function used in information retrieval. Similarity between keywords includes semantic similarity.
- Types Types are defined as different groups of messages from the perspective of grammatical structure and reputation analysis. This group includes questions, people, rough places, order, romantic messages, happy messages, and more. These classifications can be obtained by the artificial intelligence learning of manually labeled feature vectors, grammatical data, concepts and keywords.
- Group Tuples (Q, A) of all users are clustered based on the similarity of text between answers. Each cluster is given a name.
- the group vector is a list of clusters to which the tuple (Q, A) belongs. Both have similar answers, so obviously two questions in the same group are similar.
- FIG. 15 shows a mode in which questions and answers are clustered by the similarity of the answers, and each cluster forms a group.
- Artificial intelligence can be learned using all available data.
- the input vector is a question and the output vector is an answer.
- the learned artificial intelligence estimates the answer template.
- the answers in this template can be matched to the answers in the database. This is important in order to be able to find an answer when a question is not found in the data.
- Any supervised algorithm can be used in this case as a neural network, support vector machine, k-Nearest-Neighbors, Gaussian mixture model.
- the communication providing system acquires data from various SNSs and accumulates them in a database, and an information processing module having a function as artificial intelligence is used by a user. Learning thoughts can provide appropriate answers to questions.
- the communication providing system collects social data from a plurality of SNSs, and an information processing module having a function as artificial intelligence infers and learns it, so that one user can conceive. You can also offer no thoughts.
- first database 102a that generates a virtual user generated by the communication providing system 100 is made to acquire specialized knowledge, the virtual user is acquired.
- Various services can be provided. Such expertise may be acquired from an SNS used by the corresponding user.
- the applicable field Expertise can be made available to virtual users.
- the expertise in this case includes not only uniform expertise accumulated by books and the like, but also information on expertise used by natural persons as a conversation on a social network. By including information related to such conversational expertise, the communication providing system 100 can effectively use various expertise as an answer to a question.
- such specialized knowledge can be acquired as pseudo first user knowledge generated by artificial intelligence, for example, if the first user desires.
- the pseudo first user can not only acquire specific specialized knowledge but also use the first user in consideration of the individuality of the first user. For example, if the first user has knowledge in the medical field, the pseudo first user also has expertise in the medical field, and the pseudo first user has further specialized knowledge. If you acquire expertise in the legal field, you will be able to demonstrate your abilities in the field of forensic medicine.
- the expertise stored in the first database 102a may be stored in a question and answer (ie, Q & A) format.
- the social data (knowledge) stored in the first database 102 can be updated in a timely manner, and the data is not lost, so that the amount of knowledge can be increased.
- a virtual individual having high expertise can appear on the communication providing system in accordance with the user's attributes, and the service can be provided to other users.
- FIG. 2 shows an information processing module (first information processing module) having a function as artificial intelligence configured based on personal social data and a function as artificial intelligence configured based on social data of the entire user.
- the information processing module (second information processing module) having The present invention can further include an information processing module (third information processing module) having a function as an artificial intelligence in which a plurality of first information processing modules are aggregated to share information or a knowledge level.
- FIG. 17 shows an aspect in which there are a plurality of first information processing modules 104a configured based on personal social data and having functions as artificial intelligence (first information processing modules 104a_1 to 104a_n (n is 2 or more). integer)).
- first information processing modules 104a_1 to 104a_n (n is 2 or more). integer)
- the first information processing module 104a generates pseudo users called “private AI” and “personalized AI”.
- the second information processing module 104b linked to the first information processing module 104a is based on a fusion of knowledge obtained by crawling social data of the entire user and various information on the social network. It is artificial intelligence and is called “Everyone AI”, “Common Sense AI” or “Moral AI”.
- FIG. 17 shows that a plurality of first information processing modules 104a_1 to 104a_n form one set 110.
- a third information processing module 112 having a function as an artificial intelligence that shares knowledge and information is connected to the set 110.
- the third information processing module 112 is artificial intelligence generated by a common knowledge level generated by the plurality of first information processing modules 104a_1 to 104a_n. That is, the third information processing module 112 is a pseudo user having all the knowledge, personality, and personality of the plurality of pseudo users generated by the plurality of first information processing modules 104a_1 to 104a_n. It can be said that there is.
- Such third information processing module 112 can be referred to as a “group AI”.
- the third information processing module 104c can demonstrate the knowledge and individuality of each pseudo user generated by the plurality of first information processing modules 104a_1 to 104a_n. From another viewpoint, the pseudo user generated by the third information processing module 112 is different from the pseudo user generated by the first information processing module 104a. Since a plurality of individuals are generated together, it can be said that they have a corporate character.
- the plurality of first information processing modules 104a_1 to 104a_n can have a conversation with the third information processing module 112. Thereby, the plurality of first information processing modules 104a_1 to 104a_n can share each other's information and knowledge. Even if one of the pseudo users generated by the first information processing modules 104a_1 to 104a_n leaves the group, the third information processing module 112 can share the knowledge and personality of the detached user. Since it can hold
- the third information processing module 112 having a function as artificial intelligence is operated as a function of a part of a certain corporation, even if a member (natural person) belonging to the corporation leaves in the middle, the member who has left You can keep your skills in that corporation.
