Information processing device, information processing system, information processing method, and program

Abstract

In order to address the problem of improving the accuracy of responses by accurately modifying inquiries from users, this information processing device (1) comprises: an input acquisition unit (11) that acquires a first text string indicating an inquiry from a user; a generation unit (12) that refers to a term database and generates a second text string in which at least part of the first text string is replaced by a term from the term database; an evaluation acquisition unit (13) that acquires evaluation information indicating a user evaluation of inquiry results corresponding to the inquiry indicated by the first text string or the second text string; and an updating unit (14) that updates the term database on the basis of the evaluation information.

Description

[Technical Field]

[0001] The present invention relates to a technique for responding to inquiries from users. [Background technology]

[0002] There are known techniques for responding to user inquiries. For example, Patent Document 1 describes an information processing server that searches document data based on a user inquiry. The information processing server corrects a character string included in the inquiry that needs to be corrected based on words extracted from the document data that is the search result. The information processing server also outputs the results of a re-search of the document data using the correction results. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2018-190030 Summary of the Invention [Problem to be solved by the invention]

[0004] The information processing server described in Patent Document 1 extracts words to be used to correct a character string to be corrected, which is included in a user's query, from document data searched based on the query. Therefore, depending on the character string to be corrected, the document data obtained as a search result may be inappropriate or insufficient. Therefore, the information processing server may not be able to accurately extract the words to be used for correction. Therefore, the information processing server may not be able to accurately correct the character string to be corrected, and there is room for improvement in the accuracy of re-search results.

[0005] One aspect of the present invention has been made in consideration of the above-mentioned problems, and aims to provide a technique for accurately correcting a user's inquiry to improve the accuracy of the response. [Means for solving the problem]

[0006] An information processing device according to one aspect of the present invention includes an input acquisition means for acquiring a first character string indicating an inquiry from a user, a generation means for generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database, an evaluation acquisition means for acquiring evaluation information indicating an evaluation by the user of an inquiry result corresponding to the inquiry indicated by the first character string or the second character string, and an update means for updating the word database based on the evaluation information.

[0007] An information processing system according to one aspect of the present invention includes an input acquisition means for acquiring a first character string indicating an inquiry from a user; a generation means for generating a second character string by referring to a word database and replacing at least a portion of the first character string with a word in the word database; an evaluation acquisition means for acquiring evaluation information indicating an evaluation by the user of an inquiry result corresponding to the inquiry indicated by the first character string or the second character string; and an update means for updating the word database based on the evaluation information.

[0008] An information processing method according to one aspect of the present invention includes obtaining a first character string indicating an inquiry from a user; generating a second character string by referring to a word database and replacing at least a portion of the first character string with a word in the word database; obtaining evaluation information indicating an evaluation by the user of an inquiry result corresponding to the inquiry indicated by the first character string or the second character string; and updating the word database based on the evaluation information.

[0009] A program according to one aspect of the present invention is a program for causing a computer to function as an information processing device, and causes the computer to function as an input acquisition means for acquiring a first character string indicating an inquiry from a user, a generation means for generating a second character string by referring to a word database and replacing at least a portion of the first character string with a word in the word database, an evaluation acquisition means for acquiring evaluation information indicating the user's evaluation of an inquiry result corresponding to an inquiry indicated by the first character string or the second character string, and an update means for updating the word database based on the evaluation information. [Effects of the Invention]

[0010] According to one aspect of the present invention, it is possible to accurately correct a user's inquiry and improve the accuracy of the response. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a block diagram showing a configuration of an information processing device according to a first exemplary embodiment of the present invention. [Figure 2] 1 is a flowchart showing the flow of an information processing method according to a first exemplary embodiment of the present invention. [Figure 3] FIG. 2 is a block diagram showing a configuration of an information processing system according to another aspect of the first exemplary embodiment of the present invention. [Figure 4] FIG. 10 is a block diagram showing the configuration of an information processing system according to a second exemplary embodiment of the present invention. [Figure 5] FIG. 10 is a diagram showing a specific example of a document database according to the second exemplary embodiment of the present invention. [Figure 6] FIG. 10 is a diagram showing a specific example of a synonym database according to the second exemplary embodiment of the present invention. [Figure 7] FIG. 10 is a diagram showing a specific example of a word database according to the second exemplary embodiment of the present invention. [Figure 8] FIG. 10 is a flowchart showing the flow of an information processing method according to a second exemplary embodiment of the present invention. [Figure 9]FIG. 10 is a diagram showing a specific example of an evaluation information database according to the second exemplary embodiment of the present invention. [Figure 10] FIG. 10 is a flowchart showing the flow of a generation process according to a second exemplary embodiment of the present invention. [Figure 11] FIG. 10 is a flowchart showing the flow of update processing according to the second exemplary embodiment of the present invention. [Figure 12] FIG. 10 is a diagram illustrating a specific example of evaluation points according to the second exemplary embodiment of the present invention. [Figure 13] FIG. 10 is a diagram illustrating a specific example of input conversion according to the third exemplary embodiment of the present invention. [Figure 14] FIG. 10 is a block diagram showing the configuration of an information processing system according to a third exemplary embodiment of the present invention. [Figure 15] FIG. 10 is a diagram showing a specific example of a word database according to the third exemplary embodiment of the present invention. [Figure 16] FIG. 10 is a flowchart showing the flow of a generation process according to a third exemplary embodiment of the present invention. [Figure 17] 10A to 10C are diagrams illustrating a specific example of a generation process according to the third exemplary embodiment of the present invention. [Figure 18] FIG. 10 is a diagram showing a specific example of an evaluation information database according to the third exemplary embodiment of the present invention. [Figure 19] FIG. 10 is a diagram illustrating a specific example of an update process according to the third exemplary embodiment of the present invention. [Figure 20] FIG. 1 is a block diagram showing an example of a physical configuration of an information processing device according to each exemplary embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0012] Exemplary Embodiment 1 A first exemplary embodiment of the present invention will be described in detail with reference to the drawings. This exemplary embodiment is a basic form of the exemplary embodiments described below.

[0013] <Configuration of information processing device 1> The configuration of an information processing device 1 according to this exemplary embodiment will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the configuration of the information processing device 1. As shown in FIG. 1, the information processing device 1 includes an input acquisition unit 11, a generation unit 12, an evaluation acquisition unit 13, and an update unit 14. The input acquisition unit 11 is an example of a configuration that realizes the input acquisition means set forth in the claims. The generation unit 12 is an example of a configuration that realizes the generation means set forth in the claims. The evaluation acquisition unit 13 is an example of a configuration that realizes the evaluation acquisition means set forth in the claims. The update unit 14 is an example of a configuration that realizes the update means set forth in the claims.

[0014] The input acquisition unit 11 acquires a first character string indicating an inquiry from a user. The generation unit 12 references a word database and generates a second character string by replacing at least a part of the first character string with a word in the word database. The evaluation acquisition unit 13 acquires evaluation information indicating a user's evaluation of the inquiry result corresponding to the inquiry indicated by the first character string or the second character string. The update unit 14 updates the word database based on the evaluation information. Details of the first character string, the second character string, the word database, and the evaluation information will be described later in "Flow of information processing method S1" with reference to FIG. 2.

[0015] <Flow of information processing method S1> The information processing device 1 configured as above executes an information processing method S1 according to this exemplary embodiment. The flow of the information processing method S1 will be described with reference to Fig. 2. Fig. 2 is a flow diagram showing the flow of the information processing method S1. As shown in Fig. 2, the information processing method S1 includes steps S11 to S14.

[0016] (Step S11) In step S11, the input acquisition unit 11 acquires a first character string indicating an inquiry from a user. Here, the user is a user who uses a response function that responds with an inquiry result in response to an inquiry. The information processing device 1 may have such a response function, or may operate in cooperation with a response device that has such a response function.

[0017] The user inputs the first character string using, for example, an input device. The input device used by the user may be, for example, but is not limited to, a keyboard, a mouse, a tablet, a touch panel, a touchpad, a microphone, or a combination thereof. The input device may be connected to or included in the information processing device 1, or may be connected to or included in the response device described above. The input device may be connected to or included in a terminal communicatively connected to the information processing device 1 or the response device.

