JP6691340B2 - Question answering system, question acceptance answering system, primary answering system, and question answering method using them - Google Patents

Description

  The present invention relates to a question answering system that accepts a question input from a terminal device of a specific or unspecified end user and provides an answer to the question, more specifically, a question answering system using artificial intelligence. The present invention relates to a question acceptance answering system and a primary answering system having artificial intelligence, which cooperate to form the question answering system, and a question answering method using them.

  In recent years, in order to provide the optimum answer to a question, a specific end user such as a telephone operator answers a question from a customer, in order to provide an optimal answer to the question, a past response history is searched to promptly obtain a more appropriate answer. Answer assistance system that uses artificial intelligence to support users and various responses that automatically respond to questions from unspecified end users (questioners) using artificial intelligence without human intervention. The system is being developed. Furthermore, research is being conducted on natural language processing technology that analyzes the syntax and meaning of a question entered by the questioner to retrieve the answer to the question, and the extraction of the optimal answer by artificial intelligence.

In the following description in this specification, an unspecified or specific end user who asks a question is referred to as a “questioner”, and a person who introduces and uses a question answering system is referred to as a “system user”. A person who provides a question answering system to a user is called a "system provider".
In addition, the terminal device used by the end user for inquiring is referred to as a "question terminal", and the system serving as the platform for the question acceptance page through which the question answering system is accessed from the question terminal is referred to as "user system". ". In addition, the system administrator who acquires and analyzes the log information of the question answering system used in-house as a system user, maintenance, etc. is called the "user system administrator", and operates the entire system as a system provider. A person who performs overall management such as maintenance and system improvement is simply referred to as a "system administrator".

For example, in Patent Document 1, an answer candidate for a question is searched, a score indicating the degree to which each answer candidate is a correct answer for the question is calculated for each combination of the question and the answer candidate, and the calculated score is set to the calculated score. Based on this, a question answering system configured to output an answer candidate that is most likely to be a correct answer to a question is disclosed.
Further, Patent Document 2 discloses a request content identification system that analyzes an input sentence in which a request from a user is described in natural language and determines whether the input request content corresponds to predefined request content. Has been done.

  Further, Patent Document 3 discloses an information providing system including an information providing server and a suggest server. When the user inputs a question, the suggest server displays a question candidate that may be input, and assists the user in inputting a question. Further, the system of Patent Document 3 is configured so that the information management system can be maintained by the system management terminal.

JP, 2013-171550, A Patent No. 5099367 JP, 2016-99967, A

However, in the systems of Patent Document 1 and Patent Document 2, it is necessary to collect a large amount of knowledge data and perform machine learning in order to improve the accuracy of answering a question. In addition, even in the operation management after introducing the question answering system, a high degree of knowledge about artificial intelligence etc. was required.
Therefore, in introducing the question answering system, much time, labor, and huge cost were required for collecting huge amount of knowledge data and machine learning before starting operation. Furthermore, even after the system starts operating, it takes a lot of time and cost for machine learning and system maintenance to update the knowledge of artificial intelligence. It was necessary to secure human resources with specialized knowledge about intelligence.

  In Patent Document 3, in view of the problem in operation management after system operation, in the maintenance mode, the system administrator uses the system management terminal to receive the response data log or the like from the response content determination means, The input is simulated so that the validity of the knowledge data can be verified. Further, according to the invention of Patent Document 3, even if the system administrator does not have highly specialized knowledge about artificial intelligence, the assumed question character string and the response character string for artificial intelligence are changed in the maintenance mode. It will help you to do so.

  However, the input of the expected question sentence and the change of the response sentence to it change the knowledge data itself of the artificial intelligence. Therefore, even in the system of Patent Document 3, a high level of knowledge is not required for collecting and checking log data and simulating input, but the system administrator is required to have a certain degree of specialized skill regarding artificial intelligence. (See the description in the latter part of (0048) of Patent Document 3). In addition, the change of the knowledge data of the artificial intelligence requires re-learning for reconstructing the neural network, which requires a lot of time and labor in addition to the specialized knowledge of the artificial intelligence. Therefore, in the conventional question answering system using artificial intelligence, a system user who does not have specialized knowledge cannot directly participate in improving the question answering system introduced in-house and gradually growing it. It was In addition, in the conventional question answering system using artificial intelligence, there is no simple mechanism to flexibly change the concrete answer contents to the question while actually operating the system. It was difficult to evolve flexibly to meet the demands of.

  Under the above circumstances, even if a small-to-medium-sized company with limited financial resources wishes to introduce a question answering system, the conventional question answering system would require enormous cost for introduction, and it would be necessary to secure specialized human resources for system operation management. The hurdles of system introduction were extremely high because of the major obstacles such as the necessity. Moreover, even if the introduction barrier could be overcome, it was not possible for the system user to directly improve the question answering system to meet the changing needs of the company and grow the system. Furthermore, in the conventional question answering system, it is not possible to deal with questions in a plurality of types of languages only by adding a simple configuration.

The present invention has been made in view of the problems and social needs of the conventional question answering system as described above, can reduce the system introduction cost, and can easily change the final answer after the start of system operation. A question answering system utilizing artificial intelligence capable of improving and evolving a question answering system, a question acceptance answering system constituting the question answering system, a primary answering system, and a question answering method using them. The purpose is to provide.
Another object of the present invention is to have a simple structure capable of accepting questions in a variety of languages with a simple structure and providing a final answer to a question terminal in a predetermined answer language according to the language of the question. It is an object of the present invention to provide a question answering system, a question acceptance answering system constituting the question answering system, a primary answering system, and a question answering method using them.
Further, another object of the present invention is to use a question answering system utilizing artificial intelligence that can be shared by a plurality of users and reduce the introduction cost, a question acceptance answering system constituting the question answering system, and a primary answer. An object is to provide a system and a question answering method using them.

  The present invention has a configuration in which a question answering system is separated into a question accepting answering system for accepting a question and returning a final answer to the question to a questioning terminal, and a primary answering system having artificial intelligence. A first answer to a question is provided as an identifier in response to a request from the question receiving and answering system, and the question receiving and answering system associates a final answer candidate with the identifier of the first answer received from the first answering system. The final answer candidate is read out based on the first answer from the first answer system, and the final answer is returned to the question terminal device.

A question answering system according to a first aspect of the present invention includes a question acceptance answering system using a computer that accepts a question from a questioning terminal and returns a final answer to the question to the questioning terminal, and artificial intelligence. A question answering system that is provided separately from a primary answering system that returns a primary answer to a question in response to a request,
The first answer system is
A first response to the question from the artificial intelligence is input to the artificial intelligence by accepting a reply request for the primary answer to the question from the question acceptance and response system, and the first answer to the question from the artificial intelligence. Interface means for transmitting the next answer to the question acceptance answer system,
By machine learning a plurality of assumed questions and a predetermined identifier as a primary answer corresponding to the assumed questions, a first answer to the question from the question acceptance / answer system in response to a request from the question acceptance / answer system. Artificial intelligence for outputting a primary answer as the identifier,
The question acceptance answer system is
Receiving means for receiving a question from the question terminal,
A primary answer inquiry that transmits the question accepted by the question acceptance means to the primary answer system, requests the primary answer, and receives the primary answer from the primary answer system. Means and
A final answer storage unit storing a final answer candidate corresponding to the identifier in a 1: 1 association with the identifier machine-learned by the artificial intelligence;
Based on the primary response received from the primary response system, the final response candidate corresponding to the identifier of the primary response is read from the final response storage unit, and the final response candidate is read out. A final answer means for returning a final answer to the inquiry terminal based on the above.

  Here, a general-purpose artificial intelligence system provided for the purpose of commercial service can be used as the primary response system for providing the primary response, and the question acceptance response system and the primary response system are It can be connected via a communication network. By separating the question answering system into the question accepting answering system and the primary answering system in this way, since the question accepting answering system is separated from the artificial intelligence, the artificial intelligence machine of the primary answering system is separated. Learning can be designed separately from the final answer candidates of the question acceptance answering system. In addition to machine learning of artificial intelligence, it is also possible to make corrections, changes, and additions independently of the hit rate of answers to questions.

A question acceptance / answer system according to a first aspect of the present invention includes a question acceptance / answer system using a computer that accepts a question from a question terminal, an artificial intelligence, a plurality of assumed questions, and a plurality of assumed questions corresponding to the assumed questions. In a question answering system, which is machine-learned with a predetermined identifier as a primary answer, and which is separately provided with a primary answering system which outputs a primary answer to a question from the question acceptance answering system as the identifier. It is a question acceptance answer system,
Question accepting means for accepting a question from the question terminal,
A primary answer inquiry that transmits the question accepted by the question acceptance means to the primary answer system, requests the primary answer, and receives the primary answer from the primary answer system. Means and
A final answer storage unit that stores a final answer candidate corresponding to the identifier in a 1: 1 association with the machine-learned identifier in the primary answer system;
Based on the primary response received from the primary response system, the final response candidate corresponding to the identifier of the primary response is read from the final response storage unit, and the final response candidate is read out. A final answer means for returning a final answer to the question terminal based on
It is characterized by including.
The invention according to this aspect is a sub-combination invention which constitutes the question acceptance answering system in which the question acceptance answering system and the primary answering system described above are combined.

  Furthermore, in the question answering system or question acceptance answering system according to another aspect of the present invention, the artificial intelligence of the primary answering system includes a plurality of the one or more assumed questions having a common attribute as one group. Machine learning is performed so that a large number of the expected questions that form a question group and one identifier corresponding to the question group in a 1: 1 relationship are used as the primary answers of the expected questions included in each question group. Is characterized by.

In the question answering system or the question acceptance answering system, the question acceptance means can be configured to further acquire unique attribute information of the question associated with the question from the question terminal, and the final answer storage unit If necessary, it is possible to store different final answer candidates corresponding to the unique attribute information.
The unique attribute information includes identification information of the user system, language information indicating the language type of the question, and the like. The identification information of the user system can be acquired based on the communication route information when the question is accepted and the information added from the user system in the communication route. The language information of the question can be obtained by a well-known technique such as being obtained from the language designation information such as the OS of the question terminal or the application for answering the question even if the language information of the question is designated when the question is input.

A question answering system or question acceptance answering system according to another aspect of the present invention, based on the acquired unique attribute information, acquires identification information of a user system through which the question has passed,
The final answer storage unit stores, as each final answer candidate, a basic answer that is a base of the final answer stored in association with each of the identifiers, or, if necessary, the identifiers and the identification information of the user system. Stored in association with each of the basic answers and unique answers with different contents,
The final answer means reads and reads the basic answer or the unique answer from the final answer storage unit based on the identifier of the primary answer and the identification information of the user system corresponding to the question. The basic answer or the unique answer is transmitted to the inquiry terminal as a final answer.
As described above, the identification information of the user system is acquired as the unique attribute information from the identification information of the user system from the communication route information through which the question reaches the question acceptance answering system or the information given in the communication process. It is possible.

Further, in a question answering system or a question acceptance answering system according to another aspect of the present invention, the artificial intelligence of the primary answering system has a plurality of hypothetical questions having common attributes in different languages in one question group. The primary answers to all the expected questions that are composed of a plurality of question groups including and the plurality of expected questions that are included in the same question group and are in different languages are Machine learning is performed to output the same primary answer regardless of the type, and when a question is received from the question acceptance answering system, the question is answered as the primary answer to the question. It is configured to output the identifier to be output to the interface means.
By making the artificial intelligence of the primary answering system machine-learn the answers of a plurality of supposed questions in different languages having a common attribute so as to output the same single primary answer, the languages having a common attribute differ. The same primary answer is output for the question. Since the final answer candidate is stored in association with the primary answer, if the primary answer is the same, the same final answer will be sent to the question terminal even if the question language is different. Become. The answer language of the final answer is the same regardless of the language of the question All the same final answers are returned, or if the question language is a foreign language, the final answer in English is also returned. The final answer in the same answer language as the question language may be returned.
Here, “a plurality of supposed questions having common attributes in different languages” includes supposed questions of different languages and having substantially the same meaning.

In a question answering system or a question acceptance answering system according to another aspect of the present invention, the question acceptance answering system obtains the question language information included in the unique attribute information of the received question by the question accepting means, The final answer storage unit stores the final answer candidates in a plurality of languages having substantially the same content in an answer language determined based on the question language information in association with the identifier and the type of the answer language. Cage,
The final answer means, to obtain the information of the answer language corresponding to the question based on the acquired question language information, based on the identifier of the obtained primary answer, and the information of the answer language, The final answer candidate in the answer language corresponding to the question is read out, and the final answer in the answer language corresponding to the question language is returned as the final answer corresponding to the question.
With this configuration, when the language of the question is a foreign language, the answer language can be changed to English or another language, and the final answer can be returned in the same answer language as the question language.

  In a question answering system or a question acceptance answering system according to another aspect of the present invention, the primary answering system extracts a primary answer to the question and outputs an answer certainty factor thereof, However, the basic answer or the unique answer stored in association with the primary answer in accordance with the certainty factor of the answer received from the primary answer system is read from the final answer storage unit to answer the question. The final answer is transmitted to the inquiry terminal.

In addition, a question answering system or a question acceptance answering system according to another aspect of the present invention further includes a question record such as a contact record with the question terminal, a contact record with the primary answer system, and question unique attribute information. An operation history storage unit that stores various information up to the final answer, and a user system browses the various information stored in the operation history storage unit according to the authority and stores the information in the final answer storage unit. Maintenance means capable of maintenance including correction, addition and change of the final answer candidate or the unique answer.
With this configuration, the primary answering system and the question answering system can be easily maintained. Especially for maintenance of the question answering system, knowledge of artificial intelligence is not required, and it is possible to easily improve the system by changing the basic answer and the unique answer which are final answer candidates.
Further, the maintenance means limits the content of the final answer candidate, the basic answer, or the unique answer stored in the final answer storage unit to the supposed question of the primary answer system according to authority. It can also be configured to allow additions, modifications, changes, deletions, etc. without updating.

