JP7268070B2 - Answering device, answering method, answering program - Google Patents

Description

The present invention relates to an answering device, an answering method, and an answering program.

Information providing technology for providing various information to users is known. As an example of such information providing technology, there is known a technology of selecting information to be provided in response to a question from a user to whom information is provided. For example, on a so-called Q&A (Question & Answering) site, there is known a technique of selecting an ideal answer according to a question of a questioner and notifying the questioner of the selected answer.

JP 2012-033076 A

However, a fixed answer to a question is not always optimal for the questioner. For example, if the answer to the question is an answer to the question, but the expression is not polite, the questioner may be offended.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to provide users with more appropriate information.

In order to solve the above-described problems and achieve the purpose, the answering device according to the present disclosure analyzes information including the question text input by the user, and an analysis unit that sets the conditions of the answer text desired by the user. and a selection unit that selects an answer sentence based on the answer sentence conditions set by the analysis unit that the user desires.

According to one aspect of the embodiment, more appropriate information can be provided to the user.

FIG. 1 is a diagram showing an example of an information providing device according to the first embodiment. FIG. 2 is a diagram showing a configuration example of an information providing device according to the first embodiment. FIG. 3 is a diagram showing an example of information registered in a user database according to the first embodiment. FIG. 4 is a diagram showing an example of information registered in the Q&A database according to the first embodiment. FIG. 5 is a diagram showing an example of the flow of reply processing according to the first embodiment. FIG. 6 is a diagram illustrating an example of a hardware configuration;

Embodiments for implementing the answering device, answering method, and answering program according to the present application (hereinafter referred to as "embodiments") will be described in detail below with reference to the drawings. Note that the answering device, answering method, and answering program according to the present application are not limited to this embodiment. Further, in each of the following embodiments, the same parts are denoted by the same reference numerals, and overlapping descriptions are omitted.

[First embodiment]
[1-1. About an example of an information providing device]
First, with reference to FIG. 1, an example of an information providing device 10, which is an example of an answering device according to an embodiment, will be described. In the following description, as an example of answer processing performed by the information providing apparatus 10, an example of answer processing performed in a service related to a so-called Q&A site will be described, but the embodiment is not limited to this. As will become clear from the description given later, the answer processing can be applied to arbitrary services provided to users, such as various community sites, in addition to services related to Q&A sites.

Further, in the following description, a user who posts a question or the like to the Q&A site is referred to as a questioner, and a user who posts a question or the like as an answer is referred to as an answerer. In addition, in the following description, the information received from each user, such as a question sentence and an answer sentence, may be collectively referred to as posted information posted by each user.

FIG. 1 is a diagram showing an example of an information providing device according to the first embodiment. An information providing device 10 shown in FIG. 1 is an information providing device that performs reply processing, and is realized by, for example, a server device, a cloud system, or the like. For example, the information providing apparatus 10 can be connected to various wireless communication networks such as 4G (Generation), 5G, LTE (Long Term Evolution), Wifi (registered trademark), wireless LAN (Local Area Network), etc., or networks such as various wired communication networks. N (see, for example, FIG. 2), communication is performed between the terminal devices QT1, RT1 to RT3 used by each user.

Terminal devices QT1, RT1 to RT3 (hereinafter sometimes collectively referred to as “terminal devices T”) are implemented by smart devices such as PCs (Personal Computers), server devices, smart televisions, smartphones, tablets, and the like. , the network N, and can communicate with the information providing apparatus 10 . The terminal device T has a screen such as a liquid crystal display and has a touch panel function. It may have a function capable of accepting various operations on the content to be processed.

In the example shown in FIG. 1, terminal device QT1 is a terminal device used by user Q1, and terminal devices RT1 to RT3 are terminals used by different respondents R1 to R3, respectively. . Also, in the example shown in FIG. 1, one user Q1 and three respondents R1 to R3 are described, but the embodiment is not limited to this. The information providing device 10 may receive any number of questions from any number of questioners, and may receive any number of answers from any number of respondents.