[0018] (Step S12) In step S12, the generation unit 12 refers to the word database and generates a second character string by replacing at least a part of the first character string with a word in the word database. Here, the number of words included in the word database may be one or more.

[0019] (Step S13) In step S13, the evaluation acquisition unit 13 acquires evaluation information indicating a user's evaluation of the inquiry result in response to the inquiry indicated by the first character string or the second character string. Here, the "inquiry result in response to the inquiry indicated by the first character string or the second character string" refers to the inquiry result output as a response by the response function described above, with the inquiry indicated by the first character string or the second character string as input. Hereinafter, the "inquiry result in response to the inquiry indicated by the first character string or the second character string" will also be simply referred to as the "inquiry result in response to the first character string or the second character string."

[0020] For example, when an inquiry result corresponding to a first character string is output, the evaluation acquisition unit 13 acquires evaluation information for the inquiry result. Also, when an inquiry result corresponding to a second character string is output, the evaluation acquisition unit 13 acquires evaluation information for the inquiry result.

[0021] Here, the evaluation information indicates, for example, whether the evaluation is good or bad. However, the evaluation information is not limited to two levels, and may indicate three or more levels of evaluation.

[0022] The query result is output to an output device. The output device may be, for example, a display, a speaker, or a combination thereof, but is not limited to these. The output device may be connected to or included in the information processing device 1, or may be connected to or included in the response device described above. The output device may be connected to or included in a terminal communicably connected to the information processing device 1 or the response device. The user inputs the evaluation information using, for example, an input device. Specific examples of input devices used by the user are as described in step S11.

[0023] (Step S14) In step S14, the update unit 14 updates the word database based on the evaluation information. For example, if the evaluation information for the query result corresponding to the first character string indicates a "good evaluation," the update unit 14 may store a word included in the first character string in the word database. Furthermore, if the evaluation information for the query result corresponding to the first character string indicates a "bad evaluation" and a word included in the first character string is already stored in the word database, the update unit 14 may delete the word. Alternatively, if the evaluation information for the query result corresponding to the first character string indicates a "bad evaluation," the update unit 14 may store the word included in the first character string in the word database. In this case, the administrator may set which words based on evaluation information among the words stored in the word database are to be used in step S12 (for example, whether or not to use words based on a bad evaluation). A specific example of updating the word database based on evaluation information for the query result corresponding to the second character string in this step can be similarly described above by replacing the first character string with the second character string.

[0024] <Advantages of this exemplary embodiment> As described above, the information processing device 1 and information processing method S1 according to this exemplary embodiment are configured to (i) obtain a first character string indicating an inquiry from a user, (ii) refer to a word database to generate a second character string in which at least a part of the first character string is replaced with a word in the word database, (iii) obtain evaluation information indicating the user's evaluation of the inquiry result corresponding to the inquiry indicated by the first character string or the second character string, and (iv) update the word database based on the evaluation information.

[0025] In this manner, the exemplary embodiment updates the word database used to replace at least a portion of the user's query based on the user's evaluation of the query result. Using the updated word database, the user's query can be more accurately modified. As a result, the exemplary embodiment can accurately modify the user's query and improve the accuracy of the response.

[0026] Other Aspects of the Exemplary Embodiment Another aspect of this exemplary embodiment will be described with reference to FIG. 3. FIG. 3 is a block diagram showing the configuration of an information processing system 10 according to another aspect. As shown in FIG. 3, the information processing system 10 includes an input acquisition unit 11, a generation unit 12, an evaluation acquisition unit 13, and an update unit 14. The information processing system 10 includes a plurality of physically different devices, and the multiple units shown in FIG. 3 are distributed across these multiple devices. The details of the configuration and operation of each unit are as described above. As a result, the information processing system 10 achieves the same effects as the information processing device 1 described above.

[0027] Exemplary Embodiment 2 A second exemplary embodiment of the present invention will be described in detail with reference to the drawings. Note that components having the same functions as those described in the first exemplary embodiment are given the same reference numerals, and their description will be omitted as appropriate.

[0028] <Configuration of Information Processing System 10A> The configuration of an information processing system 10A according to this exemplary embodiment will be described with reference to Fig. 4. Fig. 4 is a block diagram showing the configuration of the information processing system 10A. As shown in Fig. 4, the information processing system 10A includes an information processing device 1A, a response device 2A, a document database 3A, a synonym database 4A, an evaluation information database 5A, a word database 6A, and a management device 7A.

[0029] (Information processing device 1A) The information processing device 1A is a device that corrects an inquiry input by a user to a response device 2A. As shown in Fig. 4, the information processing device 1A includes an input acquisition unit 11A, a generation unit 12A, an evaluation acquisition unit 13A, and an update unit 14A. Details of each functional block will be described later in "Flow of information processing method S1A" with reference to Fig. 8.

[0030] (Response Device 2A) The response device 2A is a device that responds to a user's inquiry with an inquiry result related to a specific field. Specifically, the response device 2A responds as the query result with a search result obtained by searching a document database 3A using a first character string or a second character string. As will be described in detail later, the document database 3A contains multiple documents related to the specific field. In other words, the query result indicates the search result obtained by searching multiple documents related to the specific field using the first character string or the second character string.

[0031] The response device 2A also acquires user evaluation information on the inquiry result. The response device 2A also modifies the inquiry from the user using the information processing device 1A and re-outputs the inquiry result according to the modification result. The response device 2A may be, for example, a device called a chatbot or an automatic response, but is not limited to these. Details of the response device 2A will be described later in "Flow of information processing method S1A" with reference to FIG. 8.

[0032] (Document Database 3A) The document database 3A is a database containing multiple documents related to a specific field. Each document contains, for example, a question and / or its answer. The document database 3A is the target of search by the response device 2A. A specific example of the document database 3A will be described with reference to FIG. 5. FIG. 5 is a diagram showing a specific example of the document database 3A. As shown in FIG. 5, the document database 3A contains documents No. 001, 002, 003, 004, ... related to the specific field "XX Society." Documents No. 001 and 002 represent answers. Documents No. 003 and 004 represent questions. The documents representing questions and documents representing answers may be linked to each other.

[0033] (Synonym Database 4A) The synonym database 4A includes sets of synonyms related to a specific field. A specific example of the synonym database 4A will be described with reference to FIG. 6. FIG. 6 is a diagram showing a specific example of the synonym database 4A. As shown in FIG. 6, the synonym database 4A includes sets of synonyms related to a specific field, "XX academic society." Specifically, for example, set No. 001 indicates that "lecture," "speech," and "seminar" are synonyms for each other. Similarly, sets No. 002 to No. 004 indicate that the illustrated words are synonyms for each other. Note that while FIG. 6 shows an example in which one synonym set includes two or three words, the number of words included in one set may be four or more.

[0034] (Evaluation Information Database 5A) The evaluation information database 5A includes evaluation information. The evaluation information indicates the user's evaluation of the inquiry result from the response device 2A. Specific examples of the evaluation information will be described later in the "Flow of information processing method S1A" section with reference to FIG. 8.

[0035] (Word Database 6A) The word database 6A includes a plurality of words and a priority of each word, and also includes some or all of (i) words included in a plurality of documents related to a specific field, (ii) a set of synonyms related to a specific field, and (iii) words included in a first character string or a second character string.

[0036] A specific example of the word database 6A will be described with reference to Fig. 7. Fig. 7 is a diagram showing a specific example of the word database 6A. In the example shown in Fig. 7, the word No. 001, "lecture," is linked to a priority of "4." Each of the words No. 002 to No. 004 is similarly linked to a priority. Details of the priority will be described later in "Flow of information processing method S1A" with reference to Fig. 8.

[0037] (Management device 7A) The management device 7A is a device used by an administrator who manages the information processing system 10A. The management device 7A presents information stored in the evaluation information database 5A to the administrator. As will be described in detail later, the evaluation information database 5A stores a first or second character string and evaluation information in association with each other. The administrator refers to the presented information and inputs information instructing updating of one or both of the document database 3A and the synonym database 4A. The management device 7A updates one or both of the document database 3A and the synonym database 4A based on the administrator's input. For example, the administrator may add a set of synonyms to the synonym database 4A by referring to a word contained in the first or second character string. The administrator may also modify an existing set of synonyms by referring to a word contained in the first or second character string. For example, the administrator may add a document indicating a query indicated by the first or second character string and its answer to the document database 3A. Furthermore, for example, the administrator may refer to the query indicated by the first character string or the second character string to correct the document included in the document database 3A.