  In the operation history storage unit, the number of questions / answers for the entire system and each system user, the average answer confidence, the number and ratio of the answer confidence below the reference value, the number of questions and answers for each primary response identifier, and the average It is possible to include information such as an answer certainty factor, a question list with a low answer certainty factor, and a list of answers. These pieces of information can be browsed by the maintenance means, and it becomes possible to easily grasp the operating status of the system such as the usage status of the system and the accuracy of the reply based on the information. Further, since the question acceptance answer system and the final answer storage unit are separated from the primary answer system, the final answer stored in the final answer storage unit by the maintenance means even if there is no expert knowledge about artificial intelligence. And amendment (in the present specification, unless specifically stated to the contrary, the term “amendment” includes not only partial correction or alteration, but also addition of Can be used as). Therefore, not only system administrators but also system users who do not have a high level of knowledge about artificial intelligence can correct, change, or add the final answer such as basic answer or unique answer within the given authority. I was able to do it.

  Further, the maintenance means may be configured to add, to the primary answer system, a machine learning function by correcting, changing, or adding an assumed question, a common attribute, a question group, or a primary answer. With such an operation history storage unit and maintenance means, it is possible to output a list that extracts records of questions and answers with a low answer certainty, and to assume that the answer certainty is low, depending on the analysis data from aggregation or statistical data. It becomes possible to consolidate questions and groups of questions and improve and evolve them into a question answering system that provides more appropriate answers.

A first answer system according to a first aspect of the present invention includes a question acceptance and answer system using a computer that accepts a question from a question terminal and returns a final answer to the question to the question terminal, and a plurality of assumed questions. And a primary answer system having artificial intelligence for outputting the primary answer to the question from the question acceptance answer system as the identifier by machine-learning a predetermined identifier as a primary answer corresponding to the assumed question. A primary answering system in a question answering system comprising:
A question from the question acceptance and answer system and a request for the first answer are accepted, the question is input to the artificial intelligence, and a primary answer to the question from the artificial intelligence is given to the question acceptance and answer system. Interface means for sending
A plurality of expected questions in different languages and a plurality of the expected questions as a plurality of question groups in which the plurality of the expected questions having a common attribute regardless of the language are one group and each of the question groups are 1: Each identifier corresponding to 1 is machine-learned as a primary answer common to the plurality of assumed questions belonging to the individual question group, and when the question is received from the question acceptance answer system, the question is answered. Artificial intelligence that outputs the first answer to the interface means as the identifier.

The present invention relates to the primary answering system which constitutes the above-mentioned question acceptance answering system, and is a sub-combination invention of the question answering system.
According to this configuration, the primary answering system is made to perform machine learning so as to output one common primary answer (identifier) for an assumed question having a common attribute even if the question language is different. As a result, the same identification number is output as the primary answer for a question having a common attribute even if the question language is different. Therefore, even if the question language is different, the same final answer can be returned to the question terminal. Since it is relatively difficult to create an accurate question document in a language other than the native language, it becomes possible to make an accurate question by asking only the question in the native language, and it is possible to obtain an accurate answer. On the other hand, in general, the ability to read and understand documents is higher than the ability to create documents, so there is a high possibility that the answer language can be understood even if it is not the native language.
In addition, the answer language for the final answer is decided in advance according to the language of the question, and the final answer candidate for each answer language is stored, so that the final answer language is decided according to the language information of the question. It can also be configured to send back an answer.

A question answering method of a question answering system according to a first aspect of the present invention comprises a question acceptance answering system for accepting a question from a questioning terminal by a computer and returning an answer to the question to the questioning terminal, and artificial intelligence. In this case, a plurality of expected questions and a predetermined identifier that is an answer to the plurality of expected questions are machine-learned as a primary answer, and the identifier is returned as a primary answer according to the input question. A question answering method in a question answering system comprising a system,
(A) in the question acceptance and answer system, accepting a question from a question terminal and transmitting the question to a primary answer system;
(B) In the primary answer system, accepts the question from the question acceptance answer system, inputs the question to the artificial intelligence, and outputs the primary answer to the input question as the identifier. And transmitting the identifier of the primary answer output from the artificial intelligence to the question acceptance answer system,
(C) In the question receiving and answering system, when the identifier is received as the primary answer from the primary answering system, the final answer candidate is associated with the identifier that the artificial intelligence is machine learning. Reading the final answer candidate corresponding to the received identifier from the final answer storage unit that stores, and returning the read final answer candidate as a final answer to the question terminal,
It is characterized by including.

A question answering method of a question acceptance answering system according to a first aspect of the present invention comprises a question acceptance answering system for accepting a question from a questioning terminal by a computer and returning an answer to the question to the questioning terminal, and artificial intelligence. A plurality of expected questions and a predetermined identifier that is an answer to the plurality of expected questions are machine-learned as a primary answer, and the identifier is returned as a primary answer according to the input question. A question answering method of the question acceptance answering system in a question answering system comprising an answering system,
(A1) In the question acceptance and answer system, a step of accepting a question from a question terminal and transmitting the question to the first answer system,
(B1) When the identifier is received from the primary response system as the primary response, the final response is stored as a final response candidate in association with the identifier machine-learned by the artificial intelligence. Reading the final answer candidate corresponding to the received identifier from the storage unit and returning the read final answer candidate to the question terminal as a final answer;
It is characterized by being equipped with.

In the question answering method of the question answering system or question acceptance answering system according to another aspect of the present invention, the step (a) or (a1) is
(A-1) a step of accepting a question from a question terminal,
(A-2) a step of acquiring and storing unique attribute information of the question associated with the question,
(A-3) sending the question to the primary answering system,
Equipped with
The step (c) or (b1) includes
(C-1) receiving the identifier as the primary response from the primary response system,
(C-2) a step of acquiring the identification information of the user system from the unique attribute information stored in the step (a-1),
(C-3) Identification of the primary answer and the user system from the answer storage unit that stores the final answer candidate corresponding to the identifier by dividing it into a basic answer and a user-specific unique answer. Reading the corresponding basic answer or the specific answer based on information;
(C-4) returning the read basic answer or unique answer to the question terminal,
It is characterized by including.

A multilingual question answering method of a question answering system according to a first aspect of the present invention includes a question acceptance answering system for accepting a question from a questioning terminal by a computer and returning an answer to the question to the questioning terminal, and an artificial intelligence. And all said expected questions consisting of multiple question groups containing multiple expected questions having common attributes of different languages in one question group and different languages included in the same question group The primary answers to the plurality of expected questions are machine-learned so that the primary answers having the same identifier are output to all of the expected questions regardless of the language type. A multilingual question answering method in a question answering system comprising a primary answering system that returns the identifier as a primary answer according to Te,
(A) In the question receiving and answering system, a question from a question terminal is received, language information of the question is acquired and stored from unique attribute information of the question associated with the question, and the question and the primary If the answering system needs the language information of the question, sending the question together with the language information of the question to the primary answering system;
(B) In the primary answer system, accepts the question from the question acceptance answer system, inputs the question to the artificial intelligence, and outputs the primary answer to the input question as the identifier. And transmitting the identifier of the primary answer output from the artificial intelligence to the question acceptance answer system,
(C) In the question receiving and answering system, when the identifier is received as the primary answer from the primary answering system, the final answer candidate is associated with the identifier that the artificial intelligence is machine learning. Reading the final answer candidate corresponding to the received identifier from the answer storage unit storing, and returning the read final answer candidate as a final answer to the question terminal,
It is characterized by including.
Here, “a plurality of supposed questions having common attributes in different languages” includes supposed questions of different languages and having substantially the same meaning.

A multilingual question acceptance and response method of a question acceptance and answer system according to a first aspect of the present invention includes a question acceptance and answer system that accepts a question from a question terminal by a computer and returns an answer to the question to the question terminal. It has artificial intelligence, and all the expected questions that are composed of multiple question groups that include multiple expected questions that have a common attribute of different languages in one question group, and different languages that are included in the same question group. The primary answers to the plurality of expected questions consisting of are all machine-learned so as to output the primary answers having the same identifier to any of the expected questions regardless of the language type. A question answering and answering system in a question answering system, which comprises a primary answering system that returns an identifier as a primary answer in response to a question A question accepted response method,
(A1) In the question acceptance and answer system, the question from the question terminal is accepted, the language information of the question is acquired and stored from the unique attribute information of the question associated with the question, and the question and the first If the next answer system requires the language information of the question, sending the question together with the language information of the question to the first answer system;
(B1) When the identifier is received from the primary response system as the primary response, the final response is stored as a final response candidate in association with the identifier machine-learned by the artificial intelligence. Reading the final answer candidate corresponding to the received identifier from the storage unit and returning the read final answer candidate to the question terminal as a final answer;
It is characterized by including.
The present invention is a sub-combination invention of the invention according to claim 16. Here, “a plurality of supposed questions having common attributes in different languages” includes supposed questions of different languages and having substantially the same meaning.

According to another aspect of the present invention, a question answering system or a question accepting and answering system of a multilingual question accepting and answering method comprises the step (c) or (b1):
(C-1) receiving the identifier as the primary response from the primary response system,
(C-2) a step of obtaining answer language information determined according to a decision rule of an answer language based on the language information stored in the step (a) or (a1),
(C-3) The final answer candidate stored in the final answer storage unit based on the answer language designated by the answer language information, which is associated with the machine-learned identifier and the answer language information, is the primary answer. Reading the final answer candidate corresponding to the identifier received as an answer and the acquired answer language information;
(C-4) returning the read final answer candidate in the answer language to the question terminal as a final answer,
It is characterized by including.

The answering method of the primary answering system according to the first aspect of the present invention is a question accepting and answering system that accepts a question from a questioning terminal and returns an answer to the questioning terminal, and responds to a request from the question accepting and answering system. And a response method of the primary answer system in a question answering system using a primary answer system having artificial intelligence for outputting the primary answer to the question as an identifier,
(A) A large number of supposed questions forming a plurality of question groups including a plurality of supposed questions having a common attribute regardless of language in one question group correspond to each of the plurality of question groups 1: 1. Causing the artificial intelligence to perform machine learning by using an identifier as a primary answer to each of the assumed questions;
(B) inputting the question received from the question receiving and answering system to the artificial intelligence, and outputting the identifier as a primary answer from the artificial intelligence;
(C) transmitting the identifier output as the first answer to the question acceptance and answer system as a first answer to the question;
In the step (a), the plurality of expected questions having common attributes included in each question group are configured to include expected questions having substantially the same meaning but different languages. Preferably.
The present invention is a sub-combination invention of the invention according to claim 16.

    In the question answering system of the present invention, the question accepting answering system and the primary answering system having artificial intelligence are separated, and the final answer candidates for the question are given to the question accepting answering system and the first answering system of the primary answering system. The answer is stored in association with it, and the final answer candidate of the question acceptance answering system is read out and the final answer is returned according to the primary answer from the primary answering system. As a result, the content of the final answer candidate can be freely modified and changed regardless of artificial intelligence. That is, the final answer can be freely changed within the scope of the predetermined authority even if there is no specialized knowledge about artificial intelligence. Therefore, for example, when the confidence level of the primary answer from the primary answer system is low. Even if there is, by changing the final answer to the question to be the final answer that matches the question, it is possible to improve and maintain the answer confidence substantially without machine learning. .

  In addition, in order to improve the accuracy of the response to the question, it is desirable to machine-learn as many possible questions as possible from the beginning of system introduction, but increasing the number of expected questions and answers to machine-learn will reduce the system introduction cost. Get higher On the other hand, in the present invention, since the final answer candidate is stored in the question acceptance answering system separate from the primary answering system, the content of the answer is expanded to a wider range than the content of directly answering the assumed question. It is possible to store the final answer candidate. As a result, in the range in which the first answer can obtain a certain degree of certainty, the answer of the assumed question that is not machine-learned by artificial intelligence is also stored in the answer content of the final answer candidate. It is possible to increase the satisfaction level of the answer. Therefore, according to the present invention, it is possible to enhance the answer content of the final answer candidate while reducing the number of assumed questions and answers to be machine-learned by the primary answer system as compared with the conventional method. In other words, it is possible to reduce the number of hypothetical questions that artificial intelligence uses for machine learning than before, reduce the cost of introducing a question answering system at the start of operation, and reduce the load of artificial learning machine learning at the start of operation. It becomes possible to do.

  In addition, in the question answering system according to another aspect of the present invention, after the question acceptance answering system and the primary answering system are separated, different final answer candidates for each system user are stored together, so System users can share one question answering system. This makes it possible for individual system users to reduce the system installation cost.

  In addition, in a question answering system according to another aspect of the present invention, the question acceptance answering system and the primary answering system are separated, and even if a plurality of supposed questions in different languages have one supposed question having a common attribute, It is possible to accept questions in different languages with a simple configuration by performing machine learning so that the same first answer can be obtained, and storing final answer candidates in different answer languages in the question acceptance and answer system. The answer language can be configured to answer in one language, or can be configured to answer in a plurality of predetermined answer languages according to the language of the question.