For example, in the example shown in FIG. 1, the information providing apparatus 10 uses a data set of question sentences and answer sentences to learn the relationship between the characteristics of question sentences and the characteristics of answer sentences, and generates a learning model (step S1). The information providing device 10 receives input of a question from the user (step S2). The information providing device 10 sets the conditions of the answer sentence requested by the questioner from the question sentence received from the user (step S3). The information providing device 10 accepts input of reply sentences from other users (step S4). The information providing device 10 inputs the question sentence of the questioner to the learning model and obtains the output of the answer sentence. (Step S5). The information providing device 10 selects an answer sentence based on the condition of the answer sentence requested by the questioner (step S6). The information providing device 10 converts the style of the selected answer (step S7). The information providing device 10 causes the user's terminal device QT1 to display the response text with the converted style (step S8).

[1-2. Regarding response processing]
Here, on the Q&A site, by improving the quality of the answers to each question, it is possible to encourage the posting of new questions, and as a result, the number of views of questions and answers will increase, and the Q&A site as a whole will be activated. It is considered possible.

Therefore, the information providing apparatus 10 acquires information including a question from the user, analyzes the information including the question input by the user, and sets the conditions for the answer requested by the user. After setting the conditions of the answer sentence requested by the user, the answer sentence is selected based on the conditions of the answer sentence requested by the user.

For example, user Q1 asks question A: “I was able to purchase a Uniqlo mask. Processing when a question is input to the terminal device QT1 will be described. In this case, the information providing device 10 acquires the information of the question text A from the terminal device QT1, analyzes the question text A, and sets the conditions of the answer text requested by the user Q1. Specifically, for example, the information providing device 10 determines from the question text A that the condition of the answer text requested by the user Q1 is Condition 1: "Whether free shipping is available only to first-time purchasers of UNIQLO masks online. include information about Further, the information providing apparatus 10 may set the condition 2 "Information about the shipping fee when purchasing a UNIQLO mask at another online mail order site is included" as an additional condition. When the conditions 1 and 2 are set as the conditions of the answer sentences requested by the user Q1, the information providing apparatus 10 selects the answer sentences from the answer sentence set indicating information collected from the answers posted in the past by the users of the Q&A site. answer sentences that satisfy the conditions 1 and 2 are selected. When answer sentences satisfying conditions 1 and 2 are selected, the information providing device 10 generates content information for displaying the selected answer sentences on the terminal device QT1, and transmits the content information to the terminal device QT1.

As a result, it is possible to increase the possibility that the user who entered the question text will obtain the information that the user wants to obtain by entering the question, so the user's reliability of the Q & A site will increase. It is possible to improve the frequency of user's questions. As a result, the Q&A site can be activated.

[1-3. Concerning the conversion of the writing style of the answer sentence]
On the other hand, even if the content of the answer to the question is correct, depending on the style of the answer sentence, the questioner may be offended, and as a result, the questioner's frequency of questioning may be lowered, and the Q&A site may not be activated.

Therefore, when the answer sentence is selected based on the conditions of the answer sentence requested by the user, the information providing apparatus 10 changes the style of the selected answer sentence.

For example, when the user Q1 inputs the above-mentioned question text A as a question text, the information providing device 10 outputs the answer text A, "3 UNIQLO masks, free shipping is a good deal," and the answer Sentence B Suppose that the response sentence "Site A charged for the shipping of UNIQLO masks, so free shipping is a good deal!" is selected. In this case, the information providing apparatus 10 converts the style of the answer sentence A and the answer sentence B. FIG. For example, the information providing apparatus 10 responds to an answer sentence A and an answer sentence B written in ordinary sentences (“da”/“desuru” tone) in honorific sentences (“desu”/“masu” tone). ), and convert it to the answer sentence AX, "For 3 Uniqlo masks, free shipping is a good deal," and the answer sentence BX, "Site A charged for the shipping of Uniqlo masks, so free shipping is a good deal." . Note that the writing style converted by the information providing apparatus 10 is not limited to normal style sentences to honorific sentences, and may be converted to other arbitrary style styles.

As a result, users can obtain the information they wish to obtain by inputting a question text after converting it into an appropriate style. harm can be eliminated. Therefore, the frequency of questions asked by users can be improved, and the Q&A site can be activated.

[1-4. Concerning the conversion of the writing style of the answer sentence to the writing style of the user's preference]
Here, it is thought that user satisfaction can be improved by converting the writing style of the answer sentences to the writing style of the user's preference, instead of uniformly converting the writing style of the answer sentences into a specific writing style. be done.

Therefore, the information providing apparatus 10 estimates the user's preferred writing style based on the information including the question sentence input by the user, and converts the selected answer sentence into the estimated preferred writing style. .