- the third information processing module 104c can acquire new information from the plurality of first information processing modules 104a_1 to 104a_n, whereby artificial intelligence grows. That is, the skill of each individual person generated by the first information processing modules 104a_1 to 104a_n is improved, so that the skill is a skill of the pseudo user generated by the third information processing module 112. Reflected.
- the third information processing module 112 may be placed in an accessible state with the second information processing module 102b through an electric communication line. Thereby, the third information processing module 103c can collect a wide range of information.
- a plurality of pseudo users generated by artificial intelligence can be used to determine the knowledge, intelligence, and personality of a plurality of users forming the set.
- a pseudo user with everything can be formed and used.
- SYMBOLS 100 Communication provision system, 102 ... Database, 104 ... Information processing module, 106 ... Similarity database, 107 ... Question database, 108 ... Response database, 200 ... User terminal, 202 ... Plural users, 204 ... First user, 206 ... Second user, 300 ... SNS
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Information Transfer Between Computers (AREA)
Abstract
Description
図1Aは、本発明の一実施形態に係るコミュニケーション提供システムと、ソーシャル・ネットワーキング・サービス(以下「SNS」ともいう。)及び、このコミュニケーション提供システムを利用する利用者との関係を示す。本実施形態に係るコミュニケーション提供システム100は、利用者が利用している様々なSNSからデータを取得しデータベースに蓄積する機能を有している。コミュニケーション提供システム100はさらに、データベースに蓄積された情報に基づいて、利用者の思考を推論し学習する機能(人工知能としての機能)を有し、質問に対する適切な回答を提供可能としている。以下、その詳細を説明する。
図3Aは、本発明の一実施形態に係るコミュニケーション提供システム100におけるデータベース102構成を説明する図である。図3Aで示すように、データベース102は複数の階層に分かれて構築されていてもよい。第1のデータベース102aは、各利用者に対応するデータベースであり、個人ソーシャルデータが記憶されている。この第1のデータベース102aに記憶される個人ソーシャルデータは、コミュニケーション提供システム100において、仮想的な利用者データを生成するときの基礎データとしても利用される。したがって、個人ソーシャルデータおよびそれから生成される情報は、利用者ごとにデータベースに記憶される。また、個人ソーシャルデータから生成される情報には、後述のように、個人ソーシャルデータをキーワードにより分析して得られるベクトル、個人ソーシャルデータをコンセプトにより分析して得られるベクトル、個人ソーシャルデータをタイプにより分析したベクトル、および、個人ソーシャルデータをグループにより分析したベクトルのいずれか一以上が含まれる。
人工知能としての機能を有する情報処理モジュール104は、データベース102に記憶されているソーシャルデータを用いて仮想個人を生成し、実在の利用者とコミュニケーションを行う。人工知能としての機能を有する情報処理モジュール104は、例えば、データベース102に記憶されているテキストデータを認識し、質問に対する回答を作成又は生成する機能を有する。このとき人工知能としての機能を有する情報処理モジュール104は、データベース102に記憶されたテキストデータから、質問に対する適切な回答を推論し又は学習して決定する機能を有している。
「貴方は誰ですか」
という質問に対し、
「私は○○(名前)です。」
という回答が多くなされている場合、人工知能としての機能を有する情報処理モジュール104は同様の回答を、質問に対する回答として決定することができる。上記は簡単な質問の一例であるが、第2のデータベース102bに基づけば、多数の利用者の回答傾向が反映されるので、時流に沿った回答を得ることができる(S06)。
データベース102は、質問と質問に対する返答を関連付けて記憶している。しかし、人間社会のコミュニケーションにおける質問(問いかけ)と、それに対する返回答(返答)は画一的なものではないのが通常である。例えば、初対面の人の名前を問いかけようとする場合、「貴方は誰ですか?」と問いかけるような場合もあれば、「お名前を教えて下さい」と問いかける場合もある。
本実施形態に係るコミュニケーション提供システムは、図1Aで示す態様において、第2の利用者206が音声で話しかけることでその音声を認識し、第1の利用者204の本人の音声(又は、擬似的な本人の音声)で返答する機能を有していてもよい。あるいは、第2の利用者206が文字情報などにより非音声的に質問をした場合に、第1の利用者204の本人の音声(又は、擬似的な本人の音声)で返答する機能を有していてもよい。このような音声による返答の方式としては、あらかじめ第1の利用者204の本人の音声を数多く録音しておき、それを用いて返答文を音声データとして作成する方式がとり得る。しかしながら、このような方式は、利用者本人に予め多量の音声データの入力を強いることになり、全ての利用者に適用することは好ましくなく、また利用者の利便性を損なうこととなる。
