Information processing device, information processing method, and program

Abstract

To provide answers that are in line with the questioner's intent. [Solution] An information processing device 1 has an acquisition unit 131 that acquires text indicating an inquiry from a user, an extraction unit 132 that extracts a main structure contained in the text, the main structure having as its components element information indicating the element being inquired about in the text, target information indicating the target of the inquiry in the text, and processing information indicating the content of the processing that is the target of the inquiry in the text, a selection unit 133 that selects a template corresponding to the main structure extracted by the extraction unit 132, and a response text indicating a response to the text based on the template selected by the selection unit 133.

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an information processing method, and a program. [Background technology]

[0002] BACKGROUND ART There is known a device for reducing the workload of replying by extracting reply sentences based on keywords contained in inquiries (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-161375 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in conventional technology, answer text is searched for by matching the question with pre-stored text in predetermined units, but there was a problem that depending on the keywords contained in the question, an answer that differs from the intent of the question could be extracted.

[0005] The present invention has been made in view of the above circumstances, and aims to provide an answer that is in line with the questioner's intent. [Means for solving the problem]

[0006] An information processing device of a first aspect of the present invention includes an acquisition unit that acquires text indicating an inquiry from a user, an extraction unit that extracts a main structure contained in the text, the main structure having components: element information indicating the element to be inquired about in the text, target information indicating the target of the inquiry in the text, and processing information indicating the content of the processing that is the target of the inquiry in the text; a selection unit that selects a template corresponding to the main structure extracted by the extraction unit; and a response generation unit that generates a response text indicating a response to the text based on the template selected by the selection unit.

[0007] The system may further include a storage unit that stores a plurality of templates in which reference element information, which is reference element information, reference target information, which is reference target information, reference processing information, which is reference processing information, and answer texts corresponding to the reference element information, the reference target information, and the reference processing information are associated, and the selection unit may select a template from the plurality of templates stored in the storage unit that corresponds to the element information, the target information, and the processing information extracted by the extraction unit.

[0008] The extraction unit may further extract premise information, which is an element that modifies any of the components that make up the main topic structure included in the text, and the selection unit may select a template corresponding to the main topic structure based further on the premise information extracted by the extraction unit.

[0009] The storage unit may store a plurality of templates that are premise information, which is an element that modifies one of the components that make up the main structure, and that are further associated with reference premise information, which is reference premise information; the extraction unit may further extract the premise information about the main structure contained in the text acquired by the acquisition unit; and the selection unit may select from the plurality of templates a template that corresponds to the element information, target information, processing information, and premise information extracted by the extraction unit.

[0010] The storage unit may further store a plurality of supplementary templates that associate the template with reference premise information, which is reference premise information, and supplementary answer text corresponding to the template and the reference premise information, the selection unit may further select a supplementary template from the plurality of supplementary templates that corresponds to the main subject structure and the premise information, and the response generation unit may generate the response text based on the template and the supplementary template.

[0011] The system may further include a determination unit that determines whether any of the components of the main structure extracted by the extraction unit are missing, and a message generation unit that, when the determination unit determines that there are missing components, generates a message to inform the user of the missing components.

[0012] The information processing device may further include a search unit that searches for candidate text that will be an answer candidate responding to the text based on the main subject structure when the selection unit is unable to select a template that corresponds to the main subject structure; a display control unit that displays the candidate text searched for by the search unit on a terminal used by an administrator of the information processing device and displays on the terminal a screen that prompts the administrator for approval; and a template registration unit that, when the administrator approves the candidate text, associates the main subject structure with the candidate text and stores them in the storage unit.

[0013] An information processing method of a second aspect of the present invention includes an acquisition step executed by a computer to acquire text indicating an inquiry from a user; an extraction step to extract a subject structure contained in the text, the subject structure having as its components element information indicating the element to be inquired about in the text, target information indicating the target of the inquiry in the text, and processing information indicating the content of the processing that is the target of the inquiry in the text; a selection step to select a template corresponding to the subject structure extracted in the extraction step; and a response generation step to generate a response text indicating a response to the text based on the template selected in the selection step. [Effects of the Invention]