For example, the information providing apparatus 10 estimates the user's preferred writing style from information such as a question input by the user Q1, a word searched by the user Q1, and a website viewed by the user Q1. For example, assume that the information providing device 10 estimates that the user Q1's favorite writing style is "shonen manga style" based on the information on the website browsed by the user Q1. In addition, as in the above example, the information providing apparatus 10 provides the answer sentence A, "3 Uniqlo masks, free shipping is advantageous" and the answer sentence B, "Since the shipping fee for the Uniqlo mask is paid for by the site A. , free shipping is a good deal!". In this case, the information providing apparatus 10 converts the writing style of the answer sentence A and the answer sentence B into the "shonen manga style" which is the estimated favorite writing style of the user Q1. For example, the fonts of the answer sentence A and the answer sentence B are changed to a comic style font. The writing style converted by the information providing apparatus 10 is not limited to the "shonen manga style", and may be converted into any other writing style according to the user's preference.

As a result, users will be able to obtain the information they wish to obtain by entering a question text after converting it into the style of writing preferred by the user. can be enhanced. Therefore, the frequency of questions asked by users can be improved, and the Q&A site can be activated.

[1-5. Regarding the use of a learning model that learns the relationship between the characteristics of question sentences and the characteristics of answer sentences]
Here, there are cases where an answer requested by the user is not included in the answers registered in the Q&A database 32 provided in the information providing apparatus 10 in response to the questions input by the user. Therefore, it is considered that the Q&A database 32 can be supplemented by outputting an answer sentence from a question sentence using a learning model that has learned data of answers to various questions.

Therefore, the information providing apparatus 10 learns the relationship between the features of question sentences and the features of answer sentences using data sets of answers to various questions, and generates an answer learning model 34 . The information providing device 10 inputs the question text input by the user to the answer learning model 34 and obtains the output of the answer text.

For example, assume that the Q&A database 32 of the information providing apparatus 10 lacks data on questions and answers regarding "manga". In this case, the information providing apparatus 10 learns the characteristics of the question sentences and the characteristics of the answer sentences by using the data set of questions and answers related to “manga” acquired from the outside, and generates the answer learning model 34 . Then, if the user inputs a question about "manga", the information providing apparatus 10 inputs the question input by the user to the answer learning model 34, and outputs the answer from the answer learning model 34. obtain. The information providing device 10 executes the processing described in 1-2 based on the question text input by the user, which may be parallel processing or continuous processing, and the Q&A database 32 Select an answer from After that, the information providing device 10 determines whether or not the answer sentence output by the answer learning model 34 satisfies the condition of the answer sentence requested by the user. When it is determined that the answer sentence output by the answer learning model 34 satisfies the conditions of the answer sentence required by the user, the information providing device 10 outputs the answer sentence selected from the Q&A database 32 and the answer learning model 34. Contents information to be displayed on the terminal device of the user is generated by using the answer sentence, and is distributed to the terminal device of the user.

As a result, even if the information that the user wants to obtain by inputting the question is insufficient in the question and answer data registered in the Q&A database 32 of the information providing apparatus 10, the external By outputting an answer sentence using a learning model learned using a data set, it is possible to improve the possibility of obtaining the answer sentence desired by the user. Therefore, the frequency of questions asked by users can be improved, and the Q&A site can be activated.

[1-6. Assisting other users in answering user questions]
Here, when another user enters an answer to the question entered by the user, the other user who enters the answer is assisted to enter the answer that the user wants. This will make it easier for other users to input the answer that the user wants.

Therefore, the information providing apparatus 10 prompts the user to input information requested by the user when another user inputs an answer to the question input by the user.

For example, as described in 1-2, the user inputs the question text A, and the condition of the answer text requested by the user is Condition 1: Is free shipping available only to first-time purchasers of UNIQLO masks online? and condition 2 "Information about the shipping fee when purchasing UNIQLO masks at another online mail order site is included." In this case, when another user inputs an answer to question text A input by the user, the information providing device 10 sets condition 1 and condition 2 to the terminal device used by the other user. to display.

As a result, when another user enters an answer to a user's question, the user who entered the question to the other user can assist the user to enter an answer requested by the user. can be done. Therefore, it is possible to increase the possibility that other users will input an answer text desired by the user who input the question text. Therefore, the question frequency of the user is improved. In addition, it is possible to improve the probability of obtaining the best answer from other users who have input the answer sentence. As a result, the Q&A site can be activated.