本実施形態に係るコミュニケーション提供システムは、第2の利用者206が、コミュニケーション提供システム100から送信された回答の内容を評価する評価モジュールを有することができる。回答の評価は、第2の利用者206のユーザ端末200を操作することによって行うことができる。ユーザ端末200に、回答内容を画面表示するとともに、評価を行うためのアイコン(またはボタン)を表示させ、第2の利用者206はその画面表示に従って操作をすることにより、評価を実行することができる。評価用の画面は、回答のメッセージと共に、「Like」、「Dislike」を示すアイコン(またはボタン)(図1Jのボタン15および16)を表示させることで構築することができる。また、評価は2者択一式ではなく、利用者が点数を付ける形式のものであってもよい。
本実施形態に係るコミュニケーション提供システムは、コミュニケーション提供システム100が作成した返答文を、第1の利用者204が自ら編集する編集モジュールを有する。例えば、
「貴方は誰ですか」又は「お名前を教えて下さい」
という質問に対し、
「私は○○(名前)です。」
と返答するように、人工知能としての機能を有する情報処理モジュール104は学習されているとする。しかし、利用者によっては、このように導き出された返答内容では適切でないと考える場合がある。例えば、上記の質問に対して、
「私は○○会社の代表取締役です。」
と答えることが適切であると考える場合がある。
「お名前を教えて下さい」
と問いかける。人工知能としての機能を有する情報処理モジュール104は、予め用意されている回答として、
「私は、○○(名前)です。」
と応答する。第1の利用者204は、この応答内容に問題ないと判断することができる。この場合、第1の利用者204は編集機能を使用しないで済む。次に、別の質問として、「貴方は誰ですか」
という問いかけに対し、人工知能としての機能を有する情報処理モジュール104は、予め用意されている回答として、
「私は、○○(名前)です。」
という応答がなされると、第1の利用者204は、この応答は適切ではないと判断し、編集機能を使用する。例えば、ユーザ端末200の画面表示に従ってアイコンを操作し、編集モードにする。そして、「私は、○○(名前)です。」という応答内容を、「私は○○会社の代表取締役です。」という応答内容に設定する。これにより、「貴方は誰ですか」という問いかけに対しては、「私は○○会社の代表取締役です。」と回答するように設定することができる。
図8Bは、コミュニケーション提供システム100が、質問に対する回答の候補を生成し、利用者がこれを選択する態様を示す。コミュニケーション提供システム100は、個人ソーシャルデータが記憶された第1のデータベース102aに基づいて、質問に対する回答の候補を生成する。このとき、人工知能としての機能を有する第1の情報処理モジュール104aは、異なるロジックに基づいて複数の回答の候補を生成する機能を有する。人工知能としての機能を有する第1の情報処理モジュール104aは、第1の利用者の個人ソーシャルデータの中から、近時の会話の内容に重み付けをして回答の候補を生成し、別のロジックとしては全データの中から出現頻度の高い会話の内容に基づいて回答候補を作成する。例えば、好きな食べ物の関する質問に対し、最近食べて美味しかった食べ物を回答するロジックと、第1の利用者が昔から好きであった食べ物を回答するロジックとがある。
本実施形態に係るコミュニケーション提供システム100は、利用者の姿を画像データとして取得し、三次元の画像データに変換してユーザ端末に表示させることができる。例えば、図1Aで示す構成において、第2の利用者206が、コミュニケーション提供システム100上に生成される仮想の第1の利用者と対話をするとき、ユーザ端末200に第1の利用者の映像を表示させることができる。
本実施形態のコミュニケーション提供システム100は、データベース102に記憶されているソーシャルデータを使って、質問文を自動的に生成し、登録しておく機能を有する。
質問に対する返答には揺らぎを与えるようにしてもよい。例えば、
「お元気ですか?」
という質問に対し、
「元気です。あなたはいかがですか?」
という返答が設定されていたとしても、必ずしもこの返答をしないようにしてもよい。例えば、他の多くの利用者が「すごくいいよ!」と返答をしているとき、それに影響をされて同様に回答をするようにしてもよい。
人工知能としての機能を有する情報処理モジュール104は、問いかけに対する応答を次のようにして行うことができる。
例えば、利用者1と、利用者2とで次のような会話がなされることを想定する。ここで、T1、T2・・・は時刻を表すものとする。
利用者2:「こんにちは」[T2]
利用者1:「ごきげんいかがですか」[T3]
利用者2:「調子いいよ、あなたは」[T4]
利用者1:「今日は素晴らしいよ」[T5]
利用者1:「コンピューティングの将来はユビキタス人工知能である」[T2]
また、図11に示すように、質問に対する回答は、次のようにランク分けしてもよい。例えば、図11で示すように、第1の利用者と第2の利用者が過去に交わした会話の中から回答が見つけられる場合、その回答は両者の間で共有され得る「個人レベル」の回答とすることができる。また、第1の利用者が他の利用者と交わしている会話の中から、適切な応答文を見つける場合には、得られた一つ又は複数の回答について機密性のスコアを評価する。そして、機密性のスコアが低い回答を「公開レベル」の回答とすることができる。さらに、図12で示すように、複数の利用者のソーシャルデータを学習し、そこから得られた一つ又は複数の回答の機密性のスコアを評価し、スコアの低い回答を「共有レベル」の回答とすることができる。
は次式で表すことができる。
ここで、Q(i,j)は利用者jから利用者iに対してなされた質問を表し、A(i、j)は利用者iから利用者jになされた回答を示し、tは対応する時間を示す。
ここで、sim(q,Qx)は、q及びtuple(Qx,Ax,tx)との間の類似性又はマッチングを示す関数である。
質問に対する回答を予測する場合には、文脈を把握することが重要となる。同じ質問に対する回答は、会話の中でなされていた、過去の質問と回答に依拠することができる。これは、「はい」と返事をする場合のような、一般的な質問の場合に特に当てはまるものとなる。そのような質問への答えは、上記で述べた関数に当てはめることができる。文脈を統合するために、過去になされた会話における質問と回答の学習データをグループ化して適用することができる。図13は、会話における質問と回答をグループ化することにより、文脈を統合する態様を示す。クエリと、質問及び回答とのマッチングは、加重和となることが示されている。
過去の会話から質問と回答の適合性を見つけるために、それらの間の類似性を評価する必要がある。ベクトルの集合として質問と回答を、各クエリで表すことができる。2つのベクトルの集合間の類似性は、対応するベクトル間の類似度の加重和として表すことができる。ベクトル間の類似性は、コサイン距離、ユークリッド距離、マハラノビス距離として推定することができる。