[0038] Here, presentation to the administrator is realized, for example, by output to an output device. The output device may be, for example, a display, a speaker, or a combination thereof, but is not limited to these. The output device may be connected to or included in the management device 7A. The output device may also be connected to or included in a terminal communicatively connected to the management device 7A. The administrator inputs information indicating the above-mentioned instruction to the management device 7A using an input device. The input device may be, for example, a keyboard, a mouse, a tablet, a touch panel, a touchpad, a microphone, or a combination thereof, but is not limited to these. The input device may also be connected to or included in the management device 7A. The input device may also be connected to or included in a terminal communicatively connected to the management device 7A.

[0039] The management device 7A updates the document database 3A and the synonym database 4A at any timing.

[0040] <Flow of information processing method S1A> The information processing system 10A configured as above executes an information processing method S1A according to this exemplary embodiment. The flow of the information processing method S1A will be described with reference to Fig. 8. Fig. 8 is a flow diagram showing the flow of the information processing method S1A. As shown in Fig. 8, the information processing method S1A includes steps S21 to S28 executed by the response device 2A and steps S31 to S35 executed by the information processing device 1A.

[0041] (Step S21) In step S21, the response device 2A acquires a first character string input by the user that indicates an inquiry from the user, and transmits the acquired first character string to the information processing device 1A. As a result, the input acquisition unit 11A of the information processing device 1A acquires the first character string. As a specific example, the acquired first character string is, for example, "I would like to know the cost of the speech." The input device used by the user to input the first character string is as described in the first embodiment.

[0042] (Step S22) In step S22, response device 2A uses the first character string to search for multiple documents contained in document database 3A. In the example shown in Fig. 5, document database 3A does not store any documents related to the first character string "I would like to know the cost of a speech." Therefore, in this case, the search result is 0 (the number of documents contained in the search result is zero).

[0043] (Step S23) In step S23, the response device 2A responds with the search result in step S22 as the query result. For example, the response device 2A outputs the query result and information requesting input of evaluation information for the query result to the output device.

[0044] For example, the response device 2A may display the query result and a user interface component for selecting either "Like" or "Dislike" on a display. The response device 2A also transmits the query result to the information processing device 1A.

[0045] (Step S24) In step S24, the response device 2A acquires evaluation information in response to the inquiry result sent in step S23, and transmits the acquired evaluation information to the information processing device 1A. However, it is assumed that the user may or may not input evaluation information. That is, the response device 2A may or may not be able to acquire evaluation information. If the user's input of evaluation information is not obtained, the response device 2A transmits information indicating that evaluation information has not been obtained to the information processing device 1A. Hereinafter, "information indicating that evaluation information has not been obtained" will also be referred to as evaluation information indicating "no evaluation." As a result, the evaluation acquisition unit 13A of the information processing device 1A acquires evaluation information in response to the inquiry result corresponding to the first character string. The input device used by the user to input evaluation information is as described in the first embodiment.

[0046] (Step S31) In step S31, the evaluation acquisition unit 13A of the information processing device 1A associates the acquired evaluation information with the first character string acquired by the input acquisition unit 11A and stores the information in the evaluation information database 5A. Note that the processing of this step may be executed by the response device 2A.

[0047] A specific example of the evaluation information database 5A will now be described with reference to FIG. 9. FIG. 9 is a diagram showing a specific example of the evaluation information database 5A. In the example shown in FIG. 9, evaluation information No. 001, "Bad," is linked to the first character string, "I would like to know the cost of the speech." This indicates that the user's evaluation of the query results (0 search results) in response to the first character string, "I would like to know the cost of the speech," was bad.

[0048] Furthermore, the evaluation information No. 002 in FIG. 9, "No evaluation", is linked to the first character string "I want to know about the lecture." Furthermore, the evaluation information No. 003 in FIG. 9, "No evaluation", is linked to the first character string "Is there a support group?" These indicate that the user did not input evaluation information into the response device 2A in response to the inquiry result corresponding to the first character string. The evaluation information No. 004 will be explained in step S28 below.

[0049] (Step S32) In step S32, the generation unit 12A of the information processing device 1A determines whether to generate a second character string. For example, the generation unit 12A may determine to generate a second character string when the amount of information included in the query result corresponding to the query indicated by the first character string is equal to or less than a threshold. The amount of information included in the query result is, for example, the number of documents included in the query result. The threshold is, for example, zero, but is not limited to this. Furthermore, for example, the generation unit 12A may determine to generate a second character string when evaluation information for the query result corresponding to the query indicated by the first character string satisfies a predetermined condition. Examples of evaluation information satisfying the predetermined condition include, but are not limited to, being "bad" or no evaluation information being obtained. Note that the processing of this step may be executed by the response device 2A. In this case, the information processing device 1A receives the determination result from the response device 2A.

[0050] (Step S33) In step S33, the generation unit 12A refers to the word database 6A and generates a second character string by replacing at least a part of the first character string with a word in the word database 6A. The generation unit 12A also transmits the second character string to the response device 2A. Here, it is assumed that the second character string "I would like to know the cost of the lecture" has been generated from the first character string "I would like to know the cost of the speech." Details of the processing of this step will be described later with reference to another drawing.

[0051] (Step S25) In step S25, the response device 2A determines whether or not the second character string has been received from the information processing device 1A. If the determination in step S25 is No, the operation of the response device 2A ends. If the determination in step S25 is Yes, the response device 2A executes the process of step S26.

[0052] (Step S26) In step S26, response device 2A uses the second character string to search for multiple documents contained in document database 3A. In the example shown in Fig. 5, document database 3A stores document No. 002 related to the second character string "I would like to know the cost of the lecture." Therefore, in this case, document No. 002 is included in the search results.

[0053] (Step S27) In step S27, the response device 2A responds with the search result in step S26 as the query result. At this time, the response device 2A outputs the query result and information requesting input of evaluation information for the query result to the output device. Specific examples of the output contents are as described in step S23.

[0054] (Step S28) In step S28, the response device 2A acquires evaluation information for the inquiry result responded in step S27, and transmits the acquired evaluation information to the information processing device 1A. As in step S24, it is assumed that the user may input evaluation information voluntarily. Specific examples of input of evaluation information are as described in step S24. If the user does not input evaluation information, the response device 2A transmits information indicating that evaluation information has not been acquired (evaluation information indicating "no evaluation") to the information processing device 1A. As a result, the evaluation acquisition unit 13A of the information processing device 1A acquires evaluation information for the inquiry result corresponding to the second character string.

[0055] (Step S34) In step S34, the evaluation acquisition unit 13A of the information processing device 1A associates the acquired evaluation information with the second character string generated by the generation unit 12A and stores the information in the evaluation information database 5A. Note that the processing of this step may be executed by the response device 2A.

[0056] A specific example of information that is added to and stored in the evaluation information database 5A will now be described with reference to Fig. 9. In the example shown in Fig. 9, the evaluation information No. 004, "Like," is linked to the second character string, "I would like to know the cost of the lecture." This indicates that the user's evaluation of the inquiry result (one search result) in response to the second character string, "I would like to know the cost of the lecture," was good.

[0057] (Step S35) In step S35, the update unit 14A refers to the evaluation information database 5A and updates the word database 6A. Details of the processing in this step will be described later with reference to another drawing.

[0058] <Details of the generation process (step S33)> The details of the generation process in step S33 described above will be described with reference to Fig. 10. Fig. 10 is a flow diagram showing the flow of the generation process in step S33. As shown in Fig. 10, the generation process in step S33 includes steps S331 to S333.

[0059] (Step S331) In step S331, the generation unit 12A extracts words related to words included in the first character string from the word database 6A. The number of words to be extracted may be one or more. In other words, the generation unit 12A identifies, as correction parts, words that are related to words in the word database 6A among the words that make up the first character string. Here, specific examples 1 and 2 of the process of extracting related words will be described.