[1-7. Regarding the calculation of the degree of fulfillment of requirements]
Here, when a numerical value indicating how much the answer text selected by the information providing device 10 satisfies the request of the user for the question text input by the user is displayed on the terminal device of the user, It is advantageous for the user to be able to immediately determine which of the answer sentences selected by the information providing apparatus 10 should be referred to first.

Therefore, the information providing apparatus 10 makes a request indicating how close the selected answer sentence is to the answer sentence desired by the user, based on the relationship between the condition of the answer sentence requested by the user and the selected answer sentence. Calculate sufficiency.

For example, assume that the information providing apparatus 10 has set four conditions, condition A to condition D, as the conditions of the reply sentence. Assume that the answer text A selected by the information providing device 10 satisfies the three conditions A to C of the four conditions A to D, but does not satisfy the condition D. In this case, the information providing device 10 calculates the request sufficiency level of the answer text A as 75%. Then, the information providing device 10 causes the terminal device of the user to display the calculated request sufficiency degree together with the information of the answer sentence A. FIG.

As a result, when the answer to the question entered by the user is displayed on the user's terminal device, the request sufficiency of the selected answer is also displayed. It is possible to immediately make a decision as to whether or not it should be referred to. Therefore, user satisfaction is improved, and the Q&A site can be activated.

[2. Configuration of Information Providing Device]
An example of the functional configuration of the information providing apparatus 10 described above will be described below. Note that the following description shows an example of the functional configuration of the information providing apparatus 10 that executes answer processing on the Q&A site. FIG. 2 is a diagram showing a configuration example of an information providing device according to the first embodiment. As shown in FIG. 2, the information providing device 10 has a communication section 20, a storage section 30, and a control section 40. FIG.

The communication unit 20 is realized by, for example, a NIC (Network Interface Card) or the like. The communication unit 20 is connected to the network N by wire or wirelessly, and transmits and receives information to and from the terminal device T. FIG.

The storage unit 30 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disc. The storage unit 30 also stores a user database 31, a Q&A database 32, a writing style learning model 33, and an answer learning model . An example of information registered in the user database 31 and the Q&A database 32 will be described below with reference to FIGS. 3 and 4. FIG.

In the user database 31, information of users who are questioners or answerers is registered. For example, FIG. 3 is a diagram showing an example of information registered in the user database according to the first embodiment. As shown in FIG. 3, in the user database 31, information having items such as "user ID (identifier)", "question history", "answer history", "search history", and "browsing history" is registered. be.

Here, "user ID" is an identifier for identifying a user. "Question history" is a history of questions posted by users. "Answer history" is a history of answers posted by users. A "search history" is a history of phrases (words, phrases, compound words, etc.) searched by the user. Also, the “browsing history” is the history of information browsed by the user among the information displayed in the search results.

For example, in the example shown in FIG. 3, the user database 31 stores user ID "U1", question history "question history #U1", answer history "answer history #U1", search history "search history #U1", Information such as browsing history “browsing history #U1” is registered. Such information includes the question indicated by the question history "question history #U1", the answer indicated by the answer history "answer history #U1", and the search history. It shows that the phrase indicated by "search history #U1" was searched and the information indicated by the browsing history "browsing history #U1" was browsed.

In the example shown in FIG. 3, information registered in the user database 31 is described using conceptual expressions such as question history #U1, answer history #U1, search history #U1, and browsing history #U1. In practice, character information used as questions and answers posted by users, character information indicating searched words and phrases, and character information indicating browsed information are registered. In addition to the information shown in FIG. 3, any other information related to the user may be registered in the user database 31. FIG.

The Q&A database 32 contains a set of questions posted by users, answers posted by other users to the questions, and questions from the set of answers posted by other users. Information related to the answer selected as the best answer by the input user is registered. For example, FIG. 4 is a diagram showing an example of information registered in the Q&A database according to the first embodiment. As shown in FIG. 4, in the Q&A database 32, information having items such as "question ID", "question sentence", "answer sentence set", and "selection information" is registered. It should be noted that an answer sentence set is information in which answer sentences posted by one or more other users to a question sentence posted by a user are grouped into one.

In the example shown in FIG. 4, information such as the question ID "Q1", the question sentence "question sentence #1", the answer sentence set "answer sentence set #1", and the selection information "answer sentence #1-1" are associated and registered. It is Such information includes the question ID "Q1", the question text "question text #1", and the question text "question text #1" from other users. 1” has been input, and the user who has entered the question has selected the answer indicated by the selection information “answer #1-1” as the best answer.