The conceptual diagram for demonstrating the utilization aspect of the question answering system which concerns on the 1st Embodiment of this invention. The schematic diagram which shows the structural example of other embodiment (2nd Embodiment) which concerns on the question answering system of this invention. FIG. 3A is a functional block diagram illustrating an embodiment of the question acceptance / answer system of the present invention, and FIG. 3B is a functional block showing a configuration example of the final answer means shown in FIG. The functional block diagram which shows the structural example of the maintenance means shown in FIG. The flowchart which illustrates the flow of the response with respect to the question which concerns on this invention. The figure which illustrates the specific example of a common attribute about the question groups Q1-QN which the primary answer system, such as AI, used in this invention learns, and the correspondence of a primary answer and a common attribute. The figure for exemplifying and explaining the question group and primary answer concerning the present invention. The figure for exemplifying and explaining another question group and the primary answer concerning the present invention. The figure for exemplifying and explaining another question group and the 1st answer. The figure which illustrates the relationship between the primary reply identifier memorize | stored in the final reply memory | storage part which concerns on this invention, and a final reply. The figure which illustrates the content of the basic answer of the final answer corresponding to the identifier "NNN" shown in FIG. The flowchart which illustrates the basic procedure of initial setting. Flowchart exemplifying a procedure for the vendor to maintain the final answer in the final answer storage unit The flowchart which illustrates the procedure of the maintenance which re-learns of a question answering system, and reviews a basic answer regularly or irregularly. The figure which shows the specific example of the common attribute of the assumption question of each question group about the question groups Q1-QN in an administrative body (for example, city hall or ward office etc.), and the correspondence of the 1st answer corresponding to a question group. The figure which shows the specific example of the assumption question of the question group Q1 of multiple languages, and the relationship of the 1st answer. The figure which shows the specific example of the assumption question of the question group Q2 of multiple languages, and the relationship of the 1st answer. The figure which shows the specific example of the assumption question of the question group QN of multiple languages, and the relationship of the 1st answer. The figure which illustrates the relationship of the primary response identifier, the language identifier, and the final response in multiple languages. The flowchart which illustrates the processing procedure of multiple languages. The figure which illustrates the specific example of the common attribute about question group Q1-QN for demonstrating an entity, and the correspondence with a 1st answer. The figure which shows the relationship between the assumption question of question group Q1, and a 1st answer, in order to demonstrate an entity. The figure which shows the relationship between the assumption question of question group Q2, and a 1st answer, in order to demonstrate an entity. The figure which shows the relationship between the assumption question of question group Q3, and a 1st answer, in order to demonstrate an entity. The figure which shows the relationship of the assumption question and primary answer of question group QN-1 in order to explain an entity. FIG. 11 is a diagram showing a relationship between an assumed question and a first answer of a question group QN for explaining an entity. The figure which illustrates a primary reply identifier and a final reply relationship. Diagram for explaining the correspondence of the primary answers corresponding to the intent (intent) question group

  A question answering system, a question acceptance answering system, a primary answering system, and a question answering method according to the present invention will be described below with reference to the drawings. FIG. 1 is a diagram for schematically explaining a usage pattern and an overall schematic configuration of a question answering system according to a first exemplary embodiment of the present invention. In FIG. 1, the question answering system 10 and the primary answering system 30 constituting the question answering system 25 according to the present invention, and the question transmitted from the questioning terminal 90 are accepted by the question answering system 25 of the present invention, and the question is asked. A part of various devices and a communication network which cooperate with each other until a final answer is returned to the terminal 90 is illustrated.

  A question from the question terminal 90 is transmitted via a question acceptance page on a system (hereinafter referred to as “user system”) 40 which is a platform of a user (trader etc.) connected via a communication network 91. It is transmitted to the question acceptance / answer system 10. When the question terminal 90 is a terminal device such as a desktop PC or a smartphone connected to the user system 40, the question is directly transmitted from the user system 40 to the question acceptance / answer system 10. Here, it is assumed that the communication network 91 includes a wireless communication network, equipment such as a relay station, and a plurality of networks (not shown) (including a LAN of a system provider). 1 and 2, the user system 40 is connected to an external network, but it may be connected to a local area network of a provider who provides a question answering system. .

  Note that, in the description of FIGS. 1 to 14, traders A to N are illustrated as users of the question answering system. However, users of the question answering system of the present invention are not limited to so-called commercial or industrial operators. That is, it is not limited to a for-profit corporation or a for-profit organization, but any organization or individual, including, for example, universities, hospitals, local governments, unions, etc., can be a user (system user) of the question answering system according to the present invention.

  In FIG. 1, 25 is a question answering system according to the first embodiment of the present invention, which includes a question acceptance answering system 10 and a primary answering system 30. The first answer system 30 includes a knowledge database such as artificial intelligence, and the final answer storage unit 18 forming a part of the question acceptance answer system stores the final answer candidates which are the answers to the question. The final answer storage unit 18 can be configured by a storage device such as a server or a database. The final answer storage unit 18 may be an internal storage device of the question acceptance answering system 10, may be connected via a LAN, or may be connected via an external communication network. In FIG. 1, the final answer storage unit 18 is indicated by a broken line in the sense that it may exist as an internal storage device of the question acceptance / answer system 10 or a storage device connected to a network. . FIG. 3 shows a functional block diagram as a configuration example showing the main functions of the question receiving and answering system 10.

The question acceptance / answer system 10 includes hardware such as a server having a CPU, a ROM, a RAM, etc., and basic software and various other types of software. The computer system includes a primary reply inquiry unit for inquiring a primary reply to the primary reply system 30, a final reply unit for returning a final reply to the question terminal 90, and a final reply unit.
The primary answer system 30 includes interface means for receiving a question and returning an answer from the artificial intelligence, and artificial intelligence for outputting the primary answer as an identifier by machine learning the assumed question and the answer. Computer system.

In the conventional question answering system, it is necessary to machine-learn a huge amount of supposed questions and answers in order to improve the satisfaction of the answers. Therefore, much time and labor are required for collecting enormous assumed question data and answer data used for machine learning of artificial intelligence and for machine learning of these data. In addition, machine learning of artificial intelligence was essential to improve and update the accuracy of question answering systems.
On the other hand, in the question answering system 25 of the present invention, the question answering system 10 and the primary answering system 30 having artificial intelligence are clearly separated, and the primary answering system 30 makes the primary answering. And the final answer that is actually returned is separated. As a result, the primary answering system 30 having artificial intelligence and the question acceptance answering system 10 for returning the final answer can be separated and improved separately. That is, the question answering system 25 according to the present invention is capable of improving and evolving the system by machine learning of artificial intelligence, and in addition to the final answer candidate of the question answering system 25 based on the analysis of the question answering log or the like. It is possible to improve and improve the question answering system 25 so that a more appropriate answer can be given to the question by appropriately adding and changing the contents of the question.

  Further, in the present invention, since the system can be constructed on the assumption that the final answer candidate can be easily changed and improved in this way, at the time of introduction, the number of supposed questions to be machine-learned by artificial intelligence and the number of primary answers It is also possible to reduce and compared with the conventional method and to enhance the final answer candidates. As a result, the cost at the time of introduction and the machine learning load at the time of introduction can be reduced.

First, an assumed question to be machine-learned by the primary answering system 30 having artificial intelligence and a primary answer output as an answer to the question from the primary answering system 30 will be described.
For conventional artificial intelligence, machine learning (supervised learning) is performed on a plurality of supposed questions and answers to the supposed questions. To explain using an example of a question answering system in city hall, in general, questions such as "Where should I go to pay taxes?" Or "Which window should I pay taxes?" In response, machine learning is performed so as to output an answer such as "the third window on the second floor is a tax window". However, in the primary answer system 30 of the present invention, machine learning is performed so that a specific identifier is output as the primary answer instead of the specific answer of the supposed question. Here, the form of the identifier may be any unique identifier as long as it can identify the primary response, but a short and simple configuration is desirable as long as it can identify each primary response. In this specification, description will be made assuming that a simple 3-digit numerical identifier such as "100" is output.

  In other words, in the primary response system 30, with respect to a plurality of expected questions having some common attribute (hereinafter, simply referred to as “common attribute”), one same identifier is used as the primary answer for all the expected questions. Machine learning to output. The size and range of common attributes can be set arbitrarily.

  In this specification, a bundle of supposed questions having one common attribute is referred to as a question group. That is, a plurality of supposed questions having a common meaning belong to one question group, and one common primary answer (concept) to the concept (attribute) common to each supposed question included in one question group ( Machine learning is performed to output (identifier). In this case, the number of primary answers (identifiers) is the same as the number of question groups, and the same primary answers (identifiers) are output for each putative question belonging to the same question group. For example, in the example of the assumed question of tax payment described above, when the answer corresponding to the question about “tax” is set to one question group, all the assumed questions regarding tax have one common primary identifier. It will be machine learned so that it can be output as an answer.

What is used as a common attribute can be decided according to the question group. For example, in the example of the assumed question regarding “payment of tax”, “place for paying tax” is a common attribute. Here, it is possible to freely decide whether the “tax refund” and the “tax consultation” are in the same question group or in different question groups. For example, it is possible to broaden the range of common attributes so that all the matters related to taxes can be treated as one answer. In this case, one question group is configured with "tax" as one common attribute, and the primary answers to many assumed questions regarding taxes are all output as one identifier.
When the question is received from the question receiving and answering system 10, the primary answering system 30 returns a predetermined identifier together with the answer certainty to the question accepting and answering system as the primary answer to the question by the artificial intelligence.

  In the question acceptance / answer system 10, the final answer storage unit 18 stores the final answer candidate in association with the identifier (first answer) machine-learned by the first answer system. It is preferable that each final answer candidate includes all answers to a plurality of expected questions that have at least a common attribute included in each question group, and further, only direct answers to the machine-learned expected questions are included. Alternatively, it may be configured to include answers to related questions. This makes it possible to reduce the system introduction period and the introduction cost while maintaining the satisfaction of the answers. Therefore, since the final answer candidate includes many answers such as answers to a plurality of assumed questions and other predicted answers, the final answer candidate in the present invention tends to be long. When the question acceptance answering system 10 receives the primary answer (identifier) from the primary answering system 30, the final answer candidate corresponding to the identifier is read out. 18 is stored.

  The final answer storage unit 18 stores, as final answer candidates, at least a basic answer that serves as a basis for answers to all assumed questions belonging to each question group, associated with an identifier corresponding to the primary answer of each question group. ing. When only the basic answer is stored as the final answer candidate corresponding to each identifier, the basic answer becomes the final answer of the identifier as it is. It is preferable that the contents of the basic answer include answers to all assumed questions related to the question group corresponding to the associated identifier. More preferably, although not included in the expected questions of the question group, it is preferable to include answers to expected questions having the same common attribute as that question group. The final answer storage unit 18 can store a unique answer as a final answer candidate in addition to the basic answer. The unique answer is a final answer candidate that has the same basic framework as the basic answer, but is obtained by replacing the basic answer or adding new information to the basic answer. Therefore, when the question acceptance answering system 10 receives the primary answer (identifier) to the question from the primary answering system 30, it determines from the final answer candidates of the final answer storage unit 18 as the final answer candidates corresponding to the received identifier. , Either the basic answer or the specific answer is read out and returned as the final answer to the question terminal 90.

As will be described later using a specific example, the unique answer is used when the answer content to the question differs depending on the system user.
For example, based on the unique attribute information of the question transmitted from the question terminal 90, the information (identification number, etc.) for identifying the system user is acquired and received from the final answer storage unit 18 from the primary answer system 30. The "basic answer or unique answer" associated with the primary answer (identifier) and the identification number of the system user is read. As a result, even if the primary answers have the same identifier, the unique answer corresponding to the identification number of the system user can be read out and a different final answer can be returned for each system user.

  Here, the “question unique attribute information” is information unique to each question received from the question terminal 90, for example, information regarding the question terminal 90, information other than the question entered by the questioner, and the question terminal 90. Can include various information such as communication information to the question answering system and communication path information, information about the other party to the question (identification number of the system user, etc.), information about the language of the question and answer language. Even if the "question unique attribute information" is automatically acquired from the communication information or the communication route information by the question receiving / answering system 10 or the like, the questioner is requested to input the information as necessary. It may be configured. Further, the identification information of the system user may be configured to add the identification number of the system user in the user system 40 when the question is transmitted via the user system 40.

  As a typical usage pattern in which a unique answer is used, there is a case where it is desired to send back a final answer, which is different from the content of the basic answer, and which adds the business hours and location unique to the system user. For example, there is a case where it is desired to inform a system user who is a real estate agent, from a question terminal, to a question regarding a real estate agent about the business hours and place unique to the system user (real estate agent). In such a case, in addition to the basic answer, a unique answer including the business hours and location of the system user (real estate agent) is stored in the final answer storage unit as the answer corresponding to the first answer. As a result, the unique answer of the system user is read out based on the primary answer and the identification number of the system user, and the business hours and place of the system user are returned to the question terminal 90 as the final answer. Will be.

  Further, as will be described later in detail, the artificial intelligence of the primary answer system includes assumed questions of different languages having a common attribute in the question group to machine-learn the assumed questions and the primary answers. This makes it possible to accept questions in a plurality of types of question languages. The language of the answer to the question does not necessarily have to be the same as the language of the question, and the final answer may be returned in the language standardly set in the question answering system. Since the ability to read documents is relatively high compared to the ability to write documents, there are many people who can read and understand the answers even if they are not in their native language, so it is not necessary to answer in the same language as the question language. Absent. In order to reduce the cost of creating answers and shorten the introduction period, it is useful to make only the question language multilingual. When the introduction cost and the introduction time are a little longer, it is also useful to limit the answer language to Japanese, English, Chinese or the like, which is likely to be used, so that the answer can be provided. It can also be configured to give the final answer in the same language as the question. By associating the final answer in each answer language with each identifier of the primary answer and the answer language information corresponding to the question language and storing the final answer in various answer languages in the final answer storage unit 18, It is possible to send back the final answer in multiple answer languages.

  As is apparent from the above description, the present invention separates the primary answering system 30 and the question receiving and answering system 10 into machine learning of artificial intelligence in the primary answering system 30 and the question answering and answering system 10. The final answer is created and updated separately and independently. In addition, the primary answer system 30 and the question acceptance answer system 10 are connected by the concept of the common attribute of the assumed question and the identifier of the primary answer. Therefore, the expected question to be machine-learned, the primary answer (identifier), and the final answer can be matched, and the expected question and the final answer candidate can be connected to each other. Therefore, it is possible to construct a final answer candidate having a flexible structure with respect to the supposed question to be machine-learned, and regarding the maintenance of the question answering system 25 and the system improvement and updating, the first answering system 30 and the question answering system can be used. The system 25 and the system 25 can be independently operated.

As described above, since the primary answer system 30 and the question acceptance answer system 10 can be separated and the final answer candidates can be flexibly constructed, the system introduction cost and the load are reduced depending on the field of the question answer system to be applied. Various measures and efforts can be made to reduce the number.
For example, by narrowing down to a specialized field or a specific field of a specific industry, the prediction of the question and the required answer become relatively clear. The final answer candidates of the acceptance answering system can include answer contents corresponding to the questions that the artificial intelligence has not machine-learned. This makes it possible to improve the quality of the final answer while suppressing the number of machine learning assumed questions and the number of primary answers. That is, since the load at the time of introduction can be significantly reduced, the introduction cost and the time until the start of operation can be shortened.
Further, the number of supposed questions can be narrowed down by narrowing down the range of the target question to the specialized field, the specific industry, and the question of the specific field. By narrowing down to a specialized field or a specific field of a specific industry, the question can be predicted relatively easily, so that the number of types of common attributes can be reduced.