[0060] (Example 1) For example, the generation unit 12A may refer to the synonym database 4A and extract synonyms for the words included in the first character string from the word database 6A. For example, referring to the set No. 001 in the synonym database 4A shown in FIG. 6, synonyms for the word "speech" included in the first character string "I would like to know the cost of the speech" are the word "lecture" and the word "seminar." Therefore, the generation unit 12A extracts the word "lecture" and the word "seminar" from the word database 6A. In this case, the word "speech" is the part to be corrected in the first character string.

[0061] (Example 2) Note that words in the word database 6A that are related to words included in the first character string are not limited to synonyms. For example, the generation unit 12A may extract words in the word database 6A that have similar spellings to words included in the first character string. For example, if the first character string includes the word "speach" and the word database 6A includes the word "speech," the words "speach" and "speech" have similar spellings. Such spelling similarity includes, for example, a case where a word included in the first character string is misspelled with respect to a word in the word database 6A. Therefore, in this case, the generation unit 12A may extract the word "speech" from the word database 6A. Note that known techniques can be used for the process of determining the similarity in spelling. Note that the relevance between words is not limited to synonyms in Specific Example 1 or spelling similarity in Specific Example 2, but may also be other relevances.

[0062] (Step S332) In step S332, the generation unit 12A selects one of the multiple words extracted from the word database 6A based on the priority. Specifically, the generation unit 12A may select the word with the highest priority from the multiple words extracted from the word database 6A. Note that if the number of extracted words is one, the generation unit 12A is deemed to have selected the extracted word as is.

[0063] (Step S333) In step S333, the generation unit 12A replaces the word identified as the correction part in the first character string (i.e., at least a part of the first character string) with the word selected in step S332. As a result, the generation unit 12A sets the character string after the replacement as the second character string.

[0064] <Details of Update Process (Step S35)> The update process in step S35 described above will be described in detail with reference to Fig. 11. Fig. 11 is a flow diagram showing the flow of the update process in step S35. As shown in Fig. 11, the update process in step S35 includes steps S351 to S355.

[0065] (Step S351) In step S351, the update unit 14A extracts words from multiple documents included in the document database 3A. For example, if the documents are written in a language (such as English) that includes word delimiters (such as spaces), the update unit 14A extracts words based on the delimiters. Alternatively, if the documents are written in a language (such as Japanese) that does not include word delimiters, the update unit 14A extracts words using a known technique for extracting words from sentences in such a language. As a specific example, the update unit 14A extracts words such as "lecture," "apply," "below," and "inquiry" from document No. 001 shown in FIG. 5. Note that the update unit 14A may extract words from documents that indicate questions among the multiple documents included in the document database 3A, but not from documents that indicate answers.

[0066] The document database 3A may be updated by the management device 7A at any time. If the document database 3A has not been updated between the previous execution of the update process and the current execution of the update process, the update unit 14A may omit the process of step S351. Alternatively, the update unit 14A may execute the process of step S351 only for the part of the document database 3A that has been updated between the previous execution of the update process and the current execution of the update process.

[0067] (Step S352) In step S352, the update unit 14A extracts words from the sets of synonyms included in the synonym database 4 A. As a specific example, the update unit 14A extracts words such as “lecture,” “speech,” and “seminar” from the set of synonyms No. 001 shown in FIG.

[0068] The synonym database 4A can be updated by the management device 7A at any timing. If the synonym database 4A has not been updated between the previous execution of the update process and the current execution of the update process, the update unit 14A may omit the process of step S352. Furthermore, the update unit 14A may execute the process of step S352 only for the part of the synonym database 4A that has been updated between the previous execution of the update process and the current execution of the update process.

[0069] (Step S353) In step S353, the update unit 14A refers to the evaluation information database 5A and extracts words from the first character string or the second character string. In the example shown in Fig. 9, the update unit 14A extracts words such as "speech," "cost," and "know" from the first character string No. 001, "I would like to know the cost of the speech."

[0070] (Step S354) In step S354, the update unit 14A extracts words that satisfy a condition from among the words extracted in steps S351 to S353. The condition is, for example, "verb or noun," but is not limited to this.

[0071] (Step S355) In step S355, the update unit 14A updates the priority of each word extracted in step S354 based on the evaluation information.

[0072] Here, the update unit 14A updates the priority of each word based on the evaluation information associated with that word. Here, the words extracted in step S354 can be considered to be associated with evaluation information. Specifically, a word extracted from the first or second character string is associated with evaluation information associated with the first or second character string that contained that word. Furthermore, a word extracted from the document database 3A or the synonym database 4A is considered to be equivalent to no evaluation information being obtained. In other words, a word extracted from the document database 3A or the synonym database 4A is associated with evaluation information "no evaluation."

[0073] More specifically, the update unit 14A assigns evaluation points to each word extracted in step S354 in the order of good evaluation (good) indicated by the evaluation information, no evaluation information available, and bad evaluation (bad) indicated by the evaluation information. The update unit 14A also updates the priority of each word in the word database 6A based on the evaluation points. The evaluation points are an example of the "weighting coefficient" described in the claims. Details of the evaluation points will be explained below.

[0074] Here, the evaluation information "Like" indicates an active intention by the user. In this case, it is considered that the response device 2A is able to respond with the inquiry result that the user expects. Therefore, it is considered that the word associated with the evaluation information "Like" is highly important as a word to be used to improve the accuracy of the response. Therefore, it is desirable that the evaluation score of the word is high.

[0075] Furthermore, the evaluation information "bad" indicates the user's active intention. In this case, it is considered that the response device 2A is unable to respond with the inquiry result the user expected. In such a case, the inquiry itself may be in a different field from the specific field assumed by the response device 2A. For example, a response device 2A assuming a specific field, "XX Society," is likely to not obtain the inquiry result the user expected in response to an inquiry about a park. Furthermore, even if the inquiry itself is in the specific field, the inquiry result may not be what the user expected. For example, if an inquiry about the date of a lecture results in an inquiry about lecture tickets, the inquiry result is not what the user expected. Therefore, words associated with such evaluation information are considered to be less important as words to use to improve the accuracy of the response. Therefore, it is desirable to give the word a low evaluation score.

[0076] Furthermore, when the evaluation information "no evaluation" is obtained, that is, when the user's active intention is not obtained, it is considered that this includes a case where the response device 2A responds with the inquiry result the user expects, but the user does not enter an evaluation. Therefore, it is considered that the word associated with the evaluation information "no evaluation" may be highly important as a word to be used to improve the accuracy of the response. Therefore, it is desirable that the evaluation score of the word be lower than when the evaluation information is "good" and higher than when the evaluation information is "bad."

[0077] A specific example of the evaluation score will be described with reference to Fig. 12. Fig. 12 is a diagram for explaining a specific example of the evaluation score. In the example shown in Fig. 12, when the evaluation information is "good", an evaluation score of 2 is determined, when it is "bad", an evaluation score of 0 is determined, and when it is "no evaluation", an evaluation score of 1 is determined.

[0078] Furthermore, when multiple instances of the same word are extracted in step S354, the updating unit 14A may set the priority to a value obtained by adding up the evaluation points determined for each word. Furthermore, when the word is already stored in the word database 6A, the value obtained by adding up the evaluation points may be used as the priority after the update, replacing the priority before the update. In other words, the updating unit 14A may overwrite the priority before the update with the value obtained by adding up the evaluation points.

[0079] As a specific example, the update unit 14A determines an evaluation point of 1 for each of the words "lecture" extracted from documents No. 001, 002, and 003 shown in FIG. 5, assuming that the word is associated with the evaluation information "no evaluation." The update unit 14A also determines an evaluation point of 1 for the word "lecture" extracted from the synonym set No. 001 shown in FIG. 6, assuming that the word is associated with the evaluation information "no evaluation." The update unit 14A also determines an evaluation point of 1 for the word "lecture" extracted from the first character string No. 002 shown in FIG. 9, assuming that the word is associated with the evaluation information "no evaluation." The update unit 14A also determines an evaluation point of 2 for the word "lecture" extracted from the second character string No. 004 shown in FIG. 9, assuming that the word is associated with the evaluation information "good." The update unit 14A then calculates a total of 7, which is the sum of the evaluation points 1, 1, 1, 1, 1, and 2 determined for the six words "lecture." Furthermore, the update unit 14A updates the word database 6A by overwriting the pre-update priority of 4 for the word No. 001 "lecture" shown in FIG. 7 with the updated priority of 7 in total evaluation points. Note that if the update unit 14A calculates a priority for an unregistered word that is not saved in the word database 6A, it adds the word to the word database 6A. However, if the calculated priority (total evaluation points) for an unregistered word is zero, the update unit 14A does not have to add the word to the word database 6A.