In the example shown in FIG. 4, conceptual expressions such as "question sentence #1", "answer sentence set #1", and "answer sentence #1-1" are described. Information, character information of a set of answer sentences answered to the question sentence, is registered. In addition to the information shown in FIG. 4, for example, the Q&A database 32 stores various questions such as the user ID of the user who asked the question and the user ID of the user who answered the question. and any information about the answer may be registered.

The writing style learning model 33 stores an unlearned writing style learning model and a learned writing style learning model. As the learned writing style learning model, a writing style learning model generated by execution of learning processing by the writing style learning unit 421, which will be described later, is stored. When the writing style learning unit 421 to be described later executes the learning process, the learning process is executed using the unlearned writing style learning model stored in the writing style learning model 33 .

The answer learning model 34 stores an unlearned answer learning model and a learned answer learning model. As the learned answer learning model, an answer learning model generated by executing learning processing by the answer learning unit 422, which will be described later, is stored. When the answer learning unit 422, which will be described later, performs the learning process, it uses an unlearned answer learning model stored in the answer learning model 34 to perform the learning process.

The control unit 40 is, for example, a controller, and various programs stored in the storage device of the information providing apparatus 10 are executed by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like using the RAM as a work area. It is realized by being executed. Also, the control unit 40 is a controller, and may be implemented by an integrated circuit such as an ASIC (Application Specific Circuit) or an FPGA (Field Programmable Gate Array).

As shown in FIG. 2, the control unit 40 includes a collection unit 41, a learning unit 42, a receiving unit 43, an estimating unit 44, an analyzing unit 45, a selecting unit 46, a converting unit 47, a calculating unit 48, an auxiliary unit 49, and a providing unit. a portion 50;

The collection unit 41 collects various data used for generating a learning model. For example, the collection unit 41 collects search queries input by each user in the past, questions and answers input by the user in the past, and answers selected as the best answer by the user as past logs.

The learning unit 42 includes a writing style learning unit 421 and an answer learning unit 422 . The writing style learning unit 421 learns the relationship between the writing style characteristics of the question sentence and the writing style characteristics of the answer sentence satisfied by the user, and generates the writing style learning model 33 . The writing style learning unit 421 may use, for example, question sentences registered in the Q&A database 32 and selection information indicating the best answer selected by the user as learning data. Also, the writing style learning unit 421 performs learning using, for example, an RNN (Recurrent Neural Network) or an LSTM (Long Short-Term Memory) after performing morphological analysis on an answer sentence satisfied by the user. By doing so, the writing style learning model 33 that has learned the word order of the user's favorite writing style may be generated.

The answer learning unit 422 learns the relationship between the feature of the question sentence and the feature of the answer sentence, and generates the answer learning model 34 . Here, the answer learning unit 422 may use question sentences registered in the Q&A database 32 and a set of answer sentences to the question sentences as learning data. In addition, the answer learning unit 422 uses, for example, BERT (Bidirectional Encoder Representations from Transformers) to perform pre-learning that captures the structure of the language from a huge amount of sentences without a teacher. The answer learning model 34 may be generated by readjusting the data of the set of answer sentences as learning data.

The reception unit 43 receives registration of questions and answers from users. For example, upon receiving a question text from the terminal device used by the user, the receiving unit 43 registers the received text in the Q&A database 32 as a question text. In addition, when receiving a text as an answer from a terminal device used by an answerer, the receiving unit 43 associates the received text with the question ID of the corresponding question text and registers it in the Q&A database 32 .

The estimation unit 44 estimates the user's preferred writing style based on information including the question text input by the user. The estimating unit 44 may, for example, estimate the writing style of the question text input by the user as the user's favorite writing style, or estimate the user's question based on information indicated by the search history registered in the user database 31. A favorite writing style may be estimated, or a writing style used in information viewed by the user indicated by the browsing history registered in the user database 31 may be estimated as the user's favorite writing style.

The analysis unit 45 analyzes the question text received from the user and sets the conditions required for the answer text. For example, by performing morphological analysis on the question sentence, dividing the question sentence into a string of morphemes, determining the part of speech, and then extracting the keywords contained in the question sentence, the keywords extracted in the answer sentence may be included as a condition of the answer sentence. In addition, it is also possible to make it a condition for the answer sentence to include information that corresponds to a specific category (for example, place name, time, quantity, person name) from the interrogative words and surrounding words contained in the question sentence that has undergone morphological analysis. good.