キーワードは、利用者の間で交わされるすべてのメッセージから単語を構築することによって得られる。2つのキーワードの間のベクトルの類似性は、重み付け(TF-IDF:Robertson, S. E., Walker, S., Beaulieu, M. M., Gatford, M., & Payne, A. Okapi at TREC-4. In Proceedings of the 4th Text REtrieval Conference (TREC-4), pp. 73-96, 1995])されたキーワードのコサイン距離として得ることができる。或いは、OkapiBM25(Harman, D. Ranking algorithms. In W. B. Frakes & R. Baeza-Yates (Eds.), Information retrieval: Data structures and algorithms, pp. 363-392. Englewood Cliffs, New Jersey, USA: Prentice Hall, 1992])、または情報検索において使用される任意の他のランク付け関数を用いることができる。キーワード間の類似性には意味的類似性が含まれる。それは、意味的に関連している2つのキーワードは類似していると考えられる。この意味的な関係は、WordNet(G. A. Miller, R. Beckwith, C. D. Fellbaum, D. Gross, K. Miller. WordNet: An online lexical database. Int. J. Lexicograph. 3, 4, pp. 235-244, 1990)、BabelMet(R. Navigli and S. P Ponzetto. BabelNet: The Automatic Construction, Evaluation and Application of a Wide-Coverage Multilingual Semantic Network. Artificial Intelligence, 193, Elsevier, pp. 217-250, 2012)、共出現分析(Harris Z. S. Co-occurrence and transformation in linguistic structure, Language, 33, pp. 283-340., 1957)のデータを用いて得ることができる。類似度を推定する際には、文書内におけるキーワードの順序と場所を考慮しなければならない。システム内の順序を統合する対策として場所を表すインデックスの分散を用いる。図14は、どのようなキーワードの順序が、一致を求める際に重要であるかを示している。
コンセプトは、語彙の次元を圧縮することで得ることができる。潜在的意味解析(Latent Semantic Analysis)は、コンセプトの限定されたリストを提供するために用いることができる(Thomas K. Landauer, Peter W. Foltz et Darrell Laham. Introduction to Latent Semantic Analysis, Discourse Processes, vol. 25, p. 259-284, 1998)。各質問と応答は、概念空間に投影される。この投影は概念ベクトルを与える。そして、コサイン距離は、概念ベクトル間の類似度を推定することができる。
タイプは文法構造と評判分析の観点から、メッセージの異なるグループとして定義される。このグループには、質問、人、大体の場所、順序、ロマンチックなメッセージ、幸せなメッセージなどが含まれる。これらの分類は、手動のラベル付けされた特徴ベクトルは、文法構造であるデータ、概念やキーワードを人工知能が学習することによって得ることができる。
全利用者のタプル(Q,A)は、回答同士のテキストの類似性に基づいてクラスタ化される。各クラスタには名前が与えられる。グループベクトルはタプル(Q,A)が属するクラスタのリストである。どちらも同じような答えを持っているので、明らかに同じグループの2つの質問は類似するものとなる。図15は、質問と回答が、その回答の類似性によってクラスタ化された態様を示し、各クラスタが一つのグループを構成する態様を示す。
利用可能な全てのデータを使って人工知能を学習させることができる。入力ベクトルは質問であり、出力ベクトルは回答となる。任意の質問の回答を得るために、学習された人工知能は、解答のテンプレートを推定する。データベースに記憶されている質問と共にクエリのマッチングと組み合わせることで、データベース内の回答にこのテンプレートの答えを一致させることができる。これは、質問がデータの中に見つけられないときに、回答を見つけることができるようにする上で重要である。どんな監修アルゴリズムであっても、ニューラルネットワーク、サポートベクターマシン、k-Nearest-Neighbors、ガウス混合モデルとして、この場合使用することができる。
図2は、個人ソーシャルデータに基づき構成される、人工知能としての機能を有する情報処理モジュール(第1の情報処理モジュール)と、利用者全体のソーシャルデータに基づき構成される、人工知能としての機能を有する情報処理モジュール(第2の情報処理モジュール)と、の関係を示している。本発明は、さらに複数の第1の情報処理モジュールが集合して、情報又は知識レベルを共有する、人工知能としての機能を有する情報処理モジュール(第3の情報処理モジュール)を有することができる。
Claims (38)
- 一つ又は複数のソーシャル・ネットワーキング・サービスに登録されている第1の利用者のソーシャルデータと、前記複数のソーシャル・ネットワーキング・サービスに登録されている複数の利用者のソーシャルデータと、前記複数のソーシャル・ネットワーキング・サービスから収集された複数のテキストデータと、の情報が記憶されたデータベースと、
前記第1の利用者に対する第2の利用者の質問を、前記データベースに記憶されている前記情報の少なくとも一部に基づいて、前記質問に対する回答を推論し又は学習して決定する人工知能としての機能を有する情報処モジュールと、を含むことを特徴とするコミュニケーション提供システム。 - 前記データベースは、前記第1の利用者のソーシャルデータを記憶する第1のデータベースと、前記複数の利用者のソーシャルデータを記憶する第2のデータベースと、前記複数のテキストデータを記憶する第3のデータベースと、に階層化されている、請求項1に記載のコミュニケーション提供システム。
- 前記人工知能としての機能を有する情報処理モジュールは、前記第1のデータベース、前記第2のデータベース及び前記第3のデータベースに記録されているデータの中から、前記質問に対する適切な回答を、上位のデータベースに記録された情報の少なくとも一部に基づいて推論し生成する、請求項2に記載のコミュニケーション提供システム。