[0080] Note that the specific evaluation score values ​​and the method for updating the priority based on the evaluation score are not limited to the values ​​and methods described above.

[0081] <Advantages of this exemplary embodiment> As described above, in this exemplary embodiment, the following configuration is adopted: (i) the query result indicates the search results obtained by searching multiple documents related to a specific field contained in document database 3A using the first character string or the second character string; and (ii) word database 6A includes some or all of the words contained in the multiple documents, sets of synonyms related to the specific field, and words contained in the first character string or the second character string.

[0082] Here, words contained in multiple documents stored in document database 3A are suitable for queries targeting those documents. Also, a set of synonyms related to a specific field stored in synonym database 4A is suitable for queries targeting that specific field. Also, a first character string entered by a user or a second character string obtained by correcting the first character string may contain words that appropriately reflect the intent of the user's query. And, because these words are updated based on evaluation information from users, they become more suitable for queries targeting a specific field.

[0083] Therefore, according to the above configuration, the exemplary embodiment can modify a user's inquiry into a character string suitable for an inquiry related to a specific field, which increases the possibility that the response device 2A can retrieve an appropriate document, thereby improving the accuracy of the response.

[0084] In addition, in this exemplary embodiment, the following configuration is adopted: (i) the word database 6A includes a plurality of words and the priority of each word; (ii) the generation unit 12A replaces at least a portion of the first string with a word selected from the plurality of words in the word database 6A based on the priority; and (iii) the update unit 14A updates the priority of each word in the word database 6A based on the evaluation information.

[0085] According to the above configuration, appropriate words selected from among words contained in the first character string entered by the user or the second character string obtained by correcting the first character string according to the user's evaluation of the query result are used when generating the second character string, thereby improving the accuracy of the response.

[0086] In addition, in this exemplary embodiment, the update unit 14A is configured to (i) assign increasing evaluation points to each word in the word database 6A in the order of good evaluation indicated by the evaluation information, no evaluation information available, and bad evaluation indicated by the evaluation information, and (ii) update the priority of each word in the word database 6A based on the evaluation points.

[0087] According to the above configuration, if the user gives a good evaluation to the query result corresponding to the first character string input by the user or the second character string obtained by correcting the first character string, the priority is updated by placing importance on the words included in the first character string or the second character string, thereby enabling the generation of a more accurate second character string.

[0088] Furthermore, with the above configuration, when a user rating is not available, the words included in the first or second character string can be given the second highest importance after a good rating, and the priority can be updated. This is because when a rating is not available, it does not necessarily mean that the user rating is bad, but may also be good. Therefore, there is a possibility that the first or second character string contains words that should be given high importance.

[0089] Furthermore, with the above configuration, if the user's evaluation is poor, the priority can be updated so that the words included in the first character string or the second character string are not considered important. This is because if the evaluation is poor, it is highly likely that the content of the user's inquiry is not related to the specific field targeted by the response device 2A, or the inquiry result is not what the user expected.

[0090] Furthermore, in this exemplary embodiment, a configuration is adopted in which a second character string is generated when the amount of information included in the query result corresponding to the first character string is equal to or less than a threshold.

[0091] Here, if the amount of information (number of documents) included in the query result corresponding to the first character string is equal to or less than a threshold (for example, if the search result is zero), it is considered that the user's satisfaction is low. According to the above configuration, the second character string is generated in such a case, and the response device 2A can respond to the query result again using the generated second character string, thereby improving the user's satisfaction.

[0092] In the present exemplary embodiment, the generation unit 12A may generate a second character string when evaluation information on the inquiry result corresponding to the first character string satisfies a predetermined condition. According to the above configuration, the second character string is generated in such a case, and therefore the response device 2A can respond to the inquiry result again using the generated second character string, thereby improving the satisfaction of the user.

[0093] Exemplary Embodiment 3 A third exemplary embodiment of the present invention will be described in detail with reference to the drawings. Components having the same functions as those described in the first exemplary embodiment are denoted by the same reference numerals, and their description will be omitted as appropriate. In this exemplary embodiment, a first character string is input to a response device using an input language that involves input conversion.

[0094] <Input language with input conversion> First, we will explain the details of languages ​​that involve input conversion, to which this exemplary embodiment is applicable. Examples of such input languages ​​include "Japanese" and "Chinese." Here, input conversion refers to "converting" information "input" by a user into the type of characters used in that language. Input conversion requires various user operations. Therefore, if the user's operation during input conversion differs from the original intention, the first character string input to the response device may contain errors. Even in such cases, this exemplary embodiment can accurately correct errors in the first character string to improve the accuracy of the response.

[0095] A specific example of input conversion in Japanese will be described with reference to FIG. 13. Japanese is written using multiple types of characters, such as "kanji," "hiragana," and "katakana." "Kanji" is called an ideogram, and each character has a specific meaning. "Hiragana" and "katakana" are called phonetic characters, and each character has no meaning and represents only sound. Therefore, Japanese pronunciation is expressed in hiragana or katakana. Here, the notation of Japanese pronunciation using hiragana is also referred to as "hiragana notation." The notation of Japanese words using kanji is also referred to as "kanji notation." The notation of Japanese sentences that include some or all of "kanji," "hiragana," and "katakana" is also referred to as "Japanese notation."

[0096] Furthermore, in Japanese, "romaji" is defined, which corresponds a string of alphabetic characters to the sounds of Japanese. When inputting Japanese using a keyboard with alphabetic keys, the "romaji input method" is often used, in which Japanese pronunciation is input using these "romaji." Figure 13 is a diagram explaining a specific example of input conversion of "Japanese" using the romaji input method.

[0097] 13, G1 to G5 indicate input areas of an application on the display, and U1 to U5 indicate a series of operations in input conversion. Here, an example will be described in which a word w2 written in kanji and meaning "lecture" is input.

[0098] Input area G1 shows a display state in which no characters have been input into the input area. In this state, the user first performs operation U1. Operation U1 is an operation of pressing a switching key (for example, the "Esc" key) to switch to "Japanese input mode."

[0099] Next, the user performs operation U2. Operation U2 is an operation of pressing each key of the alphabet string "kouen" in order. Here, the alphabet string "kouen" is the "Roman letters" that correspond to the reading of the Japanese word w2. As a result, the display in input area G1 changes to look like input area G2. In input area G2, the "hiragana notation" w0 that indicates the reading of word w2 is displayed.

[0100] Next, the user performs operation U3. Operation U3 is an operation of pressing the "conversion key" (for example, assigned to the space key) and converting the "reading" displayed on the screen into "kanji" or the like. As a result, the display in input area G2 changes to that shown in input area G3. In input area G3, multiple words w1 to w3 written in kanji are displayed. Here, in Japanese, there are often multiple words with different meanings that correspond to the same "reading." In this case, these words with different meanings also have different "kanji" writings. Therefore, operation U3 often displays multiple different words w1 to w3 as candidates. Words w1 to w3 all have the same reading in Japanese.

[0101] Next, the user performs operation U4. Operation U4 is an operation of pressing a selection key (for example, assigned to an arrow key) as many times as necessary to select one of the multiple words w1 to w3. As a result, the display in input area G3 changes to look like input area G4. In input area G4, a figure F1 is displayed, indicating that word w2 is being selected.

[0102] Next, the user performs operation U5. Operation U5 is an operation of pressing the "confirm key" (for example, assigned to the Enter key) to confirm the selection. As a result, the display of input area G4 changes to look like input area G5. That is, word w2, which has been confirmed as input, is displayed in input area G5.