The selection unit 46 searches the answer sentences registered in the Q&A database 32 for an answer sentence to the question sentence input by the user based on the conditions required for the answer sentence set by the analysis unit 45, and selects the answer sentence. Select an answer sentence that satisfies the conditions required for For example, a user inputs a question "where is the capital of Japan?" , the selection unit 46 selects an answer sentence that satisfies the condition A and the condition B from among the set of answer sentences registered in the Q&A database 32. . If there is no answer sentence that satisfies all the conditions set by the analysis unit 45 in the set of answer sentences registered in the Q&A database 32, the selection unit 46 selects the answer sentence registered in the Q&A database 32. Among them, an answer sentence that satisfies at least one condition is selected.

The conversion unit 47 converts the style of the answer sentence selected by the selection unit 46 . Style conversion (style conversion) is a type of paraphrasing generation. For example, when the answer sentence selected by the selection unit 46 is in ordinary sentences (“da” or “desuaru” tone), the conversion unit 47 converts it into honorific sentences (“desu” or “masu” tone). Convert. The conversion unit 47 may use the writing style learning model 33 generated by the writing style learning unit 421 to convert the writing style of the answer sentence selected by the selection unit 46, for example. Further, the conversion unit 47 may, for example, separate the semantic information and the stylistic information of the sentence, and add the stylistic information to the separated semantic information. Further, the conversion unit 47 may convert the style of the answer sentence selected by the selection unit 46 using a learning model generated using GAN (Generative Adversal Networks) or VAE (Variational AutoEncoder). .

The calculation unit 48 calculates the answer sentence selected by the selection unit 46 based on the relationship between the condition of the answer sentence set by the analysis unit 45 and the answer sentence selected by the selection unit 46. Calculate the degree of fulfillment of the request, which indicates how close it is to the answer sentence. For example, when the analysis unit 45 sets four conditions, condition A to condition D, as the conditions for the answer sentence, the calculation unit 48 determines that the answer sentence selected by the selection unit 46 is from the condition A selected by the analysis unit 45 to Of the four conditions D, if the three conditions A to C are satisfied but the condition D is not satisfied, the requirement satisfaction is calculated as 75%.

The auxiliary unit 49 prompts the user to input information requested when another user posts an answer to the question posted by the user. For example, after the analysis unit 45 analyzes the question text posted by the user and sets the conditions for the response text requested by the user, the auxiliary unit 49 may In addition, it is possible to display the condition of the answer sentence requested by the user on the terminal device of the other user who is posting the answer sentence, thereby assisting the input of the answer sentence of the other user.

The providing unit 50 provides the answer text selected by the selecting unit 46 to the user who posted the question. In addition, the providing unit 50 provides the user who posted the question with an answer text whose style has been converted by the conversion unit 47 with respect to the answer text selected by the selection unit 46 . Further, the provision unit 50 may provide the request sufficiency level calculated by the calculation unit 48 to the user who posted the question. For example, the providing unit 50 generates content information including an answer text whose style has been converted by the conversion unit 47 for the answer text selected by the selection unit 46, and transmits the generated content information to the terminal device. To deliver.

[3. Processing procedure]
Next, a procedure of reply processing by the information providing apparatus 10 according to the first embodiment will be described with reference to FIG. FIG. 5 is a flowchart showing an example of the flow of reply processing according to the first embodiment. For example, the information providing device 10 acquires information including a question from the user (step S101). Then, the information providing device 10 sets the conditions of the reply sentence requested by the user (step S102). Then, when the conditions for the answer sentence are set, the information providing apparatus 10 selects an answer sentence based on the conditions for the answer sentence requested by the user (step S103). Then, the information providing apparatus 10 converts the style of the selected reply sentence (step S104).