  Further, basically, the final answer candidate includes all the answers of the supposed question having the common attribute, and further includes the answers related to the supposed question. Therefore, for one question, the final answer including many answers that are frequently asked in the field related to the question is simultaneously provided for one question. Therefore, there is a high possibility that the questioner's next planned answer is also included in the final answer, in which case the second and subsequent questions are unnecessary and the question is suppressed. .

  By separately configuring the primary answering system 30 and the question receiving and answering system 10 in this manner, the assumed questions and the primary answers are reduced to reduce the load of machine learning, and the construction and introduction costs of the question answering system 25 are reduced. Can be lowered. Further, by narrowing down the range of questions that can be accepted, the number of common questions and the number of primary answers (identifiers) to be machine-learned can be significantly narrowed down by narrowing down the number of common attributes while organizing the common questions. it can. For example, in order to increase the rate of correct answers, the number of supposed questions and answers that are conventionally required to be at least tens of thousands or more is narrowed down to a specific task of a specific type of industry by limiting the number of questions to 500. It became possible to start the operation of the question answering system by narrowing down to about 1000 expected questions and the first answer and machine learning them.

In addition, by narrowing down the type of business and business, it is possible to create standardized highly accurate expected questions corresponding to the type of business and business, and to build a question answering system. It is also possible to do. If one question answering system can be used in common by a plurality of peers, the introduction cost can be shared and the system introduction barrier can be further lowered.
Regarding the shared use of the one question answering system 25 by the same trader, as described above, the unique answer is stored in the final answer storage unit 18, and the final answer different for each user is returned based on the unique attribute information of the question. With such a configuration, it is possible to make an individual final response according to the individuality of each user, so that it is possible to realize a more preferable shared use form.
The “common attribute”, “assumed question”, “first answer”, “basic answer”, and “unique answer” described above will be described in more detail later using a specific example.

  In the first answer system 30 that provides the first answer of the present invention, WATSON (WATSON is a registered trademark) of International Business Machines Corporation can be used as a general artificial intelligence service. .

In the above description, as the first embodiment of the question answering system, the configuration example in which the primary answering system 30 is connected to the primary answering system 30 via the communication network 91 has been shown.
However, as shown as an example in FIG. 2, a question answering system 25 ′ according to another embodiment of the present invention does not include the question acceptance answering system 26 and the primary answering system 27 through an external communication network. Alternatively, the connection may be made via an internal line or a dedicated communication line. The configuration of FIG. 2 is different only in that the question receiving and answering system 26 and the primary answering system 27 are connected by an internal line (dedicated line or the like). The basic configuration and function are the same as those of the question acceptance answering system 10 and the primary answering system 30 of FIG. 1, respectively.

  FIG. 1 shows an example in which a management terminal (system management terminal 50) of a system provider who operates and manages the question answering system 25 is connected to the communication network 91. As is clear from the explanation that the communication network 91 also includes a LAN, the system management terminal 50 may be configured to be connected to the question answering system 25 (25 ') by an internal communication system (LAN).

Next, a main configuration example of the question acceptance / answer system 10 (26) will be described with reference to FIG. FIG. 3A is a functional block diagram illustrating the configuration of the question acceptance / answer system 10 (26) of the present invention. For example, CPU, bus control and registers, other control circuits, input / output devices, memories, hard disks, etc. Hardware, software for controlling these hardware, and various programs.
The question acceptance / answer system 10 (26) has, as a basic configuration, a communication interface 11 having a transmission / reception function, an acceptance unit 12 for accepting an external question, a question terminal 90, and an operation history storage unit that stores information regarding the question. 13, a first answer inquiry means 14, and a final answer means 15. Further, it is desirable to provide maintenance means 16 as well.

  The acceptance means 12 of the question acceptance / answer system 10 (26) accepts a question from the question terminal 90 and stores the unique attribute information of the question in, for example, the operation history storage unit 13 or the like. The primary answer inquiry means 14 inquires of the primary answer system 30 (27) about the question received from the accepting means 12, and the primary answer system 30 (27) makes a primary answer and if possible. Will receive the answer certainty factor. When the primary response inquiry means 14 receives the primary response and the response certainty from the primary response system 30 (27), it outputs the information to the final response means 15. The final answering means 15 determines the unique attributes of the primary answer and the question when the primary answer received from the primary answering means satisfies a predetermined condition (for example, when the answer certainty factor is equal to or more than a predetermined threshold value). The final answer candidate (basic answer or unique answer) corresponding to is read from the final answer storage unit 18 (28) and transmitted to the question terminal 90. The operation history storage unit 13 stores an operation history of each process from acceptance of a question to a final answer, for example, a peculiar attribute of the question, a question content, a first answer, an answer certainty degree, a final answer, and the like as necessary. . In the above description and the following description, an example is shown in which the primary answer system 30 (27) always receives the primary answer together with the reliability of the response. It is not mandatory to receive a degree. Further, instead of the answer certainty factor, other similar information may be received.

  In the above, an example of receiving one primary reply (identifier) from the primary reply system 30 (27) is shown, but a list of a plurality of primary reply candidate identifiers from the primary reply system is shown. It is also possible to select an appropriate only primary answer (identifier) in the receiving and answering system 10 (26) for the question and obtain the final answer.

  The maintenance means 16 receives a request for collection, analysis, and maintenance of log information and various operation information from the system administrator or system user, and executes processing such as information collection, analysis, and maintenance. By the maintenance means 16, for example, even a system user who does not have specialized knowledge about artificial intelligence can store the final answer separately from the system administrator of the question answering system 25 (25 ') within a predetermined authority range. Part of the content of the final answer of the part 18 (28) can be independently managed and corrected / changed.

FIG. 3B is a functional block diagram showing an example of the configuration of the final answer means 15 of FIG. 3A.
When the primary reply is received from the primary reply system 30 (27), the primary reply is transmitted from the primary reply inquiry means 14 to the final reply means 15. When the final answer control unit 35 of the final answer means 15 receives the primary answer, it acquires the unique attribute information of the question stored in the operation history storage unit 13 via the unique attribute information acquisition unit 36, and finally It is transmitted to the answer control unit 35. Based on the identifier of the primary response and the unique attribute information (system user information, etc.), the final response control unit 35 sends the primary response from the final response storage unit 18 (28) to the final response acquisition unit 37. Command to get the basic or specific response corresponding to. The basic answer or the specific answer acquired by the final answer acquisition unit 37 is sent to the final answer transmission unit 38, and is transmitted to the question terminal 90 via the communication interface 11 as the final answer under the control of the final answer control unit 35. It

FIG. 4 is a functional block diagram showing a configuration example of the maintenance means 16 shown in FIG. The maintenance unit 16 includes an initial setting unit 21, an analysis unit 22, and a data updating unit 23. An operation from the system management terminal 50 is accepted via the initial setting means 21, a plurality of expected questions and primary answers (identifiers) are input and set when the system is constructed, and machine learning of the primary answer system 30 (27) is performed. Work such as data setting is performed, and final answer candidates (basic answer and unique answer if necessary) are stored in the final answer storage unit 18 (28).
The analysis unit 22 includes statistical information such as questions stored in the operation history storage unit 13, primary answers, final answers, and answer certainty factors, for example, questions and answer numbers for each system user, and questions for each answer group. Various statistical data such as the number of responses, the average response certainty, the number and ratio of the response certainty below the reference value are created. These statistical data can be monitored only by a person having a predetermined authority.

The data updating unit 23 receives access from the management terminal of the user system 40 and the system management terminal 50, and updates desired data.
From the system management terminal 50, it is possible to update the knowledge data of the primary answer system 30 (27) by machine learning and to update the content of the final answer candidate in the final answer storage unit 18 (28). it can. For example, when the degree of certainty of the first answer to the question is low, a new assumed question or a first answer (identifier) is added to cause the artificial intelligence machine of the first answer system 30 (27) to perform machine learning. be able to. Further, it is possible to create a final answer candidate corresponding to a new identifier in the final answer storage unit of the question acceptance answering system, and to add / correct / change an already existing basic answer or unique answer. The initial setting means 21 can also integrate its function into the data updating means 23.

The system user can read the access log information of the system user from the management terminal of the user system (40) and browse the operation history and analysis data, and within the predetermined permitted range, the final answer storage unit 18 The final answer data stored in (28) can be corrected or changed.
This allows the system user (trader) to change, modify, add, etc., the content of the final answer in the final answer storage unit 18 (28) by himself / herself, regardless of whether or not he / she has specialized knowledge about artificial intelligence. That is, the system user (trader, etc.) makes a final response using the management terminal of the user system 40 connected via the communication network within a predetermined permitted range without requesting the system provider. The final answer (basic answer and unique answer) stored in the storage unit 18 can be corrected and updated.

With reference to FIG. 5, a basic flow of answering a question from the question terminal 90 in the question answering system 25 (25 ′) according to the embodiment of the present invention will be described. In the following description, an example of a processing procedure for providing a unique answer by using the unique attribute information of a question will be described. You can choose whether or not to provide an answer.
When the question q1 is received from the question terminal 90 (step S1), the accepting unit 12 stores the peculiar attribute information of the question, the question content, and the like in the operation history storage unit 13 (may be another storage unit) (step S1). S2), the primary response inquiry means 14 inquires the primary response of the question q1 to the primary response system 30 (27) (step S3).

  The primary answer system 30 (27) extracts the primary answer (final answer identifier: a1 to an in FIG. 5) to the question q1 and sends it back to the primary answer inquiry means 14. The primary response system 30 (27), for example, one or more primary responses (identifiers) in descending order of the response certainty, and the response certainty corresponding to the primary response, the primary response inquiry means. Send to 14. When the primary response inquiry means 14 receives the primary response from the primary response system 30 (27) (step S4), it is sent to the final response means 15. In the final answering means 15, whether or not the primary answer (identifier) satisfies a predetermined condition for replying to the questioner, for example, the first answer with the highest degree of certainty in the received primary answers. It is verified whether or not the answer certainty factor of the next answer exceeds a predetermined threshold value (step S5).

When the primary answer satisfies the predetermined condition (step S5; Yes), the final answer means 15 acquires the unique attribute information of the question to obtain the identification information of the system user and the like (step S6). . The peculiar attribute information of the question is stored in the operation history storage unit 13 or the like, and by confirming the communication path (origin) of the question from the peculiar attribute information, for example, which system user (provider) the question passes through It is possible to identify the system user (trader) by checking whether it has come. Next, based on the acquired primary response and the information (identification information) of the system user (trader), the basic answer candidate corresponding to the primary answer or the unique answer candidate corresponding to the system user (trader) is selected. It is acquired (step S7). When there are a plurality of received primary answers, the final answer candidates (basic answer candidate or unique answer candidate) to be returned as the final answer are narrowed down. For example, the basic answer candidate or the unique answer candidate corresponding to the highest answer certainty factor of the received first answer (identifier) is acquired from the final answer storage unit 18 (28) (step S7).
After that, the final answer candidates thus obtained are transmitted from the final answer means 15 to the question terminal 90 having a question (step S9).

  When the primary answer does not satisfy the predetermined condition in step S5 (step S5; No), for example, when the answer certainty factor is equal to or less than the predetermined threshold value, there is no suitable final answer candidate for the question q1. Is obtained (step S8), and transmitted from the final answer means 15 to the question terminal 90 (step S9). . Here, in step S5, the final answer when the primary answer does not satisfy the predetermined condition can be stored in the final answer storage unit 18 (28) in association with the special identification number. However, it can be embedded as data in the program. In addition, in FIG. 5, the processing steps of step S5 and step S6 may be interchanged in order.

  When the final answer is transmitted to the question terminal 90, various information including a log regarding the question and the final answer and other log data is stored in the operation history storage unit 13 (step S10). When a plurality of primary answers are presented from the primary answer system 30 (27) in descending order of the answer certainty, all the answer certainties or up to a predetermined rank are stored in the operation history storage unit 13. It is desirable to store the answer certainty factor.

As is clear from the above description, from the primary answer system 30 (27), the primary answer (identifier) corresponding to the question group to which the input question approximates the input question belongs If so, the answer certainty factor is output.
By the way, the questioner thinks that he has asked the system user (one of the vendors A to N) he has accessed. Therefore, depending on the content of the question, it is necessary to return different answers depending on the user systems 40 of the operators A to N who relayed the question of the question terminal 90. For example, in the case of a question about inquiring about business hours, it is necessary to answer the correct business hours for each business operator who has passed the question. In such a case, a unique answer candidate of the system user is prepared, the identification number of the system user is acquired based on the unique attribute information of the question stored in the operation history storage unit 13 or the like, By acquiring the unique answer corresponding to the system user (traders A to N) via the question, the final answer notifying the business hours of each system user is returned.

With reference to FIG. 6 to FIG. 11, as a final answer candidate, an expected question to be machine-learned by the primary answering system 30 (27), a primary answer associated with a question group, and a final answer storage unit 18 (28). The stored basic answer and unique answer will be described using a specific example.
In the examples of FIGS. 6 to 11, the trader is a system user who wants to use the question answering system, and is a trader who performs business related to real estate transactions. In the example of FIGS. 6 to 11, a plurality of contractors (contractors A to N) have contracted to use the question answering system 25 (25 ′), and the respective question pages of the user systems 40 of the respective contractors A to N. The question answering system 25 (25 ') can be accessed through the. 6 to 9 show specific examples of question groups, and FIGS. 10 and 11 show examples of basic answer candidates and unique answer candidates stored in the final answer storage unit 18 (28). .

  The question answering system 25 (25 ') is related to the question based on the information indicating which user system 40 each question came through (communication information: part of the unique attribute information of the question). A system user (trader) is specified and a unique answer candidate peculiar to the system user is acquired. When the system user (trader) does not request unique additional information, basic answer candidates are acquired and transmitted to the question terminal 90 as the final answer. Although not shown in the examples of FIGS. 6 to 11, it is also possible to adopt a simple configuration in which only the basic answer is the final answer for all questions, without allowing the unique answer for each system user (trader). Is.