[0103] 13, a series of operations U1 to U5 are required to input the word w2 written in kanji and meaning "lecture" through input conversion. Common operational errors in this type of input conversion include: (i) performing operations U2 to U5 without performing operation U1, resulting in the alphabetic string representing the reading of the word you want to input in romanized characters being confirmed as the input; (ii) performing operation U5 without performing operations U2 to U4 after operation U1, resulting in the hiragana notation being confirmed as the input; and (iii) performing an incorrect selection operation in operation U4, resulting in a word with the same reading but a different meaning (i.e., a homonym) being confirmed as the input.

[0104] The present exemplary embodiment improves response accuracy even when queries to answering machines contain language-specific errors involving input transformations such as those described above.

[0105] <Configuration of Information Processing System 10B> The configuration of an information processing system 10B according to this exemplary embodiment will be described with reference to Fig. 14. Fig. 14 is a block diagram showing the configuration of the information processing system 10B. As shown in Fig. 14, the information processing system 10B includes an information processing device 1B, a response device 2B, a document database 3B, a synonym database 4B, an evaluation information database 5B, a word database 6B, and a management device 7B.

[0106] The response device 2B, document database 3B, synonym database 4B, evaluation information database 5B, and management device 7B are configured in substantially the same manner as the devices of the same names in exemplary embodiment 2. However, the response device 2B accepts input of a first character string using a language involving input conversion and outputs information in that language. In other words, the first character string is input by a user using an input language involving input conversion. The management device 7B accepts input in a language involving input conversion and outputs information in that language. Furthermore, the document database 3B, synonym database 4B, and evaluation information database 5B store information expressed in a language involving input conversion. As other aspects of each of these devices are as described in exemplary embodiment 2, detailed description thereof will not be repeated.

[0107] (Information processing device 1B) The information processing device 1B includes an input acquisition unit 11B, a generation unit 12B, an evaluation acquisition unit 13B, an update unit 14B, and an identification unit 15B. The identification unit 15B is an example of a configuration that realizes the identification means described in the claims. The input acquisition unit 11B and the evaluation acquisition unit 13B are configured in substantially the same manner as the units of the same names in the second exemplary embodiment. The generation unit 12B, the update unit 14B, and the identification unit 15B will be described in detail below.

[0108] (Word Database 6B) The word database 6B contains almost the same information as the word database 6A in the second exemplary embodiment. However, the word database 6B differs from the word database 6A in that the words are written in Japanese and that the "reading" of the words is linked to the words in addition to the priority. A specific example of the word database 6B will be described with reference to FIG. 15. FIG. 15 is a diagram showing a specific example of the word database 6B. In the example shown in FIG. 15, the information No. 001 is information linking the Japanese word "lecture" written in kanji with the Japanese reading and the priority. The information Nos. 002 to 004 will be similarly described.

[0109] <Details of the generation process (step S33)> 8 in substantially the same manner as the information processing system 10A according to the exemplary embodiment 2. However, the details of the process of step S33 differ from the process of step S33 according to the exemplary embodiment 2.

[0110] Details of the process of step S33 in this exemplary embodiment will be described with reference to Fig. 16. Fig. 16 is a flow diagram showing the flow of the generation process in step S33. As shown in Fig. 16, the process of step S33 includes steps S331B to S334B.

[0111] (Step S331B) In step S331B, the identification unit 15B identifies a third character string before input conversion for the first character string. Identification of the third character string will be described with reference to Fig. 17. Fig. 17 is a diagram illustrating a specific example of the generation process in step S33 in this exemplary embodiment.

[0112] 17, the character string str0 is a sentence written in Japanese meaning "I want to know about the lecture information," and indicates the character string that the user originally intended to input in step S21. Here, it is assumed that there was an error in inputting the character string str0, and therefore the first character string str1 or the first character string str2 was input.

[0113] Specifically, the first character string str1 is determined as an input of the "Romanized notation" corresponding to the reading of the character string str0, as described with reference to Fig. 13. The identification unit 15B interprets such a first character string str1 as the "Romanized notation" and thereby identifies the third character string str3 in hiragana notation, which is the corresponding reading in Japanese.

[0114] 13, the first character string str2 is determined as a part of the character string str0 in which a homonym different from the original intention has been selected. In detail, the kanji character string str2a (meaning "park") included in the character string str2 is a homonym of the kanji character string str0a included in the character string str0. The identification unit 15B identifies the third character string str3 in hiragana, which is the "reading" of the entire sentence, for the first character string str2 in Japanese.

[0115] (Step S332B) In step S332B, the generation unit 12B extracts, from the word database 6B, words that can be input converted from at least a part of the third character string. Specifically, for example, the generation unit 12B extracts, from the word database 6B, one or more words that are linked to character strings included in the third character string as their "readings." Each of these words can be said to be a word that can be input converted from at least a part of the third character string. For example, in the word database 6B shown in FIG. 15, the word No. 001 "lecture" and the word No. 003 "sponsor" are linked to str3a, which is part of the third character string str3 shown in FIG. 17, as their "readings." Therefore, the generation unit 12B extracts, from the word database 6B, the words "lecture" and "sponsor" that can be input converted from the character string str3a included in the third character string str3.

[0116] (Step S333B) The process of step S333B is executed when multiple words are extracted in step S332B. In step S333B, the generation unit 12B selects one of the multiple words based on priority. For example, the generation unit 12B may select the word with the highest priority from the multiple words. Note that, when only one word is extracted in step S332B, the generation unit 12B is deemed to have selected that word. In the example shown in FIG. 15, between the word No. 001 "lecture" (priority 3) and the word No. 003 "sponsorship" (priority 1), the word No. 001 "lecture" is selected as it has the higher priority.

[0117] (Step S334B) In step S334B, the generation unit 12B replaces at least a part of the first character string with the word selected in step S333B. Specifically, the generation unit 12B sets the portion of the first character string corresponding to the word extracted in step S333B as the correction portion. The generation unit 12B then generates, as the second character string, a character string in which the correction portion in the first character string has been replaced with the word selected in step S333B. In the example shown in FIG. 17, the generation unit 12B generates, as the second character string str4, the character string "I want to know about the lecture" in which the correction portion str2a in the first character string str2 has been replaced with the word "lecture."

[0118] <Specific examples of evaluation information database 5B> A specific example of the evaluation information database 5B in this exemplary embodiment will be described with reference to FIG.

[0119] Fig. 18 is a diagram showing a specific example of the evaluation information database 5B in this exemplary embodiment. In the example shown in Fig. 18, the evaluation information No. 1, "Bad," is linked to the first character string str1. The evaluation information No. 2, "Bad," is linked to the first character string str2. The first character strings str1 and str2 are as described with reference to Fig. 17, and were each entered due to an operational error when entering the originally intended character string str0.

[0120] The response device 2B searches the document database 3B but is unable to find any documents related to the first character string str1 or str2 that contain the error, and so responds with a query result that the search result is 0. The evaluation information for No. 001 and No. 002 indicates that users have given these items poor evaluations.

[0121] Furthermore, the evaluation information of No. 003, "Like," is linked to the second character string str4. As explained with reference to FIG. 17, the second character string str4 was generated by correcting the first character string str2, which contained an error. The response device 2B searches the document database 3B for documents No. 001 to No. 003 related to the second character string str4, and responds as the query result. The evaluation information of No. 003 indicates that the user's evaluation of it was good.

[0122] Furthermore, the evaluation information for No. 004, "No evaluation," is linked to the first character string str5. The first character string str5 is a Japanese expression meaning "Is there a support group?" and is a character string entered by the user. The response device 2B searches the document database 3B for the document No. 004 related to the first character string str5 and responds as the query result. The evaluation information for No. 004 indicates that no evaluation was obtained from the user for this.

[0123] <Specific example of update process (step S35)> A specific example of the update process in step S35 in this exemplary embodiment will be described with reference to Fig. 19. Fig. 19 is a diagram showing a specific example of the update process in this exemplary embodiment.