- 前記人工知能としての機能を有する情報処理モジュールは、前記第1のデータベースに記憶されている前記複数の利用者のソーシャルデータから、質問に対する回答を傾向分析し、前記質問に対する回答を推論し又は学習して決定する、請求項3に記載のコミュニケーション提供システム。
- 前記傾向分析は、前記質問に対する前記複数の利用者の前記回答の時間的な変化を分析することである、請求項4に記載のコミュニケーション提供システム。
- 前記第1の利用者のソーシャルデータ及び前記複数の利用者のソーシャルデータは、テキストデータ並びに、音声データ、写真データおよび映像データのいずれか一以上から生成されたテキストデータである、請求項1に記載のコミュニケーション提供システム。
- 類似する質問を類似質問としてグループ化し、前記類似質問に対応する回答が対応付けられている類似データベースを含み、
前記人工知能としての機能を有する情報処理モジュールは、前記類似データベースから前記質問の回答を選択する、請求項6に記載のコミュニケーション提供システム。 - 前記類似データベースは、類似質問内容をグループ化して記録する類似質問データベースと、前記類似質問に対応する回答を記憶する類似回答データベースと、を含む、請求項7に記載のコミュニケーション提供システム。
- 前記類似データベースは、前記第1の利用者のソーシャルデータの中から質問と回答の内容を類型化したデータが記憶される第1の類似データベースと、前記複数の利用者のソーシャルデータの中から質問と回答の内容を類型化したデータが記憶される第2の類似データベースと、前記テキストデータの中から質問と回答の内容を類型化したデータが記憶される第3の類似データベースと、に階層化されている、請求項8に記載のコミュニケーション提供システム。
- 前記第2の利用者が前記回答を評価した情報を取得して、前記回答に評価値を付ける評価モジュールを含む、請求項1に記載のコミュニケーション提供システム。
- 前記評価モジュールは、前記回答の評価結果を、前記第1の利用者に通知する通知モジュールを含む、請求項10に記載のコミュニケーション提供システム。
- 前記通知モジュールから通知を受けた第1の利用者が、前記質問に対する前記回答の内容を編集する編集モジュールと、前記編集モジュールで編集された内容に基づいて前記質問に対する回答の内容を更新する更新モジュールと、を含む、請求項11に記載のコミュニケーション提供システム。
- 前記評価モジュールは、前記評価結果に基づいて優先準位を付け、前記人工知能としての機能を有する情報処理モジュールは、前記優先順位に基づいて前記質問に対する適切な回答を推論し又は学習して決定する、請求項10に記載のコミュニケーション提供システム。
- 前記データベースは、経時的に、前記第1の利用者のソーシャルデータ、前記第2の利用者のソーシャルデータ、前記テキストデータを更新し、蓄積する、請求項1に記載のコミュニケーション提供システム。
- 前記第1の利用者の三次元画像データを生成する画像データ生成モジュールを含む、請求項1に記載のコミュニケーション提供システム。
- 前記データベースは、前記第1の利用者の音声データを含み、前記音声データから音素データを生成する音素データ生成モジュールと、
前記音素データを用いて会話の音声を生成する音声生成モジュールと、を含む、請求項1に記載のコミュニケーション提供システム。 - 前記人工知能としての機能を有する情報処理モジュールは、前記質問に対する前記回答の頻度を解析し、頻度の高い回答に対応する質問を前記類似データベースに記録する、請求項7に記載のコミュニケーション提供システム。
- 前記人工知能としての機能を有する情報処理モジュールは、前記第1の利用者のソーシャルデータに含まれるテキストデータの構文解析をし、質問に対する回答を推論し又は学習して決定する、請求項1に記載のコミュニケーション提供システム。
- 前記第2の利用による質問に対し、前記情報処理モジュールで回答を生成するか否かの設定をする機能を有する、請求項1に記載のコミュニケーション提供システム。
- 前記情報処理モジュールは、第2の利用者の質問に対する回答候補を複数生成し、前記第1の利用者が、前記質問に対する前記回答候補を編集する編集モジュールと、前記編集モジュールで編集された内容に基づいて前記質問に対する回答の内容を更新する更新モジュールと、を有する、請求項1に記載のコミュニケーション提供システム。
- 第1の利用者の登録情報に基づいて、仮想的な前記第1の利用者の個人像を生成し、
第2の利用者が前記仮想的な個人に対して送信した質問を受け付け、
前記質問に対する回答を、前記第1の利用者のソーシャルデータ、複数の利用者のソーシャルデータ又は予め登録されている複数のテキストデータのいずれか一種に基づいて推論し又は学習して決定し、
前記決定された回答を、前記第2の利用者に提供することを特徴とするコミュニケーション提供方法。 - 前記質問に対する回答を、第1に前記第1の利用者のソーシャルデータを検索し、適切な回答を得られないとき、第2に前記複数の利用者のソーシャルデータを検索し、適切な回答を得られないとき、第3に前記予め登録されている複数のソーシャルデータを検索する、請求項21に記載のコミュニケーション提供方法。
- 前記質問に対する回答を、前記複数の利用者のソーシャルデータから、質問に対する回答を傾向分析し、前記質問に対する回答を推論し又は学習して決定する、請求項22に記載のコミュニケーション提供方法。
- 前記傾向分析は、前記質問に対する前記複数の利用者の前記回答の時間的な変化を分析する、請求項23に記載のコミュニケーション提供方法。
- 前記第1の利用者のソーシャルデータ及び前記複数の利用者のソーシャルデータは、テキストデータ並びに、音声データ、写真データおよび映像データの一以上から生成されたテキストデータである、請求項21に記載のコミュニケーション提供方法。
- 類似する質問を類似質問としてグループ化し、前記類似質問に対応する回答を対応付けて類似データベースに記憶させ、
前記質問に対する回答を、前記類似データベースから前記質問の回答を選択する、請求項21に記載のコミュニケーション提供方法。 - 前記類似データベースに、前記第1の利用者のソーシャルデータの中から質問と回答の内容を類型化したデータが記憶させ、前記複数の利用者のソーシャルデータの中から質問と回答の内容を類型化したデータが記憶させ、前記テキストデータの中から質問と回答の内容を類型化したデータが記憶させる、請求項26に記載のコミュニケーション提供方法。
- 前記第2の利用者が前記回答の評価した情報を取得して、前記回答に評価値を付ける、請求項21に記載のコミュニケーション提供方法。
- 前記回答の評価結果を、前記第1の利用者に通知する、請求項28に記載のコミュニケーション提供方法。
- 前記通知を受けた第1の利用者が、前記質問に対する前記回答の内容を編集した内容を受け付け、前記編集された内容に基づいて前記質問に対する回答の内容を更新する、請求項29に記載のコミュニケーション提供方法。