First, with reference to FIGS. 6 to 9, assumed questions and primary answers to be learned by the primary answer system 30 (27) will be described using specific examples. In the following description, a configuration example is shown in which one question group includes a plurality of supposed questions.
FIG. 6 is a diagram illustrating the content of the common attribute of each question group in the question groups Q1 to QN and the relationship of the primary answers corresponding to each question group Q1 to QN in a 1: 1 ratio. In the following figures, question groups are appropriately displayed as "question groups". In FIG. 6, the primary answers are identifiers 101 to NNN. It is sufficient for this identifier to specify the primary answer, and the expression format is arbitrary. Each question group Q1 to QN has a plurality of assumed questions that are grouped under the concept of a common attribute for each group, and a plurality of supposed questions included in each of the question groups Q1 to QN in FIGS. Some of the questions are given as examples.

In FIG. 6, the question group Q1 is a group having a common attribute of “real estate agent”, and it is illustrated that the primary answer is the identifier “101”. It is illustrated that the question group Q2 has a common attribute of “indication of mortgage loan”, and the identifier of the primary answer is “102”. Similarly, it is illustrated that the question group QN has a common attribute of “points when collecting regional information” and the identifier of the primary answer is “NNN”.
This identifier is the same as the identifier used when creating the final answer in the question acceptance / answer system 10 or 26 described later, and is associated with the final answer candidate described later with reference to FIGS. 10 and 11. Is.

  7 to 9 show specific examples of a plurality of assumed questions of the question groups Q1, Q2, and QN. FIG. 7 specifically illustrates a plurality of assumed questions Q1 (1) to Q1 (n) corresponding to “real estate agent”, which is a common attribute of the question group Q1. Further, FIG. 8 shows a specific example of a plurality of assumed questions Q2 (1) to Q2 (9) corresponding to "a guideline for borrowing a mortgage" which is a common attribute of the question group Q2. Similarly, FIG. 9 shows a specific example of a plurality of assumed questions QN (1) to QN (12) corresponding to "points in collecting area information", which is a common attribute of the question group QN.

  In FIG. 7, all the identifiers "101" are designated as the primary answers to the n possible questions of the question group Q1. That is, all the primary answers to the n supposed questions in the question group Q1 having “real estate agent” as a common attribute are one identifier 101. Correspondingly, the basic answer candidate corresponding to the identifier 101 of the primary answer is stored in the final answer storage unit 18 (28), and the basic answer candidate corresponding to the identifier 101 is one of the basic answer candidates. Is a bundle of answers including all the answers corresponding to the common attribute in the plurality of assumed questions Q1 (1) to Q1 (n) in the question group. A specific example of the final answer candidate stored in the final answer storage unit 18 (28) will be described later with reference to FIGS. 10 and 11.

  Similarly, FIG. 8 specifically illustrates a plurality of supposed questions Q2 (1) to Q2 (9) corresponding to the question group Q2, and the plurality of supposed questions Q2 (1) to Q2 (9). The same single identifier “102” is designated as the first answer to the above. Similarly, FIG. 9 specifically exemplifies twelve assumed questions QN (1) to QN (12) related to the question group QN regarding the common attribute “points in collecting information on the area”. The same identifier "NNN" is designated as the primary answers corresponding to these 12 assumed questions.

10 and 11 exemplify specific examples of basic answer candidates and unique answer candidates stored in the final answer storage unit 18 (28). The display in the column of the primary response (identifier) shown in FIGS. 10 and 11 corresponds to the primary responses from the primary response system 30 (27) to the identifiers 101 to NNN shown in FIGS. ing. In the final answer storage unit 18 (28), the "basic answer candidate" and the system user (corresponding to each of the identifiers 101 to NNN of the primary answer received from the primary answer system 30 (27) are associated with each other. When the traders A to N) desire, the "unique answer candidate" corresponding to the desired system user is stored.
In the user (provider) identifier column of FIGS. 10 and 11, "supplier A", "supplier B", etc. are displayed as the system user identifier, but an identifier that can identify the system user (supplier) If so, you can freely use numbers, letters, and symbols. 10 and 11, the information that identifies the user is the identifier of the system user, but this may be any identifier that identifies each user, and is "identification information of the user system". Is also good.

The system user (trader) identifier is a unique identifier assigned to each of the traders A to N, and a unique answer candidate for each trader is associated with the identifier of this trader and stored in the final answer storage unit 18 (28). Remembered 10 and 11 exemplify the contents of the basic answer candidates assigned as the final answer. Referring to the primary response identifiers “102” and “NNN” in FIGS. 10 and 11, there is no unique response candidate corresponding to the identifier of the trader B in the primary response identifier “102” column. There is no unique answer candidate corresponding to the identifiers of the traders A and C in the field of the primary reply identifier “NNN”. As described above, the setting of the unique answer candidate of the system user (trader) is not essential, and can be set arbitrarily according to the request of the system user (trader). In addition, all the primary answers or the specific primary answers may be configured to have all the unique answer candidates and only the basic answer candidates.
The identifier of the system user (trader) is determined by the identifier of the system user (trader) in each question from the unique attribute of the question for each question, and the basic answer candidate corresponding to each question or the corresponding unique answer candidate Is read from the final answer storage unit 18 (28).

Specific answer contents of the basic answer candidates are illustrated in FIGS. 10 and 11. As can be seen from the contents of the basic answer candidates, the basic answer candidate and the specific answer candidate, which are the final answers, are relatively long answers including all the answers related to the multiple expected questions of the question group having the common attribute. . This is based on the fact that the final answer is a bundle of answers including all answers related to a plurality of supposed questions included in each question group. With this configuration, even if the number of assumed questions and primary answers to be machine-learned by the primary answer system is reduced, it is possible to maintain the questioner's satisfaction with the final answer to the question at a certain level or higher. Become. Since the assumed questions are carefully selected questions that are likely to be asked, the basic answer including all answers to the assumed questions contains a lot of useful information for the questioner. Therefore, it is possible that the basic answer to one question may include the answer to the question that the questioner was going to ask next, in which case the questioner's satisfaction may be further increased. it can.
In this way, it is possible to reduce learning data creation and machine learning time for artificial intelligence to 1/10 or 1/100 or less by giving a wide range of meanings to a plurality of assumed questions, and artificial intelligence was used. It becomes possible to efficiently provide a question answering service at a very low cost.

Further, from the first answer system 30 (27), regardless of the unique attributes of the question such as which user system the question went through, the same primary answer is always given to the supposed question belonging to the same question group. (Identifier) is output. On the other hand, in the question receiving and answering system 10 (26), the basic answer candidate stored in association with the primary answer is changed and the unique answer candidate unique to the system user is read out as necessary. , The read unique answer is transmitted to the question terminal 90 as the final answer. Therefore, the unique answer candidate is stored in association with the identifier of the primary answer and the identification information of the system user (identifier of the system user). The unique answer candidate is obtained by replacing the basic answer candidate corresponding to the identifier of the first answer in accordance with the system user.
As described above, since the final answer candidates (basic answer candidates and unique answer candidates) stored in the final answer storage unit 18 (28) are stored data separated from artificial intelligence, they are not machine learning for artificial intelligence. It is possible to appropriately change or add the content regardless of. Therefore, maintenance work such as changing or adding the final answer candidates stored in the final answer storage unit 18 (28) does not require a high degree of specialized knowledge about artificial intelligence. However, the basic answer candidate corresponds to the first answer of the first answer system, and it is the base of the final answer, and it is common to all system users, so the system user should not change it. It is preferable to limit.

  As is apparent from the above description, the question answering system 25 (25 ') according to the present invention has been accumulated through experience and achievements in order to efficiently answer the question from the end user or potential end user. Data types such as assumed questions and final answer candidates, as well as facilities for automatically providing appropriate answers in response to questions, such as advanced computer systems such as artificial intelligence systems and databases, etc. It becomes possible for a plurality of system users (traders, etc.) to jointly use it. By enabling the construction of such a question answering system that enables shared use, not only can the construction cost when the question answering system is introduced for the first time be greatly suppressed, but also the system introduction period can be shortened. In addition, when a company that shares the same data or system installs it later, the data and system environment has already been constructed, so an excellent question answering system can be installed with confidence in an extremely short period of time. Has the effect. Also, since many users share the question answering system, it is possible to receive a wide variety of questions that cannot be obtained by one company in a short period of time. Therefore, it becomes possible to grow a better system in a short period of time.

  The question acceptance / answer system 10 (26) according to the present invention includes the maintenance means 16 as described above, and the initial setting means 21 causes the primary answer system 30 (27) and the final answer storage unit 18 (28) to be stored. Perform initial settings. The initial setting procedure of the primary response system 30 (27) and the final response storage unit 18 (28) will be described with reference to FIG.

FIG. 12 is a flowchart exemplifying a basic work procedure for initial setting of the primary answer system 30 (27) and the final answer storage unit 18 (28). In order to construct a system, first, an assumed question, a common attribute of the assumed question, a question group, a first answer, a final answer (basic answer, unique answer), and a unique attribute of the question are determined (step S21).
For that purpose, the system user (industry) 's work is analyzed, a plurality of expected questions are created, and answers to all the expected questions are created. Assumed questions are a plurality of assumed questions that have a common attribute and are grouped into one question group, and the primary answer is an identifier corresponding to the question group 1: 1. Of the final answer candidates, the basic answer candidates shall include all the answers related to a plurality of assumed questions in one question group. Further, the unique attribute information is determined, and the range and the content of the unique answer candidate associated with the unique attribute information are determined. At that time, language information can be included in the unique attribute information. In such a case, different final answer candidates are prepared depending on the expected questions in various languages. Such final answer candidates are created for all question groups for each question group.

  Next, the primary answering system 30 (27) is made to machine-learn all the expected questions included in each question group Q1 to QN and the primary answer (identifier) corresponding to each question group 1: 1. (Step S22). Further, the identifier of the primary answer and the basic answer candidate are stored in the final answer storage unit 18 (28) in association with each other (step S23). As described above, in step S23, not only the basic answer candidate for one identifier corresponding to the primary answer, but also the unique answer candidate for each system user based on the common basic answer candidate You may comprise so that it may be linked | related and memorize | stored in a primary response and the identifier of a system user. The order of step S22 and step S23 may be interchanged or performed simultaneously.

  The question answering system 25 (25 ') according to the present invention includes the maintenance means 16 as described above. The maintenance means 16 regularly (or irregularly) checks the history of questions / answers, creates and analyzes statistical information, and continuously monitors the effect of use of the system so that basic answer candidates and It is possible to correct / add unique answer candidates.

The question answering system 25 (25 ′) of the present invention maintains the machine learning of the primary answering system 30 (27) and the final answer candidate stored in the final answer storage unit 18 (28) from the result of monitoring. This allows us to grow continuously. Furthermore, regularly checking questions with low confidence of answers leads to the opportunity for each user to know the "needs of existing customers and potential customers" that he / she is not aware of and simply answers the questions. Not only the original purpose of the system 25 (25 '), but it can be a means of collecting customer and market information necessary for business growth. In addition, as explained below, basic answer candidates and unique answer candidates can be maintained by system users even under constant conditions, even for system users who do not have knowledge of artificial intelligence. System users themselves can continuously improve and grow the question answering system.
Further, since the question answering system can be used jointly by a plurality of system users, it is possible to receive far more various questions than one company can use. Therefore, during maintenance of the entire question answering system 25 (25 ′) by the system provider, assumed questions and basic answer candidates are modified and improved to cope with various problems, and the growth rate of the question answering system becomes faster. It is expected. For the growth of this question-answering system, all collaborators can benefit from this.

  With reference to FIGS. 13 and 14, a procedure for maintenance of the primary reply system 30 (27) and the final reply storage unit 18 (28) by the system user (trader) or system administrator will be described. First, the procedure for modifying / changing the final answer storage unit 18 (28) by the system user (trader) will be described.

FIG. 13 is a flowchart showing a processing procedure for the system user (trader) to correct / change the final answer unique to the company.
The system user (trader) uses the management terminal of the system user who is connected to the question answering system 25 (25 ') through any communication network to maintain the question answering system 25 (25'). 16 (see FIG. 3 and FIG. 4), and from the record of the access log or the like for which there is a question to the system user, the distribution of the answer confidence of the question and the first answer, the average response confidence, etc. It is possible to check the status of other final answers and consider modifying, changing, or adding the contents of the final answer candidates (basic and specific answers). If there is a case where an appropriate final answer could not be made to the question from the question terminal, or there was a case where the final answer was made but the reliability of the primary answer was low, the system user (provider) Within the scope of the authority given to himself or by requesting the system administrator, the final answer candidate stored in the final answer storage section can be corrected or added.

  As shown in FIG. 13, the system user (trader) regularly or irregularly checks and analyzes the access log (step S31). In that case, when it is desired to correct or change the final answer (basic answer and unique answer) stored in the final answer storage unit 18 (28) (step S32; Yes), the final answer is modified / changed next. It is confirmed whether or not there is an authority to do so (step S33). Although not shown in FIG. 13, when the system user (trader) browses the operation history information and analyzes the log and the like (step S31), does the person in charge have access authority in advance? It is advisable to confirm. If the system user (trader) does not have the authority to make corrections / changes, the corrections / changes are not accepted and the maintenance process ends (step S33: No). If the right to modify / change is authorized (step S33: Yes), the final answer candidate is modified within the authority of the system user (trader) stored in the final answer storage unit 18 (28) ( (Including changes or additions) (step S34).

  For example, if the question terminal 90 asks "I want to know about business hours" as the question q1, and if the system user's unique answer has already been set, information specific to the vendor, such as temporary closure, will be used. Can be added to the unique answer candidate. In addition, when the system user (trader) desires a correction exceeding the authority of the system user based on the result of analyzing the log data and the like, for example, the change of the basic answer candidate, the first answer system 30 (27) If you wish to add an expected question to an existing question group, add a question group with a new common attribute, or modify or add basic answer candidates, ask the system administrator of the question answering system for consideration. To do.

FIG. 14 is a flowchart showing an example of a maintenance processing procedure of the entire question answering system 25 (25 ′) by the system administrator. This maintenance processing procedure is a processing procedure performed by the system administrator who manages the operation of the entire question answering system via the system management terminal 50.
When maintenance is started (step S41), it is first confirmed whether or not there is access authority to this operation history (step S42), and if there is no access authority, this processing flow ends (step S42; No). When the user has the access authority (step S42; Yes), the access status of the entire system, the question and answer log data stored in the operation history storage unit 13, and the analysis data such as the statistical data are confirmed and analyzed. (Step S43).