[0124] The update unit 14B extracts words from the first character string or the second character string (step S353 in FIG. 11). As a specific example, the update unit 14B extracts three words, namely, a word w41 (meaning "lecture"), a word w42 (meaning "information"), and a word w43 (meaning "know") written in kanji from the second character string str4 shown in FIG. 18. The update unit 14B also extracts three words, namely, a word w51 (meaning "sponsor"), a word w52 (meaning "meeting"), and a word w53 (meaning "be"), written in kanji from the first character string str5 shown in FIG. 18. Here, it is assumed that the extracted words w41 to w43 and w51 to w53 all satisfy the conditions (step S354 in FIG. 11).

[0125] As shown in FIG. 19, the update unit 14B determines, for each of the words w41 to w43 extracted from the second character string str4, an evaluation score of "2" corresponding to the evaluation information "Like" associated with the second character string str4. Furthermore, because the word w41 corresponds to word No. 001 shown in FIG. 15, the update unit 14B adds an evaluation score of "2" to the pre-update priority of "3" of the word, thereby updating the priority to "5." Furthermore, for example, if the evaluation score of the word w41 is subsequently increased by "2" as a result of a further update of the evaluation information database 5B, the update unit 14B adds the increased evaluation score of "2" to the previously updated priority of "5" of the word, thereby updating the priority to "7."

[0126] 19, the update unit 14B updates each of the words w51 to w53 extracted from the first character string str5. 1 The update unit 14B determines an evaluation score of "1" corresponding to the evaluation information "no evaluation" linked to the character string str5. Because word w51 corresponds to word No. 003 shown in FIG. 15, the update unit 14B adds an evaluation score of "1" to the pre-update priority of this word, "1," to update the priority to "2." The update unit 14B also updates the priorities of the other extracted words w42, w43, w52, and w53 based on the evaluation scores in a similar manner.

[0127] <Advantages of this exemplary embodiment> In this exemplary embodiment, the following configuration is adopted: (i) a first character string is input to a response device 2B using an input language that involves input conversion; (ii) an identification unit 15B identifies a third character string before input conversion for the first character string; and (iii) a generation unit 12B extracts a word that can be input converted from at least a part of the third character string from a word database 6B, and replaces at least a part of the first character string with the extracted word.

[0128] Here, an input language involving input conversion requires various operations associated with the input conversion, which may result in the user performing an operation different from his or her original intention. According to the above configuration, the second character string is generated by reconverting at least a portion of the third character string before the user performs the input conversion into a word in the word database 6B. Therefore, even if the user performs an operation different from his or her original intention, a more appropriate second character string can be generated to obtain a query result. As a result, when a query is input using an input language involving input conversion, the accuracy of the response by the response device 2B can be improved.

[0129] [Modification] In the above-described exemplary embodiments 2 and 3, step S33 shown in FIG. 8 can be modified as follows.

[0130] In step S33, the generation units 12A and 12B may generate multiple candidates for the second character string and select one of the multiple candidates based on a user operation to serve as the second character string. For example, in step S332 of FIG. 10 or step S333B of FIG. 16, the generation units 12A and 12B select multiple words based on priority from the multiple words extracted from the word databases 6A and 6B. For example, the multiple selected words may be a predetermined number (2 or more) in descending order of priority, or may be words with a priority equal to or greater than a threshold. In addition, in step S333 of FIG. 10 or step S334B of FIG. 16, the generation units 12A and 12B generate a candidate for the second character string by replacing the corrected portion of the first character string with the selected word for each of the selected multiple words. In addition, the generation units 12A and 12B present the multiple candidates to the user by transmitting them to the response devices 2A and 2B. Then, the generation units 12A and 12B may select one of the candidates as the second character string based on the user's operation on the response units 2A and 2B.

[0131] <Effects of this modified example> In this modification, the generation units 12A and 12B generate a plurality of candidates that can become the second character string, and select one of the plurality of candidates based on the user's operation as the second character string. According to the above configuration, the first character string can be corrected to generate a second character string that suits the user's intention, and the user's satisfaction with the response of the response units 2A and 2B can be improved.

[0132] [Other Modifications] In the above-described exemplary embodiments 2 and 3, when the amount of information included in the search results for the first character string is equal to or less than a threshold, the response devices 2A and 2B may output the query results for the second character string instead of outputting the query results for the first character string. In this case, the information processing devices 1A and 1B do not acquire evaluation information corresponding to the first character string, but acquire evaluation information corresponding to the second character string.

[0133] In the above-described exemplary embodiments 2 and 3, the response devices 2A and 2B may include at least some of the functional blocks included in the information processing devices 1A and 1B, and may execute at least some of the processes executed by the information processing devices 1A and 1B. In addition, the information processing devices 1A and 1B may include at least some of the functional blocks included in the response devices 2A and 2B, and Responders 2A and 2B It may perform at least part of the processing performed by

[0134] Furthermore, the above-described exemplary embodiment 3 has been described as an example in which the language involving input conversion is Japanese, but the same effect can be achieved even when applied to languages ​​involving input conversion other than Japanese.

[0135] [Software implementation example] Some or all of the functions of the information processing devices 1, 1A, and 1B may be realized by hardware such as an integrated circuit (IC chip), or may be realized by software.

[0136] In the latter case, the information processing devices 1, 1A, and 1B are realized, for example, by a computer that executes instructions of a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as computer C) is shown in FIG. 20. The computer C includes at least one processor C1 and at least one memory C2. The memory C2 stores a program P for operating the computer C as the information processing devices 1, 1A, and 1B. In the computer C, the processor C1 reads and executes the program P from the memory C2, thereby realizing each function of the information processing devices 1, 1A, and 1B.

[0137] The processor C1 may be, for example, a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), a micro processing unit (MPU), a floating point number processing unit (FPU), a physics processing unit (PPU), a microcontroller, or a combination thereof. The memory C2 may be, for example, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or a combination thereof.

[0138] The computer C may further include a RAM (Random Access Memory) for expanding the program P during execution and for temporarily storing various data. The computer C may also include a communication interface for transmitting and receiving data to and from other devices. The computer C may also include an input / output interface for connecting input / output devices such as a keyboard, mouse, display, and printer.

[0139] The program P can also be recorded on a non-transitory, tangible recording medium M that can be read by the computer C. Such a recording medium M can be, for example, a tape, a disk, a card, a semiconductor memory, or a programmable logic circuit. The computer C can acquire the program P via such a recording medium M. The program P can also be transmitted via a transmission medium. Such a transmission medium can be, for example, a communications network or broadcast waves. The computer C can also acquire the program P via such a transmission medium.

[0140] [Appendix 1] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. For example, embodiments obtained by appropriately combining the technical means disclosed in the above-described embodiments are also included in the technical scope of the present invention.

[0141] [Appendix 2] Some or all of the above-described embodiments can also be described as follows, but the present invention is not limited to the following described aspects.

[0142] (Appendix 1) an input acquisition means for acquiring a first character string indicating an inquiry from a user; a generating means for generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; evaluation acquisition means for acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; an update means for updating the word database based on the evaluation information; An information processing device comprising:

[0143] According to the above configuration, the word database used to replace at least a part of a user's inquiry is updated based on the user's evaluation of the inquiry result. By using the updated word database in this manner, the user's inquiry can be corrected more accurately. As a result, the user's inquiry can be corrected more accurately, and the accuracy of the response can be improved.

[0144] (Appendix 2) the query result indicates a search result obtained by searching a plurality of documents related to a specific field using the first character string or the second character string; The information processing device described in Appendix 1, wherein the word database includes some or all of the words contained in the plurality of documents, a set of synonyms related to the specific field, and words contained in the first string or the second string.

[0145] According to the above configuration, words contained in multiple documents are suitable for queries targeting those documents. Furthermore, a set of synonyms related to a specific field is suitable for queries targeting that specific field. Furthermore, a first string entered by a user or a second string obtained by correcting the first string may contain words that appropriately reflect the intent of the user's query. These words are updated based on evaluation information from users, making them more suitable for queries targeting a specific field. Therefore, according to the above configuration, a query from a user can be corrected to a string suitable for queries related to a specific field. As a result, the likelihood of retrieving an appropriate document that meets the query from multiple documents increases, thereby improving the accuracy of responses.

[0146] (Appendix 3) the word database includes a plurality of words and a priority of each word; the generating means replaces at least a part of the first character string with a word selected from a plurality of words in the word database based on the priority; the updating means updates the priority of each word in the word database based on the evaluation information; 3. The information processing device according to claim 1 or 2.