[0014] According to the present invention, it is possible to provide an answer that is in line with the questioner's intent. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a diagram for explaining an overview of an information processing system S. [Figure 2] 1 is a block diagram showing a configuration of an information processing device 1. FIG. [Figure 3] 3 is a diagram showing an example of the data structure of template data stored in storage unit 12. FIG. [Figure 4] FIG. 10 is a diagram for explaining the processing of an extraction unit 132. [Figure 5] 10 is a diagram showing another example of the data structure of template data stored in storage unit 12. FIG. [Figure 6] FIG. 10 is a diagram for explaining the processing of an extraction unit 132. [Figure 7] 3 is a flowchart showing the flow of processing in the information processing device 1. DETAILED DESCRIPTION OF THE INVENTION

[0016] [Outline of Information Processing System S] FIG. 1 is a diagram for explaining an overview of an information processing system S. The information processing system S is a system for responding to inquiries received from users. The information processing system S has an information processing device 1 and an information terminal 2. In the following, an example will be described in which the information processing system S provides answers to inquiries regarding administrative services provided by a city hall or the like, but the application area of ​​the information processing system S is not limited to this.

[0017] The information processing device 1 is a device for providing a response according to the intent of an inquiry received from a user. The information terminal 2 is a terminal used by a user who makes an inquiry via the information processing system S. The information terminal 2 is, for example, a smartphone, a tablet, a personal computer, or a dedicated terminal installed at a reception desk.

[0018] The processing of the information processing system S will be described. The information processing device 1 acquires an inquiry text ((1) in FIG. 1). The inquiry text is text data indicating the content of an inquiry from a user. The information processing device 1 may acquire the inquiry text sent as a text message, may acquire the inquiry text sent as an email, or may acquire the inquiry text by performing speech recognition on the speech uttered by the user. The following processing will be described using as an example a case where the inquiry text is "Please tell me the procedure for reissuing my nursing care insurance card."

[0019] The information processing device 1 extracts the subject structure contained in the query text ((2) in Figure 1). The subject structure is information indicating the subject of the query. The subject structure includes element information, target information, and process information. The element information indicates the element being queried in the text. The element information is information indicating whether the query text is inquiring about a place, a time, a procedure, a method, etc. The target information indicates the target of the query. In the above example, the target information indicates that the target of the query is "long-term care insurance card." The process information indicates the content of the process or action to be taken in relation to the query target in the text. In the above example, the process information is "reissue."

[0020] As an example, the information processing device 1 inputs a query text into an extraction model and outputs a topic structure. As an example, the extraction model is a large-scale language model (LLM). As an example, the information processing device 1 may accept input of a prompt indicating an instruction to extract a topic structure. The extraction model may be trained in advance based on training data that associates a query text for training with a topic structure corresponding to the query text for training, so that when a query text is input, the extraction model outputs a topic structure corresponding to the input query text.

[0021] The information processing device 1 selects a template corresponding to the extracted main subject structure ((3) in FIG. 1). The information processing device 1 stores template data in which the main subject structure is associated with an answer text corresponding to the main subject structure. The information processing device 1 generates a response text based on the selected template ((4) in FIG. 1). The information processing device 1 displays the generated response text on the information terminal 2 ((5) in FIG. 1).

[0022] The information processing system S is configured to extract a subject structure that indicates the intent of the query and select an answer template associated with an answer that corresponds to the subject structure, thereby making it possible to provide an answer that is in line with the intent of the questioner.

[0023] [Configuration of information processing device 1] 2 is a block diagram showing the configuration of information processing device 1. Information processing device 1 has a communication unit 11, a storage unit 12, and a control unit 13. The control unit 13 has an acquisition unit 131, an extraction unit 132, a selection unit 133, a response generation unit 134, a display control unit 135, a determination unit 136, a message generation unit 137, a search unit 138, and a template registration unit 139. The communication unit 11 is a communication interface for transmitting and receiving data to and from other devices via a network.

[0024] The storage unit 12 is a storage medium including a read-only memory (ROM), a random access memory (RAM), a solid-state drive (SSD), a hard disk drive, etc. The storage unit 12 stores in advance programs to be executed by the control unit 13.