When the final answer candidate in the final answer storage unit 18 (28) is modified / changed (including addition, the same applies hereinafter) (step S44; Yes), it is confirmed whether or not the user has the right to modify / change the final answer. (Step S45). If there is a right to modify / change, etc. (step S45: Yes), the basic reply / specific reply is changed, or a new unique reply of the vendor is added (step S46). If there is no right to modify / change, the process is terminated (step S45: No). Of the final answer candidates, the system administrator is only allowed to correct / change basic answer candidates, the system administrator is not allowed to correct / change unique answer candidates, and the system is used for unique answer candidates. It may be configured such that only the person having the authority to make corrections and changes.
This figure shows an example in which the system administrator has the final answer candidate and the machine learning authority separately, and both authorities may be covered by one authority.

  When the final answer candidate is not corrected or changed (step S44; No), or when the final answer candidate is corrected or changed in the final answer storage unit 18 (28) (step S46), then the artificial intelligence system is used. It is confirmed whether or not machine learning is performed by adding an expected question to a certain first answer system 30 (27) (step S47). If there is no need for additional machine learning (step S47; No), this maintenance process is terminated. When additional machine learning is performed (step S47; Yes), it is confirmed whether the primary answer system 30 (27) has the authority to correct or change the expected question or the primary answer (identifier). (Step S48). If the user has the right to correct or change (step S48; Yes), the desired answer, the desired answer, the desired answer are added, modified and deleted, and the primary answer system is subjected to machine learning (step S49). Ends the maintenance process. If the user does not have the correction authority (step S47; No), the maintenance process is terminated without performing the machine learning.

  In the maintenance process (FIG. 13) by the system user (trader) and the maintenance process (FIG. 14) by the system administrator described above, for example, the data is updated or changed based on the following rules. As a result, the question answering system 25 (25 ') can be evolved in sequence to surely grow a system having a high answer certainty step by step.

<Example of response rule when answer confidence is low>
(1) If the question content is completely unknown and there is no corresponding final answer, a pair of question and answer identifier is added to the first answer system 30 (27) and The final answer is also added to the answer storage unit 18 (28). This maintenance is performed by the system administrator.
(2) If the content of the question is unknown, but an existing final answer is appropriate, the pair of the question and the identifier of the existing final answer is sent to the first answer system 30 (27). To add. In this case also, the system administrator will carry out.

<Example of response rule when answer confidence exceeds the standard value, but the content of the answer is not valid>
(1) If the content of the question is known, but the content of the final answer is not appropriate, the system administrator modifies only the data of the final answer in the final answer storage unit 18 (28).
(2) When the basic answer, which is a common answer, needs to be corrected commonly to all the related businesses, the system administrator can correct the basic answer in the final answer storage unit 18 (28). .
(3) If the basic answer, which is a common answer, does not match the contents of some system users (traders), the system administrator of the system users (traders) will give the final answer corresponding to the company identification number. You can copy the basic answer to the area and modify the contents of the basic answer. At this time, when the final response only for the company that does not affect the common response is modified, it can be added or modified for each trader identifier.

<Correspondence rule when answer accuracy is supposed to be improved by dividing question group>
(1) If there is an answer identifier with a large number of questions / answers, and if it is expected that user satisfaction will be further improved by narrowing the range of answers and dividing the question group and final answer, Divide the question group and final answer that are currently registered as one. This maintenance process is performed by the system administrator.
(2) Even when the length of the answer to the question becomes too long, it is desirable to divide the question group and the final answer as in the case where the number of questions is large. This maintenance process is performed by the system administrator.

<Support for various languages>
In the embodiments and specific examples described above, the question answering system of the present invention has been described by exemplifying the supposed question and the final answer using Japanese as the question language and the answer language.
In the question answering system 25 (25 ') according to the present invention, it is possible to deal with questions in various kinds of languages by grouping assumed questions and using unique attribute information of the questions. In the following, an example of a configuration of the question answering system 25 (25 ′) according to the present invention that responds to questions and answers in a plurality of types of languages will be described by showing specific examples of assumed questions and final answer candidates.

  If the native language of the questioner and the respondent are different, smooth communication is difficult unless either the questioner or the respondent is familiar with the native language of the other. In such a situation, if there is a question answering system that can accept the question in the native language of the questioner, the questioner can create a clear question sentence in a correct document. Since the quality of the question is improved by asking the question in the native language, the probability that an appropriate answer is provided from the question answering system is significantly improved.

  Therefore, in the question answering system, a mechanism for accepting questions in various languages is desired. Note that, in general, the ability to read a document is relatively higher than the ability to create a document, even if the document is not in the native language. Therefore, the answer to the question is not necessarily in the native language of the questioner. For example, from the viewpoint of cost-effectiveness, the types of question languages are widely accepted, but the answer languages are narrowed down to the major languages in the country where the question answering system is installed, and several widely used languages such as English and Chinese in the world. It is also possible.

Such a multilingual question answering system can significantly reduce the human burden in various consultation guidance services of administrative agencies, universities, hospitals, and the like. For example, in a local government, you can answer questions in various languages about simple questions such as where to find a person who has newly moved, such as a window for relocation and acquisition, a resident's card providing window, a tax payment window, etc. Even if it is possible to respond automatically through the system, it is possible to greatly reduce the individual handling effort by humans.
In particular, at the initial consultation stage and in the initial stage of consultation, many questions are often formal or simple basic questions, so a question answering system that accepts questions in other languages is not available to foreign questioners. It will be a very useful system for both users and institutions.
In addition, the ability of a questioner to ask a question in advance from a remote location using his or her own smartphone, etc. in such a question answering system makes it more convenient for the questioner and the user organization to ask questions than before. It will be a very useful system for both parties.

In the question answering system 25 (25 ') according to the present invention, the question answering system 25 (25') capable of dealing with various kinds of languages is constructed only by adding a simple configuration to the basic configuration of the present invention. be able to.
In the question answering system 25 (25 ') according to the present invention, if the assumed questions have the same common attribute, they are configured to belong to the same question group even if the questions have different languages. Responding to questions and answers in multiple languages by recognizing them as part of the unique attributes of the question.
Therefore, regarding a plurality of supposed questions having different kinds of languages but having the same meaning content, the primary answer system 30 (27) is caused to machine-learn that the common attributes belong to the same question group. At the same time, the final answer storage unit 18 (28) stores final answer candidates (basic answer candidates or unique answer candidates) for each type of answer language. The answers in the default language set in the question answering system 25 (25 ') are registered as basic answer candidates, and the answers in other languages are registered as unique answer candidates.

When responding to a question in a large number of languages, it is necessary to extract a correct primary answer according to the language of the question and answer in an appropriate language corresponding to the questioned language. In order to determine the question language and answer language, the questioner can be asked to input the desired question language and answer language. As the simplest method, when the question answering system 25 (25 ') is accessed from the question terminal 90, a language selection button for designating the question language and the type of answer language is provided from the question input screen of the question terminal. You can In this case, a desired language type may be selected by a pull-down method, a radio button method, or the like so that only the type of language that the question answering system can support can be specified. desirable.
When the question acceptance answer system 10 (26) or the primary answer system 30 (27) has a function of inferring the type of question language from the form of the question sentence (bit configuration of character code, etc.) It is also possible to judge the language type from the question sentence by the question acceptance answering system 10 (26) or the primary answering system 30 (27). Further, the type of question language can be obtained from common information necessary for communication between the question terminal and the question answering system. Therefore, it is possible to assume the question language without input by the questioner. Regarding the answer language, it can be dealt with by making an initial setting so as to be determined by a predetermined rule according to the question language.

  The desired question language and type of answer language are sent to the question acceptance / answer system 10 (26) together with the question, and are stored in, for example, the operation history storage unit 13 as part of the unique attribute of the question. Here, it is possible to improve the convenience of the user by automatically configuring the answer language to be the same as the input language. However, in view of the fact that the document reading ability is much higher than the document creating ability, from the viewpoint of reducing the load at the time of introducing the system, and from the viewpoint of reducing the response load of the question answering system 25 (25 '), etc. The type of answer language may be English or Japanese only, and the answer may be made in a smaller number of languages than the type of question language.

As described above, first, it is necessary to accept a question in a plurality of types of languages from the primary answer system 30 (27) and obtain an appropriate primary answer regardless of the language of the question. Therefore, multiple expected questions having the same common attribute are created according to the language type, and the expected questions of multiple languages having one common attribute correspond to each question group as 1: 1 as one question group. The primary answering system is machine-learned so that the primary answer can be obtained.
For example, each of the question groups Q1 to QN, which are assumed questions in a plurality of languages such as a Japanese question, an English question, and a Chinese question having common attributes, are assumed to belong to the same question group. The hypothetical question for each language having the same common attribute here may be a hypothetical question created in one language and translated or translated into another language as a hypothetical question of the same meaning content, but if not the same, It may be an assumed question that has almost the same meaning in consideration of language and culture. Next, the question groups Q1 to QN including the expected questions in various languages are machine-learned so that the primary answers corresponding to the question groups Q1 to QN are obtained 1: 1. As a result, the same primary answer (identifier) is obtained from the primary answer system 30 (27) for the supposed questions belonging to the same question group, regardless of the type of question language.

  In the question acceptance / answer system 10 (26), final answer candidates in various languages are stored in the final answer storage unit 18 in a one-to-one correspondence with the primary answers in accordance with each answer language. Therefore, when the primary answer is received from the primary answer system 30 (27), the final answer candidate in an appropriate answer language is based on the identifier of the primary answer and the language information (identifier) of the question. Can be read and the final answer can be returned to the question terminal 90.

  A combination with a unique answer candidate other than the language such as the trader identifier described above is also possible. By using basic answer candidates corresponding to the first answer and unique answer candidates that differ for each system user (trader) according to the type of language, the system can be used in an appropriate language according to the question language. It is possible to read out the unique answer candidate for each person and use it as the final answer. Whether to prepare a unique answer candidate corresponding to a combination of a plurality of kinds of unique attributes as the final answer can be appropriately set as necessary. When the question answering system is not used by a plurality of institutions, such as when it is used independently by an administrative agency or a university, the unique answer candidate is not always necessary.

  When the question acceptance / answer system 10 (26) receives a question from the question terminal 90, the acceptance means 12 acquires the language information of the question from the unique attribute information of the question and stores it in, for example, the operation history storage unit 13. The storage location is not limited to the operation history storage unit 13, and a location for storing individual attribute information of a question may be separately provided. When the question is received, the primary answer inquiry means 14 requests the primary answer system 30 (27) for the primary answer corresponding to the question. In this case, when the primary answering system 30 (27) requests the language information of the question, the question acceptance answering system 10 (26) sends the question sentence and its language information, and does not request the language information of the question. In some cases, only the question text will be sent. Upon receiving the primary answer from the primary answer system 30 (27), the question acceptance answer system 10 (26) acquires the answer language identifier based on the language information stored in the operation history storage unit 13. . After that, the final answer candidate in an appropriate answer language is acquired based on the identifier of the first answer and the answer language information (identifier).

With reference to FIG. 15 to FIG. 19, a concrete example assuming a question and an answer in an administrative institution regarding a question group, an assumed question, and a final answer when constructing a question answering system capable of answering a question in each kind of language. Will be explained.
15 to 19 show question groups Q1 to QN, assumed questions Q1 (1) to QN (8) in Japanese, English, and Chinese in each question group, and final answer candidates in Japanese, English, and Chinese. Specific examples of (basic answer candidate and unique answer candidate) are simplified and shown.

FIG. 15 assumes that the question answering system according to the present invention is introduced in the city hall, and in the case of accepting questions in other types of languages, regarding question groups Q1 to QN, the common attributes of the expected questions of each question group and the question group. Shows a specific example of the first answer corresponding to. In the example of FIG. 15, the common attribute of the question group Q1 is “transfer procedure”, the common attribute of the question group Q2 is “acquisition of residence certificate”, the common attribute of the question group QN is “payment of tax”, and the first answer Are "101" to "NNN", respectively. 16 to 18 show assumed questions and primary answers (identifiers) included in the question groups Q1, Q2, and QN, respectively, and FIG. 19 shows specific examples of and final answer candidates.
Note that, in FIG. 19, only the final answer for each language is shown for simplification of the description, but it is possible to provide a unique answer other than the language for each language type.

  FIG. 16 shows an example in which the expected number of questions in Japanese is 4, English is 3, and Chinese is 2. The number of supposed questions for each language may be the same, but the number of supposed questions for each language does not necessarily have to be the same as shown in FIG. In the initial stage of introducing a question answering system, it is better to answer questions in English or Chinese by giving priority to being able to answer basic questions, rather than trying to prepare many expected questions. , In many cases, this will not only improve the convenience of the inquirer but also the user organization. This is because foreigners with different cultures often do not understand the basics, and it is anticipated that it will be necessary to deal with foreign inquirers. Further, the meaning of the question sentence may be a question sentence having the same content through each language, but may be a question sentence having a different meaning or nuance for each language according to customs, culture and the like.

  In the question group Q2 shown in FIG. 17, “acquisition of resident's card” is selected as the common attribute, and the primary answer of the question group Q2 is “102”. As a hypothetical question belonging to the question group Q2, FIG. 17 shows a specific example of the hypothetical question in a plurality of languages. In the question group QN, “payment of tax” is selected as a common attribute, and the primary answer of the question group QN is “NNN”. In addition, FIG. 18 shows a specific example of assumed questions in a plurality of languages as assumed questions of the question group QN.

16, 17, and 18 exemplify concrete question sentences in Japanese and English, and Chinese is schematically shown using Japanese.
As can be seen from the figure, for example, the first answer to the question in the question group Q1 is only one of the identifiers "101" regardless of the type of language, as shown in FIG. . That is, although the question sentence includes assumed questions in Japanese, English, and Chinese, the primary answer to the question group Q1 is only the identifier “101” regardless of the assumed question in any language. This is the same in the question groups Q2 and QN as shown in FIGS. 17 and 18, and the primary answer of the question group Q2 is “102” regardless of the language of the assumed question. The answer of the group QN is "NNN".
In the first-answer system, all the supposed questions and the first-answers belonging to each of these question groups are machine-learned, and the first-answer from the first-answer system is irrespective of the language. All the answers are the same, "101" to "NNN".