[0147] According to the above configuration, appropriate words selected from among words contained in the first character string entered by the user or the second character string obtained by correcting the first character string according to the user's evaluation of the query result are used when generating the second character string, thereby improving the accuracy of the response.

[0148] (Appendix 4) The update means assigning a weighting coefficient to each word in the word database in the order of a good evaluation indicated by the evaluation information, an unobtainable evaluation information, and a bad evaluation indicated by the evaluation information; updating the priority of each word in the word database based on the weighting coefficient; 4. The information processing device according to claim 3.

[0149] According to the above configuration, if the user gives a good evaluation to the query result corresponding to the first character string input by the user or the second character string obtained by correcting the first character string, the priority is updated by placing importance on the words included in the first character string or the second character string, thereby enabling the generation of a more accurate second character string.

[0150] Furthermore, with the above configuration, when a user rating is not available, the words included in the first or second character string can be given the second highest importance after a good rating, and the priority can be updated. This is because when a rating is not available, it does not necessarily mean that the user rating is bad, but may also be good. Therefore, there is a possibility that the first or second character string contains words that should be given high importance.

[0151] Furthermore, with the above configuration, if the user's rating is poor, the priority can be updated so that the words included in the first character string or the second character string are not given importance. This is because if the rating is poor, it is highly likely that the content of the user's inquiry is not related to the specific field that the user is inquiring about.

[0152] (Appendix 5) the generating means generates the second character string when an amount of information included in a query result corresponding to the query indicated by the first character string is equal to or less than a threshold. 5. An information processing device according to any one of appendices 1 to 4.

[0153] According to the above configuration, if the amount of information (number of documents) included in the query result corresponding to the first character string is equal to or less than a threshold, it is considered that the user's satisfaction is low. Therefore, since the second character string is generated in such a case, the query result can be returned again using the generated second character string, thereby improving the user's satisfaction.

[0154] (Appendix 6) the generating means generates the second character string when evaluation information on a query result in response to the query indicated by the first character string satisfies a predetermined condition. 6. An information processing device according to any one of appendices 1 to 5.

[0155] According to the above configuration, a second character string is generated in such a case, and the generated second character string can be used to respond to the inquiry result again, thereby improving user satisfaction.

[0156] (Appendix 7) the first character string is input by the user using an input language that involves input conversion; further comprising an identification means for identifying a third character string before input conversion for the first character string; the generating means extracts, from the word database, a word that can be input converted from at least a part of the third character string, and replaces at least a part of the first character string with the extracted word; 7. An information processing device according to any one of appendices 1 to 6.

[0157] According to the above configuration, the second character string is generated by converting at least a part of the third character string before the user performs input conversion into a word in the word database. Therefore, even if the user performs an operation different from his or her original intention, a more appropriate second character string can be generated to obtain a query result. As a result, when a query is input using an input language that involves input conversion, the accuracy of the response can be improved.

[0158] (Appendix 8) the generating means generates a plurality of candidates that can be the second character string, and selects one of the plurality of candidates based on an operation by the user to be the second character string. 8. An information processing device according to any one of appendices 1 to 7.

[0159] According to the above configuration, it is possible to correct the first character string and generate a second character string that matches the user's intention, thereby improving the user's satisfaction with the inquiry result.

[0160] (Appendix 9) an input acquisition means for acquiring a first character string indicating an inquiry from a user; a generating means for generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; evaluation acquisition means for acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; an update means for updating the word database based on the evaluation information; An information processing system including:

[0161] According to the above configuration, the same effects as those in Supplementary Note 1 are achieved.

[0162] (Appendix 10) obtaining a first string indicative of a query from a user; generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; and updating the word database based on the evaluation information; An information processing method including:

[0163] According to the above configuration, the same effects as those in Supplementary Note 1 are achieved.

[0164] (Appendix 11) A program for causing a computer to function as an information processing device, the program comprising: an input acquisition means for acquiring a first character string indicating an inquiry from a user; a generating means for generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; evaluation acquisition means for acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; an update means for updating the word database based on the evaluation information; A program that functions as a

[0165] According to the above configuration, the same effects as those in Supplementary Note 1 are achieved.

[0166] [Appendix 3] Some or all of the above-described embodiments can also be expressed as follows.

[0167] at least one processor; The processor: an input acquisition process for acquiring a first character string indicating an inquiry from a user; a generation process of generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; an evaluation acquisition process for acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; an update process for updating the word database based on the evaluation information; An information processing device that executes the above.

[0168] The information processing device may further include a memory, and the memory may include the Input Acquisition Processing and Generate Processing and Obtaining an evaluation Processing and , the update process and The program may be stored in a computer-readable, non-transitory, tangible recording medium. [Explanation of symbols]

[0169] 1, 1A, 1B Information processing equipment 2A, 2B answering machine 3A, 3B Document Database 4A, 4B Synonym Database 5A, 5B evaluation information database 6A, 6B word database 7A, 7B Management device 10, 10A, 10B Information Processing Systems 11, 11A, 11B Input acquisition section 12, 12A, 12B generation section 13, 13A, 13B Evaluation Acquisition Department 14, 14A, 14B update section 15B Specific part C1 processor C2 Memory

Claims

1. an input acquisition means for acquiring a first character string indicating an inquiry from a user; a generating means for generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; evaluation acquisition means for acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; an update means for updating the word database based on the evaluation information; Equipped with the word database includes a plurality of words and a priority of each word; the generating means replaces at least a part of the first character string with a word selected from a plurality of words in the word database based on the priority; the updating means updates the priority of each word in the word database based on the evaluation information; Information processing device.

2. the query result indicates a search result obtained by searching a plurality of documents related to a specific field using the first character string or the second character string; The information processing device according to claim 1 , wherein the word database includes some or all of the words contained in the plurality of documents, a set of synonyms related to the specific field, and words contained in the first character string or the second character string.

3. The update means assigning a weighting coefficient to each word in the word database in the order of a good evaluation indicated by the evaluation information, an unobtainable evaluation information, and a bad evaluation indicated by the evaluation information; updating the priority of each word in the word database based on the weighting coefficient; 3. The information processing device according to claim 1.

4. the generating means generates the second character string when an amount of information included in a query result corresponding to the query indicated by the first character string is equal to or less than a threshold. The information processing device according to claim 1 .

5. the generating means generates the second character string when evaluation information on a query result in response to the query indicated by the first character string satisfies a predetermined condition. The information processing device according to claim 1 .

6. the first character string is input by the user using an input language that involves input conversion; further comprising a specifying means for specifying a third character string before input conversion for the first character string; the generating means extracts, from the word database, a word that can be input converted from at least a part of the third character string, and replaces at least a part of the first character string with the extracted word; The information processing device according to claim 1 .

7. an input acquisition means for acquiring a first character string indicating an inquiry from a user; a generating means for generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; evaluation acquisition means for acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; an update means for updating the word database based on the evaluation information; Including, the word database includes a plurality of words and a priority of each word; the generating means replaces at least a part of the first character string with a word selected from a plurality of words in the word database based on the priority; the updating means updates the priority of each word in the word database based on the evaluation information; Information processing system.

8. A computer comprising: obtaining a first string indicative of a query from a user; generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; and updating the word database based on the evaluation information; Including, the word database includes a plurality of words and a priority of each word; In the generating step, at least a part of the first character string is replaced with a word selected from a plurality of words in the word database based on the priority; In the updating, the priority of each word in the word database is updated based on the evaluation information. Information processing methods.

9. A program for causing a computer to function as an information processing device, the program comprising: an input acquisition means for acquiring a first character string indicating an inquiry from a user; a generating means for generating a second character string by referring to a word database and replacing at least a part of the first character string with a word in the word database; evaluation acquisition means for acquiring evaluation information indicating an evaluation by the user of an inquiry result in response to an inquiry indicated by the first character string or the second character string; an update means for updating the word database based on the evaluation information; It functions as the word database includes a plurality of words and a priority of each word; the generating means replaces at least a part of the first character string with a word selected from a plurality of words in the word database based on the priority; the updating means updates the priority of each word in the word database based on the evaluation information; program.

Images