[4. Hardware configuration]
Further, the information providing apparatus 10 according to the embodiment described above is implemented by a computer 1000 configured as shown in FIG. 6, for example. FIG. 6 is a diagram illustrating an example of a hardware configuration; A computer 1000 is connected to an output device 1010 and an input device 1020, and an arithmetic device 1030, a primary storage device 1040, a secondary storage device 1050, an output IF (Interface) 1060, an input IF 1070, and a network IF 1080 are connected by a bus 1090. have

The arithmetic device 1030 operates based on programs stored in the primary storage device 1040 and the secondary storage device 1050, programs read from the input device 1020, and the like, and executes various processes. The primary storage device 1040 is a memory device such as a RAM that temporarily stores data used by the arithmetic device 1030 for various calculations. The secondary storage device 1050 is a storage device in which data used for various calculations by the arithmetic device 1030 and various databases are registered, and includes a ROM (Read Only Memory), HDD (Hard Disk Drive), flash memory, and the like. It is realized by

The output IF 1060 is an interface for transmitting information to be output to the output device 1010 that outputs various types of information such as a monitor and a printer. It is realized by a connector conforming to a standard such as HDMI (registered trademark) (High Definition Multimedia Interface). Also, the input IF 1070 is an interface for receiving information from various input devices 1020 such as a mouse, keyboard, scanner, etc., and is realized by, for example, USB.

Note that the input device 1020 includes, for example, optical recording media such as CDs (Compact Discs), DVDs (Digital Versatile Discs), PDs (Phase change rewritable discs), magneto-optical recording media such as MOs (Magneto-Optical discs), and tapes. It may be a device that reads information from a medium, a magnetic recording medium, a semiconductor memory, or the like. Also, the input device 1020 may be an external storage medium such as a USB memory.

Network IF 1080 receives data from other devices via network N and sends the data to arithmetic device 1030, and also transmits data generated by arithmetic device 1030 via network N to other devices.

The arithmetic device 1030 controls the output device 1010 and the input device 1020 via the output IF 1060 and the input IF 1070 . For example, arithmetic device 1030 loads a program from input device 1020 or secondary storage device 1050 onto primary storage device 1040 and executes the loaded program.

For example, when the computer 1000 functions as the information providing device 10 , the arithmetic device 1030 of the computer 1000 implements the functions of the control unit 40 by executing programs loaded on the primary storage device 1040 .

[5. effect〕
The information providing apparatus 10 according to the present disclosure analyzes information including a question text input by a user, analyzes the analysis unit 45 that sets the conditions of the answer text desired by the user, and the user set by the analysis unit 45 and a selection unit 46 for selecting an answer sentence based on the condition of the answer sentence required by.

According to this configuration, the question text input by the user is analyzed, the conditions for the answer text desired by the user are set, and the answer text is selected based on the set answer conditions. You can select the answer sentence that you want.

The information providing device 10 according to the present disclosure also includes a conversion unit 47 that converts the style of the answer text selected by the selection unit 46 .

According to this configuration, the style of the answer sentence selected by the selection unit 46 is changed, so that information can be appropriately provided to the user.

In the information providing apparatus 10 according to the present disclosure, the selection unit 46 selects the estimation unit 44 and the conversion unit 47 for estimating the user's preferred writing style based on information including the question text input by the user. The reply sentence is converted into the writing style of the user's preference estimated by the estimation unit 44 .

According to this configuration, based on the information including the question text input by the user, the user's favorite writing style is estimated, and the answer text selected by the selection unit 46 is answered by the user estimated by the estimation unit 44. Since the writing style is converted into the favorite style of writing, it is possible to appropriately provide information to the user.

In addition, the information providing apparatus 10 according to the present disclosure learns the relationship between the stylistic features of the question sentence and the stylistic features of the answer sentence satisfied by the user, and the writing style learning unit 421 that generates the writing style learning model 33. , and the conversion unit 47 uses the writing style learning unit 421 to convert the answer sentence selected by the selection unit 46 into an answer sentence style that satisfies the user.

According to this configuration, the writing style of the answer sentence selected by the selection unit 46 is converted into the writing style of the answer sentence that the user is satisfied with by using the writing style learning model 33, so that information is appropriately provided to the user. can do.

The information providing apparatus 10 according to the present disclosure also includes an answer learning unit 422 that learns the relationship between the characteristics of the question sentence and the characteristics of the answer sentence, and generates the answer learning model 34. Answer sentences output using the answer learning model 34 generated by the learning unit 422 are included in the selection targets.

According to this configuration, since the output answer sentence is included in the selection target of the selection unit 46 using the answer learning model 34 that has learned the relationship between the feature of the question sentence and the feature of the answer sentence, it is more appropriate Information requested by the user can be provided to the user.

In addition, the information providing apparatus 10 according to the present disclosure includes an auxiliary unit 49 that prompts the user to input information requested by another user when another user inputs an answer to a question input by the user; Prepare.