  FIG. 19 shows a specific example of the primary answer, the language identifier that is a part of the unique attribute of the question, and the final answer candidate stored in the final answer storage unit 18 (28). In addition, in the example of FIG. 19, in order to explain handling of various languages in an easy-to-understand manner, unique answer candidates other than language information are not shown, and only final answer candidates in various languages are shown. However, in the same way as the example described with reference to FIGS. 5 to 11, in addition to the final answer candidates for each language, unique answer candidates that differ for each system user (provider) are prepared for each type of answer language. Therefore, the final answer in the appropriate answer language can be configured to be returned to the question terminal 90.

When the question answering system obtains the primary answer, the final answering means 15 obtains the language identifier of the answer language from the unique attribute information of the question stored in the operation history storage unit 13, and uses it as the primary answer and the language identifier. The corresponding final answer candidate is read from the final answer storage unit 18. For example, when the identifier “101” is acquired as the primary answer and the language identifier of the answer language is English, the final answer candidate in English as shown in FIG. 19 is read from the final answer storage unit 18. It is issued and transmitted to the inquiry terminal 90 as the final answer.
In the example shown in FIG. 19, since there is no unique answer candidate for each system user (trader), the final answer candidate in English is directly returned as the final answer. By providing them in correspondence with each other, it is possible to return the English unique answer candidate as the final answer.

  In the above description, the questioner and the answer language are individually designated by the questioner. However, the information regarding the question language can be automatically discriminated by the question acceptance answering system 10 (26) or the primary answering system 30 (27) from the form of the question sentence as described above. Even if there is no input from the interrogator, it can be obtained as common information necessary for communication between the interrogation terminal and the question answering system, based on the language information set in the interrogation terminal OS, Web browser, etc. is there. Therefore, the question language and the answer language may be configured to be set by the user for each question, and the question language may be recognized based on the information automatically available in the question acceptance answering system 10 (26) and the initial setting may be performed. It is possible to reply in the specified language or to combine the specification by the questioner and the initial setting.

  For the recognition of language information, the questioner's language is designated, the language is recognized by the question acceptance answering system 10 (26) or the primary answering system 30 (27) by the automatic discrimination function from the question sentence, or the web browser, OS, etc. Regarding the language recognition based on the set language information, it is preferable to prioritize and determine the question language. The answer language may be the same as the question language, or may be initially set to answer in a language different from the question language. The answer language determination rule can be appropriately adopted based on the convenience of the questioner, the system efficiency, and the like.

FIG. 20 shows an example of the processing procedure when the types of question languages are different.
Steps S1 to S4 and S9 to S10 are the same as the steps (steps) shown in FIG. 5 although the information handled by the unique attribute information of the question is slightly different, so that the same numbers as in FIG. The explanation will be simplified. As described in FIG. 5, when the question q1 is received from the question terminal 90 (step S1), the unique attribute information of the question is acquired (language information of the question, information of the system user, etc.), for example, operation history storage. It is stored in the unit 13 (step S2). In this step S2, the language information of the question and the like are acquired and stored as the unique attribute information, which is a difference from step S2 of FIG. Then, the question received from the question terminal 90 is transmitted to the primary answer system 30 (27) to request the primary answer (step S3). When the question q1 is received, the primary answer system 30 (27) extracts and outputs one or more primary answers corresponding to the question, and the question acceptance answer system 10 (26) makes the primary answer system 30 ( The first answer from 27) is received (step S4). The processing steps up to this point are the same as those described with reference to FIG. 5, except that in step S2, the language information of the question is acquired and stored as the unique attribute information of the question.

As described above, the language information of the question is acquired based on the communication information between the question terminal 90 and the question acceptance / answer system 10 (26), even if it is acquired by the input from the questioner. It may be automatically acquired in step 26). When the answer language is designated by the questioner, it is preferable that the answer language information is also stored as one of the language information in step S2. Such language information is stored as an identifier of a predetermined language.
In the following description, the language of the question is Chinese, the questioner has not specified anything as the answer language, and if the question language is Chinese, the final answer is also the same language as the question. Explain as if you answer in words.
In addition, in order to reduce the number of languages to answer questions in multiple languages, or if the Chinese language cannot be prepared for the answer in time and the final answer storage section does not have the final answer in Chinese, answer in English. It can also be configured to do so. In this case, the final answer storage unit 18 (28) stores the final answer candidate in English in association with the Chinese identification information and the final answer when the answer language is Chinese. Various measures are possible, such as a configuration in which the final answer candidate in English in the storage unit 18 (28) is initially set to be read.

  When the primary response is received from the primary response system 30 (27) (step S4), the identifier of the response language and the identifier of the system user stored in the operation history storage unit 13 and the like are acquired (step S55). ). Since the question language is now Chinese and the questioner has not specified the answer language, the Chinese language identifier is acquired as the answer language according to the initial settings. At the same time, it is confirmed whether or not the first answer satisfies a predetermined condition (step S56).

When the primary response satisfies a predetermined condition such as having a certain degree of certainty of answer (S56: Yes), the response from the final response storage unit 18 is returned based on the identifier of the primary response and the identifier of the response language. The final answer candidate corresponding to the language is read (step S57). Here, when the unique answer candidates other than the language are stored in the final answer storage unit 18 (28), the unique answer candidates corresponding to the system user or the like are read based on the identifier of the system user or the like. .
For example, referring to the examples of FIGS. 15 to 19, if there is a question “I came to get the resident's card” in Chinese, the expected question Q2 (7) illustrated in FIG. Therefore, the question acceptance reply system 10 (26) receives the primary reply identifier “102” (step S4) from the primary reply system 30. Further, since the answer language information has not been specified at the time of the inquiry, the Chinese language identifier is acquired as the answer language in step S55. After that, as shown in FIG. 19, based on the primary response identifier “102” and the Chinese language identifier, the final response storage unit 18 (28) stores the primary response identifier “102” in Chinese. The final answer candidate is read (step S57). Although FIG. 19 does not show a concrete document of Chinese final answer candidates, it actually stores the final Chinese answer.

Note that in the example shown in FIG. 19, no unique answer candidate other than the corresponding language is prepared, so the final answer shown in FIG. 19 becomes the final answer as it is (step S57) and is transmitted to the question terminal 90 (step S9). ).
If there is no unique answer candidate other than the corresponding Chinese language, but there is a unique answer candidate in the corresponding predetermined language (eg, Japanese), the unique answer candidate in that language (Japanese) is the final answer. It may be configured to reply as. Finally, the final answer returned to the questioner and other information are stored in the operation history storage unit 13 (step S10).

If the first answer does not satisfy the predetermined condition (S56: No), the final answer candidate prepared for each answer language, for example, the final answer candidate such as "Please ask more concretely." Is acquired from the final answer storage unit 18 (28) based on the answer language information (step S58) and transmitted to the question terminal 90 (step S9). Alternatively, a similar final answer may be generated by a program (step S58) and transmitted to the question terminal 90 (step S9).
Here, instead of preparing individual final answer candidates for each answer language, for example, prepare multiple documents as "final answer questions" in all answer languages as one final answer candidate. If the primary answer does not satisfy the predetermined condition, the final answer candidates for all the questions may be read and returned.
Then, various information from the question regarding the question to the final answer is stored in the operation history storage unit (step S10).

<Question Answering System According to Another Embodiment of the Present Invention: Entity Function>
Further, another embodiment of the present invention will be described.
When the primary answering system 30 (27) equipped with artificial intelligence has a function of notifying an object (entity) included in a question, a predetermined entity is registered in advance in the primary answering system. By doing so, it becomes possible to significantly reduce the number of supposed questions to be machine learned.
When the primary answer system 30 (27) has a function of outputting the entity included in the question, by registering the entity in advance, the primary answer system sends the primary answer (identifier) together with the primary answer (identifier). , Can receive the entities included in the question.
In order to utilize the output function of this entity, create a plurality of assumed questions that have a question group that uses the "intent of question" as a common attribute in the assumed question and the target (entity) included in the intention of the question as a variable, The necessary entities are also registered in the primary response system 30 (27). In this way, by registering multiple entities corresponding to each question group and preparing multiple expected questions with entities as variables, the number obtained by multiplying the number of registered entities by the number of created expected questions. This means that only the expected questions have been prepared, and the number of expected questions that are actually machine-learned can be greatly reduced. Here, the registration of the entity can be performed as an entity common to the system, instead of registering the entity for each question group.

  21 to 27, a method of assuming a question and making a final answer in a question answering system using an entity will be described with reference to specific examples. 21 to 27 are examples of the question answering system 25 (25 ') in the city hall. FIG. 21 exemplifies common attributes of assumed questions belonging to each question group Q1 to QN and identifiers of primary answers. In the example shown in FIG. 21, the question groups Q1 to QN-2 are composed of assumed questions having the same common attribute as before.

  In each of the question groups Q1 to QN-2, the assumed questions are the same as those explained so far, and the primary answers corresponding to the respective question groups, as illustrated in FIGS. 101 ”to“ NNN-2 ”are prepared. Therefore, when the primary answer system 30 (27) receives the question and the question corresponds to any of the envisioned questions in the question groups Q1 to QN-2, the corresponding primary answers “101” to “NNN”. "-2" is provided from the primary answering system 30 (27) to the question acceptance answering system 10 (26). When any of "101" to "NNN-2" is received as the primary answer, the question acceptance answering system 10 (26) receives the final answer corresponding to the identifier of the primary answer received as illustrated in FIG. Acquire answer candidates (basic answer candidates or unique answer candidates) and reply to the question answering system.

On the other hand, the common attributes of the question groups QN-1 and QN in FIG. 21 are “how to get to the counter” and “acquisition of required documents”, respectively. Entity) is not specified. In addition, as shown in FIGS. 25 and 26, the assumed questions included in each question group QN-1 and QN include, as shown in FIG. 25 and FIG. The target is not included and the variable is “xxx”. That is, in the question groups QN-1 and QN, the common attribute is "intent to do something", and the object (entity) is the "object of what you want to do". , Or the desired “document”) is a variable. Hereinafter, such question groups QN-1 and QN will be referred to as “intent (intention) question group”.
On the other hand, if the question from the questioner is to ask for the destination or obtain documents, the question must include a specific object. Therefore, when "NNN-1" or "NNN" is output as the primary answer from the primary answer system, the entities included in the question should also be output at the same time.

In the question receiving and answering system, when the intent (intent) question group QN-1 or the identifier “NNN-1” or “NNN” corresponding to the QN is received, based on the intent (intent) included in the question. You must then create and submit an appropriate final answer candidate for your question.
Therefore, the question acceptance answering system has a correspondence table showing the final answer identifier for selecting the final answer corresponding to the registered entity, as shown in FIG. 28, and the entity included in the question is included. With reference to, the final answer candidate identifier storing the final answer candidate corresponding to the entity is acquired. When the corresponding final answer candidate exists in the identifiers "101" to "NNN-2" corresponding to the normal question groups Q1 to QN-2, these final answer candidate identifiers are designated. If there is no appropriate final answer for the identifiers "101" to "NNN-2", a final answer identifier and a final answer candidate may be separately provided.

When the first answer is the first answer “NNN-1” or “NNN” corresponding to the intent (intention) question group, the question acceptance answer system receives the first answer from the first answer system. The corresponding final answer candidate identifier is obtained from the next answer (identifier) and the entity identifier, and the final answer is obtained.
For example, suppose that the question from the question terminal is "I came to the move-in procedure." Since there is no corresponding expected question for this question in question groups QN-1 and QN, it is assumed that the intent (intention) question group QN-1 (1) becomes the primary answer "NNN-1". In that case, the identifier of the primary answer with the highest answer certainty factor is "NNN-1" and the entity "transfer procedure" is provided. Upon receiving the primary answer identifier “NNN-1” and the entity identifier “transfer procedure”, the question acceptance answering system acquires the identifier “101” of the final answer candidate from the correspondence table illustrated in FIG. 28, and FIG. The final answer candidate corresponding to the primary answer identifier “101” shown in is read and transmitted to the question terminal as the final candidate.

  If the entity does not exist even though the primary answer “NNN-1” or “NNN” corresponding to the intent (intention) question group is received, the final answer candidate cannot be acquired, and therefore FIG. As shown, the answer such as "I didn't know which contact you want to go to" or "I didn't know what kind of document was needed" was returned to the question terminal as the final answer. In such a case, it is assumed that a question or a required document not registered as an entity is specified as a question from the question terminal. By collecting and analyzing such log information and changing the envisioned question and the registration of the entity, it is possible to increase the correct answer rate to the question and evolve the question answering system 25 (25 ').

  In the above explanation, as the final answer candidates, an example in which the “basic answer candidate” and the “unique answer candidate” are directly taken as the final answers, an example in which the “basic answer candidate and unique answer candidates in various languages” are taken as the final answers, The example in which the basic answer candidate obtained from the combination of the primary answer (identifier) and the entity identifier is used as the final answer has been described, but other combinations are possible and all the configurations obtained by combining the described configurations with each other are possible. Also belongs to the scope of the present invention.

In the embodiment described above, only an example of a question group by natural language processing is shown, but the question handled by the question answering system 25 (25 ′) according to the present invention is a question based on a natural language format. Not limited. An image analysis and image recognition system may be incorporated in the question answering system 25 (25 ′), and the final answer related to the image may be transmitted by receiving the image data from the inquiry terminal. With regard to the response to the question by the image, for example, when image data of a tangible object such as a specific building or building, animal, or person is received as a question from the question terminal, information regarding the image data is returned as a final answer. It is a thing.
Specifically, when the exterior photograph of the equipment such as the kitchen handled by the remodeling company is accepted as a question, the information about the equipment may be provided as the first answer. The question answering system 25 (25 ′) can return an appropriate answer for each trader to the questioner by registering the information about the equipment as a unique answer for the company.

Further, by receiving the image data from the question terminal, the answer of the information regarding the image can be returned in a language suitable for the questioner.
As a specific application example, it can be used, for example, as a tourist guide. In this case, it is possible to receive an image of the destination from the question terminal, and guide the way to the destination (train connection method, etc.) by searching the image recognition technology and means of transportation. By acquiring the language information from the photograph and the communication information, the explanation of the transfer method can be returned in an appropriate language.
Similarly, it is possible to provide guidance in various languages for explanations of works in museums and the like.