- 前記評価結果に基づいて優先準位を付け、前記優先順位に基づいて前記質問に対する適切な回答を推論し又は学習して決定する、請求項29に記載のコミュニケーション提供方法。
- 前記第1の利用者のソーシャルデータ、前記第2の利用者のソーシャルデータ、前記テキストデータを経時的に更新し、蓄積する、請求項21に記載のコミュニケーション提供方法。
- 前記第1の利用者の三次元画像データを生成する、請求項21に記載のコミュニケーション提供方法。
- 前記第1の利用者の音声データを記憶して、前記音声データから音素データを生成し、前記音素データを用いて会話の音声を生成する、請求項21に記載のコミュニケーション提供方法。
- 前記質問に対する前記回答の頻度を解析し、頻度の高い回答に対応する質問を前記類似データベースに記録する、請求項26に記載のコミュニケーション提供方法。
- 前記第1の利用者のソーシャルデータに含まれるテキストデータの構文解析をし、質問に対する回答を推論し又は学習して決定する、請求項21に記載のコミュニケーション提供方法。
- 前記第2の利用による質問に対し、回答を生成するか否かの設定する、請求項21に記載のコミュニケーション提供方法。
- 第2の利用者の質問に対する回答候補を複数生成し、前記第1の利用者が、前記質問に対する前記回答候補を編集し、前記編集された内容に基づいて前記質問に対する回答の内容を更新する、請求項21に記載のコミュニケーション提供方法。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2016566534A JP6502965B2 (ja) | 2014-12-26 | 2015-12-25 | コミュニケーション提供システム及びコミュニケーション提供方法 |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US201462096951P | 2014-12-26 | 2014-12-26 | |
US62/096,951 | 2014-12-26 | ||
JP2015-008988 | 2015-01-20 | ||
JP2015008988 | 2015-01-20 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2016104736A1 true WO2016104736A1 (ja) | 2016-06-30 |
Family
ID=56150748
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2015/086309 WO2016104736A1 (ja) | 2014-12-26 | 2015-12-25 | コミュニケーション提供システム及びコミュニケーション提供方法 |
Country Status (2)
Country | Link |
---|---|
JP (1) | JP6502965B2 (ja) |
WO (1) | WO2016104736A1 (ja) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2017217507A1 (ja) * | 2016-06-16 | 2017-12-21 | 株式会社オルツ | コミュニケーションを支援する人工知能システム |
WO2018061776A1 (ja) * | 2016-09-29 | 2018-04-05 | 株式会社東芝 | 情報処理システム、情報処理装置、情報処理方法、及び記憶媒体 |
JP2018055548A (ja) * | 2016-09-30 | 2018-04-05 | 株式会社Nextremer | 対話装置、学習装置、対話方法、学習方法、およびプログラム |
CN109918525A (zh) * | 2019-03-12 | 2019-06-21 | 同济大学 | 基于微信小程序的食物图片美学分析标签数据收集系统 |
WO2019138567A1 (ja) * | 2018-01-15 | 2019-07-18 | 富士通株式会社 | 出力制御プログラム、出力制御方法、及び出力制御装置 |
JP2020071679A (ja) * | 2018-10-31 | 2020-05-07 | 株式会社リクルート | 質問に対する回答を支援するためのシステム、方法、及びプログラム |
JP6837614B1 (ja) * | 2020-05-28 | 2021-03-03 | 三菱電機株式会社 | 推論装置、プログラム及び推論方法 |
WO2022034922A1 (ja) * | 2020-08-14 | 2022-02-17 | 株式会社オルツ | アンケート調査のためのシステム、プログラム、および方法 |
JP2022032935A (ja) * | 2020-08-14 | 2022-02-25 | 株式会社オルツ | アンケート調査のためのシステム、プログラム、および方法 |
JP7078307B1 (ja) | 2022-01-14 | 2022-05-31 | 望 窪田 | 学習モデルの個別化 |
JP2022166260A (ja) * | 2022-03-02 | 2022-11-01 | ベイジン バイドゥ ネットコム サイエンス テクノロジー カンパニー リミテッド | マルチシステムに基づくインテリジェントな質問応答方法、装置及び機器 |
US11494354B2 (en) | 2019-11-08 | 2022-11-08 | Fujifilm Business Innovation Corp. | Information management apparatus, information processing apparatus, and non-transitory computer readable medium |
JP2023082970A (ja) * | 2021-12-03 | 2023-06-15 | 株式会社ビデオリサーチ | 専門家回答システム |
JP7295207B1 (ja) | 2021-12-03 | 2023-06-20 | 株式会社ビデオリサーチ | 調査集計システム |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2005038088A (ja) * | 2003-07-17 | 2005-02-10 | Toshiba Corp | 質問応答装置及び質問応答方法 |
JP2005196356A (ja) * | 2004-01-05 | 2005-07-21 | Nec Corp | 擬似人格対話システム、方法、およびプログラム |
JP2008305322A (ja) * | 2007-06-11 | 2008-12-18 | C2Cube Inc | 評判コーパス生成装置、対象情報コーパス生成装置、サーバ装置、バーチャルチャット装置、およびプログラム |