According to this configuration, when another user enters an answer to a question entered by the user, the user is prompted to enter information requested by the user. can provide information.

In addition, the information providing apparatus 10 according to the present disclosure provides the answer selected by the selection unit 46 based on the relationship between the answer sentence conditions set by the analysis unit 45 and the answer sentence selected by the selection unit 46. and a calculation unit 48 for calculating a request sufficiency degree indicating how close the sentence is to the answer sentence desired by the user.

According to this configuration, it is possible to calculate the degree of request sufficiency that indicates to what extent the conditions of the answer sentence desired by the user are satisfied, and use the calculated degree of sufficiency of request to estimate the degree of satisfaction of the user. can be done.

The answering method according to the present disclosure includes a step of acquiring information including a question text input by a user, and a step of setting conditions for an answer text requested by the user based on the information including the question text input by the user. , a step of selecting an answer sentence based on the condition of the answer sentence requested by the user, and a step of converting the style of the selected answer sentence.

According to this configuration, it is possible to appropriately provide the information requested by the user.

The answer program according to the present disclosure includes a step of acquiring information including a question text input by the user, and a step of setting conditions for the answer text requested by the user based on the information including the question text input by the user. A computer executes a step of selecting an answer sentence based on the condition of the answer sentence requested by the user, and a step of converting the style of the selected answer sentence.

According to this configuration, it is possible to appropriately provide the information requested by the user.

As described above, the embodiments of the present application have been described in detail with reference to the drawings, but these are examples, and various modifications and improvements, including the aspects described in the section of the disclosure of the invention, can be made based on the knowledge of those skilled in the art. It is possible to implement the invention in other forms.

Also, the above-mentioned "section, module, unit" can be read as "means" or "circuit". For example, the acquisition unit can be read as acquisition means or an acquisition circuit.

10 information providing device 20 communication unit 30 storage unit 31 user database 32 Q&A database 33 writing style learning model 34 answer learning model 40 control unit 41 collection unit 42 learning unit 43 reception unit 44 estimation unit 45 analysis unit 46 selection unit 47 conversion unit 48 calculation unit 49 auxiliary unit 50 provision unit N network T terminal device

Claims (8)

an analysis unit that analyzes information including a question text input by a user and sets conditions for an answer text that the user desires;
a selection unit that selects an answer sentence based on the conditions of the answer sentence set by the analysis unit that the user desires;
an answer assistance unit that prompts the user to enter information requested by another user when the other user enters an answer to the question entered by the user;
answering device. a conversion unit that converts the style of the answer text selected by the selection unit;
An answering device according to claim 1. an estimation unit for estimating the user's preferred writing style based on information including the question text input by the user ;
3. The answering device according to claim 2, wherein the converting unit converts the answer text selected by the selecting unit into the writing style of the user's preference estimated by the estimating unit. a writing style learning unit that learns the relationship between the writing style features of the question sentence and the writing style features of the answer sentence satisfied by the user, and generates a writing style learning model;
3. The answering device according to claim 2, wherein the converting unit converts the answer sentence selected by the selecting unit into a writing style of the answer sentence that satisfies the user by using the writing style learning unit. an answer learning unit that learns the relationship between the feature of the question sentence and the feature of the answer sentence and generates an answer learning model;
The selection unit includes, in the selection target, answer sentences output using the answer learning model generated by the answer learning unit;
An answering device according to any one of claims 1 to 4. How much the answer sentence selected by the selection unit matches the answer sentence desired by the user based on the relationship between the condition of the answer sentence set by the analysis unit and the answer sentence selected by the selection unit a calculation unit that calculates the degree of fulfillment of the request indicating whether it is close,
An answering device according to any one of claims 1 to 5 . A computer - implemented answering method comprising :
an analysis step of analyzing information including a question text input by a user and setting conditions for an answer text desired by the user;
a selection step of selecting an answer sentence based on the conditions of the answer sentence set by the analysis step set by the user;
an answer assisting step of prompting the user to enter information requested by another user when the other user enters an answer to the question entered by the user ;
Answer method including. an analysis procedure for analyzing information including a question text input by a user and setting conditions for an answer text requested by the user;
a selection procedure for selecting an answer sentence based on the conditions of the answer sentence set by the analysis procedure set by the user;
an answer assisting procedure for prompting the user to enter information requested by another user when the other user enters an answer to a question entered by the user ;
An answer program that causes a computer to run

Images