10, 26 Question acceptance answer system 11 Communication interface 12 Question acceptance means 13 Operation history storage section 14 Primary answer inquiry means 15 Final answer means 16 Maintenance means 18, 28 Final answer storage section (Final answer database)
25 25 'Question answering system 27, 30 Primary answering system 35 Final answer control section 36 Unique attribute information obtaining section 37 Final answer obtaining section 38 Final answer sending section 40 User system 90 Questioning terminal 91 Communication network (Complex including LAN Network: Wired and unrelated)

Claims (20)

A question receiving and answering system using a computer that receives a question from a questioning terminal and returns a final answer to the question to the questioning terminal; and an artificial intelligence, which returns a primary answer to the question on request A question answering system provided separately from a primary answering system,
The first answer system is
A first response to the question from the artificial intelligence is input to the artificial intelligence by accepting a reply request for the primary answer to the question from the question acceptance and response system, and the first answer to the question from the artificial intelligence. Interface means for transmitting the next answer to the question acceptance answer system,
By machine learning a plurality of assumed questions and a predetermined identifier as a primary answer corresponding to the assumed questions, a first answer to the question from the question acceptance / answer system in response to a request from the question acceptance / answer system. Artificial intelligence for outputting a primary answer as the identifier,
The question acceptance answer system is
Question accepting means for accepting a question from the question terminal,
A primary answer inquiry that transmits the question accepted by the question acceptance means to the primary answer system, requests the primary answer, and receives the primary answer from the primary answer system. Means and
A final answer storage unit storing a final answer candidate corresponding to the identifier in a 1: 1 association with the identifier machine-learned by the artificial intelligence;
Based on the primary response received from the primary response system, the final response candidate corresponding to the identifier of the primary response is read from the final response storage unit, and the final response candidate is read out. And a final answer means for returning a final answer to the question terminal based on the question answering system. A question receiving and answering system using a computer that receives a question from a questioning terminal, artificial intelligence, and machine learning of a plurality of supposed questions and a predetermined identifier as a primary answer corresponding to the supposed questions, The question receiving and answering system in a question answering system, which is separately provided with a primary answering system that outputs a primary answer to a question from the question accepting and answering system as the identifier,
Question accepting means for accepting a question from the question terminal,
A primary answer inquiry that transmits the question accepted by the question acceptance means to the primary answer system, requests the primary answer, and receives the primary answer from the primary answer system. Means and
A final answer storage unit that stores a final answer candidate corresponding to the identifier in a 1: 1 association with the machine-learned identifier in the primary answer system;
Based on the primary response received from the primary response system, the final response candidate corresponding to the identifier of the primary response is read from the final response storage unit, and the final response candidate is read out. A final answer means for returning a final answer to the question terminal based on
A question reception and answering system.   The artificial intelligence of the first-order answering system has a plurality of the expected questions that form one or a plurality of the expected questions having a common attribute as one group, and a plurality of the expected questions and the question group 1: 1. 3. The question answering system according to claim 1 or claim 2, wherein machine learning is performed such that one identifier corresponding to the question is used as a primary answer of the assumed question included in each question group. Question acceptance answer system. The question accepting unit further obtains unique attribute information of the question associated with the question from the question terminal,
The question answering system according to claim 1 or 3, or the claim 2 or 3 wherein the final answer storage unit stores a final answer candidate corresponding to the unique attribute information as necessary. Question acceptance answer system described in. The question receiving means, based on the acquired unique attribute information, acquires the identification information of the user system through which the question has passed,
The final answer storage unit stores, as each final answer candidate, a basic answer that is a base of the final answer stored in association with each of the identifiers, or, if necessary, the identifiers and the identification information of the user system. Stored in association with each of the basic answers and unique answers with different contents,
The final answer means reads and reads the basic answer or the unique answer from the final answer storage unit based on the identifier of the primary answer and the identification information of the user system corresponding to the question. The question answering system according to claim 4 or the question acceptance answering system according to claim 4, wherein the basic answer or the unique answer is transmitted to the question terminal as a final answer. The question accepting means obtains the identification information of the user system as the unique attribute information from the communication route information that has passed through the question acceptance answering system or the information given in the communication process. The question answering system according to claim 5 or the question acceptance answering system according to claim 5. The artificial intelligence of the first answer system is
All the expected questions composed of a plurality of question groups including a plurality of expected questions having a common attribute of different languages in one question group, and the plurality of expected questions of different languages included in the same question group The primary answer to the above is machine learning to output all the primary answers with the same identifier regardless of the type of language to any of the assumed questions. 7. The system according to claim 3, wherein when the question is answered, the identifier corresponding to the question is output to the question acceptance / answer system as the primary answer to the question. The question answering system according to claim 1 or the question acceptance answering system according to any one of claims 3 to 6. The question acceptance answer system is
The question accepting means acquires question language information included in the unique attribute information of the received question,
The final answer storage unit stores the final answer candidates in a plurality of languages having substantially the same content in an answer language determined based on the question language information in association with the identifier and the type of the answer language. Cage,
The final answer means, to obtain the information of the answer language corresponding to the question based on the acquired question language information, based on the identifier of the obtained primary answer, and the information of the answer language, 8. The final answer candidate in the answer language corresponding to the question is read out, and the final answer in the answer language corresponding to the question language is returned as the final answer corresponding to the question. The question answering system according to claim 1 or the question acceptance answering system according to claim 7. The primary answer system extracts the primary answer to the question and outputs the answer certainty factor,
The question acceptance answering system may be a base of a final answer stored in association with each of the identifiers stored in association with the primary answer according to an answer certainty factor received from the primary answer system. Basic answer, or if necessary, read from the final answer storage unit a unique answer that is stored in association with each of the identifiers and the identification information of the user system through which the question has passed and that is different in content from each of the basic answers , The question answering system according to any one of claims 1 or 3 to 8, or the claim answering system according to any one of claims 2 to 8, wherein a final answer to the question is transmitted to the inquiry terminal. Question acceptance answer system. further,
An operation history storage unit for storing various information from the question acceptance to the final answer , including the communication record with the question terminal, the communication record with the first answer system, and the question unique attribute information.
Depending on the authority, the user system browses the various information stored in the operation history storage unit, and corrects or adds the final answer candidate or the unique answer stored in the final answer storage unit, Maintenance means capable of maintenance including changes,
The question answering system according to any one of claims 1 or 3 to 9, or the question acceptance and answering system according to any one of claims 2 to 9. The maintenance means limits the content of the final answer candidate, the basic answer, or the unique answer stored in the final answer storage unit to the supposed question of the primary answer system according to authority. 11. The question answering system according to claim 10 or the question accepting and answering system according to claim 10, characterized in that it can be added, modified, changed, or deleted without being updated. A question receiving and answering system using a computer that receives a question from a questioning terminal and returns a final answer to the questioning terminal to the questioning terminal, a plurality of supposed questions, and a predetermined identifier as a primary answer corresponding to the supposed question Primary answer system in a question answering system that is provided with a primary answer system having artificial intelligence that outputs a primary answer to a question from a question acceptance answer system by machine learning And
A question from the question acceptance and answer system and a request for the first answer are accepted, the question is input to the artificial intelligence, and a primary answer to the question from the artificial intelligence is given to the question acceptance and answer system. Interface means for sending
A plurality of expected questions in different languages and a plurality of the expected questions as a plurality of question groups in which the plurality of the expected questions having a common attribute regardless of the language are one group and each of the question groups are 1: Each identifier corresponding to 1 is machine-learned as a primary answer common to the plurality of assumed questions belonging to the individual question group, and when the question is received from the question acceptance answer system, the question is answered. Artificial intelligence for outputting the first answer as the identifier to the interface means;
A primary answering system including. A question accepting and answering system for accepting a question from a questioning terminal by a computer and returning an answer to the question to the questioning terminal, and artificial intelligence, and a plurality of expected questions and a predetermined answer to the plurality of expected questions A question answering method in a question answering system comprising: a primary answering system that machine-learns an identifier as a primary answer and returns an identifier as a primary answer in response to an input question,
(A) in the question acceptance and answer system, accepting a question from a question terminal and transmitting the question to a primary answer system;
(B) In the primary answer system, accepts the question from the question acceptance answer system, inputs the question to the artificial intelligence, and outputs the primary answer to the input question as the identifier. And transmitting the identifier of the primary answer output from the artificial intelligence to the question acceptance answer system,
(C) In the question receiving and answering system, when the identifier is received as the primary answer from the primary answering system, the final answer candidate is associated with the identifier that the artificial intelligence is machine learning. Reading the final answer candidate corresponding to the received identifier from the final answer storage unit that stores, and returning the read final answer candidate as a final answer to the question terminal,
Question answering method for question answering system. A question accepting and answering system for accepting a question from a questioning terminal by a computer and returning an answer to the question to the questioning terminal, and artificial intelligence, and a plurality of expected questions and a predetermined answer to the plurality of expected questions A question answering method of the question acceptance answering system in a question answering system, comprising: a primary answering system that machine-learns an identifier as a primary answer and returns an identifier as a primary answer in response to an input question. There
(A1) In the question acceptance and answer system, a step of accepting a question from a question terminal and transmitting the question to the first answer system,
(B1) When the identifier is received from the primary response system as the primary response, the final response is stored as a final response candidate in association with the identifier machine-learned by the artificial intelligence. Reading the final answer candidate corresponding to the received identifier from the storage unit and returning the read final answer candidate to the question terminal as a final answer;
Question answering method of a question acceptance answering system. The step (a) or (a1) includes
(A-1) a step of accepting a question from a question terminal,
(A-2) a step of acquiring and storing unique attribute information of the question associated with the question,
(A-3) sending the question to the primary answering system,
Equipped with
The step (c) or (b1) includes
(C-1) receiving the identifier as the primary response from the primary response system,
(C-2) a step of acquiring the identification information of the user system from the unique attribute information stored in the step (a-1),
(C-3) Identification of the primary response and the user system from the response storage unit that stores the final response candidate corresponding to the identifier by dividing it into a basic response and a user-specific unique response. Reading the corresponding basic answer or the specific answer based on information;
(C-4) returning the read basic answer or unique answer to the question terminal,
The question answering method of the question answering system according to claim 13 or the question acceptance answering system according to claim 14 . A question receiving and answering system for receiving a question from a question terminal by a computer and returning an answer to the question to the question terminal;
It has artificial intelligence, and all the expected questions that are composed of multiple question groups that include multiple expected questions that have a common attribute of different languages in one question group, and different languages that are included in the same question group. The primary answers to the plurality of expected questions consisting of are machine-learned to output the primary answers of the same identifier for any of the expected questions regardless of the language type, and are input. A method for answering a multilingual question in a question answering system comprising a primary answering system that returns the identifier as a primary answer in response to the question
(A) In the question receiving and answering system, a question is received from a question terminal, the language information of the question is acquired and stored from the unique attribute information of the question associated with the question, and the question and the primary If the answering system needs the language information of the question, sending the question together with the language information of the question to the primary answering system;
(B) In the primary answer system, accepts the question from the question acceptance answer system, inputs the question to the artificial intelligence, and outputs the primary answer to the input question as the identifier. And transmitting the identifier of the primary answer output from the artificial intelligence to the question acceptance answer system,
(C) In the question receiving and answering system, when the identifier is received as the primary answer from the primary answering system, the final answer candidate is associated with the identifier that the artificial intelligence is machine learning. Reading the final answer candidate corresponding to the received identifier from the answer storage unit storing, and returning the read final answer candidate as a final answer to the question terminal,
A method for answering multilingual questions in a question answering system, comprising: A question receiving and answering system for receiving a question from a question terminal by a computer and returning an answer to the question to the question terminal;
It has artificial intelligence, and all the expected questions that are composed of multiple question groups that include multiple expected questions that have a common attribute of different languages in one question group, and different languages that are included in the same question group. The primary answers to the plurality of expected questions consisting of are all machine-learned so as to output the primary answers having the same identifier to any of the expected questions regardless of the language type. A question acceptance response method of the question acceptance answer system in a question answering system comprising a primary answer system that returns an identifier as a primary answer in response to a question,
(A1) In the question acceptance and answer system, the question from the question terminal is accepted, the language information of the question is acquired and stored from the unique attribute information of the question associated with the question, and the question and the first If the next answer system requires the language information of the question, sending the question together with the language information of the question to the first answer system;
(B1) When the identifier is received from the primary response system as the primary response, the final response is stored as a final response candidate in association with the identifier machine-learned by the artificial intelligence. Reading the final answer candidate corresponding to the received identifier from the storage unit and returning the read final answer candidate to the question terminal as a final answer;
Question answering method of question answering system. The step (c) or (b1) includes
(C-1) receiving the identifier as the primary response from the primary response system,
(C-2) a step of acquiring answer language information determined according to a decision rule of an answer language based on the language information stored in the step (a) or (a1),
(C-3) The final answer candidate stored in the final answer storage unit based on the answer language designated by the answer language information, which is associated with the machine-learned identifier and the answer language information, is the primary answer. Reading the final answer candidate corresponding to the identifier received as an answer and the acquired answer language information;
(C-4) returning the read final answer candidate in the answer language to the question terminal as a final answer,
The question answering method of the question answering system according to claim 16 or the question answering method of the question acceptance answering system according to claim 17 . A question receiving and answering system that receives a question from a questioning terminal and returns an answer to the questioning terminal; and an artificial intelligence that outputs a primary answer to the question as an identifier in response to a request from the question receiving and answering system A response method of the primary response system in a question response system using a primary response system, comprising:
(A) A large number of supposed questions forming a plurality of question groups including a plurality of supposed questions having a common attribute regardless of language in one question group correspond to each of the plurality of question groups 1: 1. Causing the artificial intelligence to perform machine learning by using an identifier as a primary answer to each of the assumed questions;
(B) inputting the question received from the question receiving and answering system to the artificial intelligence, and outputting the identifier as a primary answer from the artificial intelligence;
(C) transmitting the identifier output as the first answer to the question acceptance and answer system as a first answer to the question;
An answering method of the primary answering system comprising: 20. In the step (a), the plurality of expected questions having a common attribute included in each question group include expected questions having substantially the same meaning but different languages. Reply method of the first answer system.