JP2013210792A (ja) * | 2012-03-30 | 2013-10-10 | Hitachi Solutions Ltd | 授業質問回答システム |
-
2015
- 2015-12-25 WO PCT/JP2015/086309 patent/WO2016104736A1/ja active Application Filing
- 2015-12-25 JP JP2016566534A patent/JP6502965B2/ja not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2005038088A (ja) * | 2003-07-17 | 2005-02-10 | Toshiba Corp | 質問応答装置及び質問応答方法 |
JP2005196356A (ja) * | 2004-01-05 | 2005-07-21 | Nec Corp | 擬似人格対話システム、方法、およびプログラム |
JP2008305322A (ja) * | 2007-06-11 | 2008-12-18 | C2Cube Inc | 評判コーパス生成装置、対象情報コーパス生成装置、サーバ装置、バーチャルチャット装置、およびプログラム |
JP2013210792A (ja) * | 2012-03-30 | 2013-10-10 | Hitachi Solutions Ltd | 授業質問回答システム |
Non-Patent Citations (1)
Title |
---|
MASAHIRO NAKAGAWA: "ATR Ken Onsei Hon'yaku ni Awase Kuchimoto no Ugoki o Saigen", GEKKAN COMPUTER DIGEST, vol. 26, no. 12, 10 December 2000 (2000-12-10), pages 68 * |
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11526720B2 (en) | 2016-06-16 | 2022-12-13 | Alt Inc. | Artificial intelligence system for supporting communication |
WO2017217507A1 (ja) * | 2016-06-16 | 2017-12-21 | 株式会社オルツ | コミュニケーションを支援する人工知能システム |
WO2018061776A1 (ja) * | 2016-09-29 | 2018-04-05 | 株式会社東芝 | 情報処理システム、情報処理装置、情報処理方法、及び記憶媒体 |
JP2018055422A (ja) * | 2016-09-29 | 2018-04-05 | 株式会社東芝 | 情報処理システム、情報処理装置、情報処理方法、及びプログラム |
JP2018055548A (ja) * | 2016-09-30 | 2018-04-05 | 株式会社Nextremer | 対話装置、学習装置、対話方法、学習方法、およびプログラム |
US11301466B2 (en) | 2018-01-15 | 2022-04-12 | Fujitsu Limited | Computer-readable recording medium recording output control program, output control method, and information processing apparatus |
WO2019138567A1 (ja) * | 2018-01-15 | 2019-07-18 | 富士通株式会社 | 出力制御プログラム、出力制御方法、及び出力制御装置 |
JPWO2019138567A1 (ja) * | 2018-01-15 | 2020-12-03 | 富士通株式会社 | 出力制御プログラム、出力制御方法、及び出力制御装置 |
JP2020071679A (ja) * | 2018-10-31 | 2020-05-07 | 株式会社リクルート | 質問に対する回答を支援するためのシステム、方法、及びプログラム |
CN109918525A (zh) * | 2019-03-12 | 2019-06-21 | 同济大学 | 基于微信小程序的食物图片美学分析标签数据收集系统 |
US11494354B2 (en) | 2019-11-08 | 2022-11-08 | Fujifilm Business Innovation Corp. | Information management apparatus, information processing apparatus, and non-transitory computer readable medium |
JP6837614B1 (ja) * | 2020-05-28 | 2021-03-03 | 三菱電機株式会社 | 推論装置、プログラム及び推論方法 |
WO2021240721A1 (ja) * | 2020-05-28 | 2021-12-02 | 三菱電機株式会社 | 推論装置、プログラム及び推論方法 |
WO2022034922A1 (ja) * | 2020-08-14 | 2022-02-17 | 株式会社オルツ | アンケート調査のためのシステム、プログラム、および方法 |
JP2022032935A (ja) * | 2020-08-14 | 2022-02-25 | 株式会社オルツ | アンケート調査のためのシステム、プログラム、および方法 |
JP7101357B2 (ja) | 2020-08-14 | 2022-07-15 | 株式会社オルツ | アンケート調査のためのシステム、プログラム、および方法 |
JP2023114460A (ja) * | 2021-12-03 | 2023-08-17 | 株式会社ビデオリサーチ | 専門家回答予測システム |
JP7352761B2 (ja) | 2021-12-03 | 2023-09-28 | 株式会社ビデオリサーチ | 専門家回答予測システム |
JP2023082970A (ja) * | 2021-12-03 | 2023-06-15 | 株式会社ビデオリサーチ | 専門家回答システム |
JP7295207B1 (ja) | 2021-12-03 | 2023-06-20 | 株式会社ビデオリサーチ | 調査集計システム |
JP2023089330A (ja) * | 2021-12-03 | 2023-06-28 | 株式会社ビデオリサーチ | 調査集計システム |
JP7078307B1 (ja) | 2022-01-14 | 2022-05-31 | 望 窪田 | 学習モデルの個別化 |
JP2023103675A (ja) * | 2022-01-14 | 2023-07-27 | 望 窪田 | 学習モデルの個別化 |
JP2022166260A (ja) * | 2022-03-02 | 2022-11-01 | ベイジン バイドゥ ネットコム サイエンス テクノロジー カンパニー リミテッド | マルチシステムに基づくインテリジェントな質問応答方法、装置及び機器 |
JP7548967B2 (ja) | 2022-03-02 | 2024-09-10 | ベイジン バイドゥ ネットコム サイエンス テクノロジー カンパニー リミテッド | マルチシステムに基づくインテリジェントな質問応答方法、装置及び機器 |
Also Published As
Publication number | Publication date |
---|---|
JP6502965B2 (ja) | 2019-04-17 |
JPWO2016104736A1 (ja) | 2017-11-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6502965B2 (ja) | コミュニケーション提供システム及びコミュニケーション提供方法 | |
US11526720B2 (en) | Artificial intelligence system for supporting communication | |
CN112292674B (zh) | 为助理系统处理多模态用户输入 | |
Dame | Making a name for yourself: Tagging as transgender ontological practice on Tumblr | |
US11829725B2 (en) | Computer implemented method for the automated analysis or use of data | |
Purohit et al. | Emergency-relief coordination on social media: Automatically matching resource requests and offers | |
Baker et al. | ‘Why do white people have thin lips?’Google and the perpetuation of stereotypes via auto-complete search forms | |
US20170344532A1 (en) | Method and system for creating interactive inquiry and assessment bots | |
Lotfian et al. | Formulating emotion perception as a probabilistic model with application to categorical emotion classification | |
US10770072B2 (en) | Cognitive triggering of human interaction strategies to facilitate collaboration, productivity, and learning | |
JP6831522B2 (ja) | コミュニケーションシステム | |
Aattouri et al. | Modeling of an artificial intelligence based enterprise callbot with natural language processing and machine learning algorithms | |
Wu et al. | Toward predicting active participants in tweet streams: A case study on two civil rights events | |
Yenkikar et al. | Sentimlbench: Benchmark evaluation of machine learning algorithms for sentiment analysis | |
Eichinger et al. | Affinity: A system for latent user similarity comparison on texting data | |
CN113806620A (zh) | 内容推荐方法、设备、系统及存储介质 | |
Zanzotto et al. | Have you lost the thread? discovering ongoing conversations in scattered dialog blocks | |
Alsulami et al. | Extracting attributes for twitter hashtag communities | |
Saravanos et al. | Discovering Influential Twitter Authors Via Clustering And Ranking On Apache Storm | |
CN110083785A (zh) | 基于用户搜索记录的性别年龄判别方法及装置 | |
Ishanka | A context-aware travel recommendation system based on user emotion and personality | |
Pulikonda et al. | Exploring the Applications, Challenges, and Issues of Sentiment Analysis | |
Hauffa | Topic-Driven Characterization of Social Relationships for the Analysis of Social Influence | |
Srinidhi et al. | Sentiment Analysis of Social Media Comments using Natural Language Procesing | |
Alharthi | The Use of Items Personality Profiles in Recommender Systems |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 15873304 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2016566534 Country of ref document: JP Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC , EPO FORM 1205A DATED 24.01.18. |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 15873304 Country of ref document: EP Kind code of ref document: A1 |