[0025] The storage unit 12 stores template data. FIG. 3 is a diagram showing an example of the data structure of a template stored in the storage unit 12. In the template data, the storage unit 12 associates reference element information, reference target information, reference process information, and reference answer text. The reference element information is element information that serves as a reference. The reference target information is object information that serves as a reference. The reference process information is process information that serves as a reference. The reference answer text is answer text that corresponds to the reference element information, reference target information, and reference process information. The storage unit 12 stores extraction models.

[0026] The control unit 13 is a processor such as a CPU (Central Processing Unit), and functions as an acquisition unit 131, an extraction unit 132, a selection unit 133, a response generation unit 134, a display control unit 135, a determination unit 136, a message generation unit 137, a search unit 138, and a template registration unit 139 by executing a program stored in the storage unit 12.

[0027] The acquisition unit 131 acquires a query text indicating an inquiry from a user. The acquisition unit 131 acquires the query text transmitted from the information terminal 2. As an example, the display control unit 135 causes the information terminal 2 to display a screen for inputting the query text. The acquisition unit 131 acquires the query text input by the user on the screen displayed on the information terminal 2. The extraction unit 132 extracts a subject structure included in the query text. Specifically, the extraction unit 132 inputs the query text to an extraction model and causes the extraction model to output the subject structure.

[0028] The selection unit 133 selects a template corresponding to the main structure extracted by the extraction unit 132. As an example, the selection unit 133 selects a template corresponding to the element information, target information, and processing information extracted by the extraction unit 132 from among a plurality of templates stored in the storage unit 12. Specifically, the selection unit 133 selects a template having reference element information, reference target information, and reference processing information that match all of the element information, target information, and processing information extracted by the extraction unit 132. The selection unit 133 may also calculate the compatibility between the element information, target information, and processing information extracted by the extraction unit 132 and the reference element information, reference target information, and reference processing information included in the template, and select a template based on the calculated compatibility. As an example, the compatibility is the similarity of semantic vectors.

[0029] The response generation unit 134 generates a response text indicating a response to the text based on the template selected by the selection unit 133. As an example, the storage unit 12 may store a template for generating a response text. In this case, the response generation unit 134 generates a response text by inserting an answer text associated with the selected template into the template for the response. If the answer text included in the template has a placeholder, the response generation unit 134 may generate a response by inputting an element extracted from the query text into the placeholder. Alternatively, the response generation unit 134 may generate a response text by inputting a question sentence, the answer text, and a prompt instructing the large-scale language model to generate a response text based on the question sentence and the answer text.

[0030] The information processing device 1 is configured to select a template containing an answer based on the subject structure extracted from the query text and generate a response text based on the selected template, thereby making it possible to provide an answer that is in line with the intent of the user's query and to provide a highly reliable answer.

[0031] By configuring the information processing device 1 to select a template based on detailed context included in the question text, the accuracy of the answer can be improved.

[0032] The extraction unit 132 further extracts premise information, which is an element that modifies one of the components constituting the main structure included in the text. FIG. 4 illustrates an example of the processing performed by the extraction unit 132 in this case. The premise information includes elements of "item number," "target," "content," and "label." "Target" indicates the element of the main structure that the premise information modifies. "Content" indicates the content that the premise information modifies. "Label" indicates the manner in which the premise information modifies. "Label" can be expressed, for example, as "possession," "state," "time," "quantity," "location," etc. The premise information may further include "link." "Link" is identification information added to identify the premise information that modifies each of the premise information when the premise information modifies other premise information. In the example of FIG. 4, the premise information of item #3 includes information of "link: #2." This indicates that the premise information of item #3 modifies the information "lost" of item #2.

[0033] Specifically, the extraction unit 132 inputs a query text to the extraction model and causes it to further output premise information. In this case, the extraction model is trained to further output premise information when a query text is input. As an example, the extraction model in this case is trained based on training data that associates a training query text, a topic structure corresponding to the training query text, and premise information corresponding to the training query text.

[0034] The selection unit 133 selects a template corresponding to the main subject structure based further on the premise information extracted by the extraction unit 132. The data structure of the template information in this case is shown in Fig. 5. The storage unit 12 stores a plurality of templates that are further associated with reference premise information, which is premise information that is an element that modifies one of the components that make up the main subject structure and is reference premise information.

[0035] The selection unit 133 selects, from the plurality of templates, a template that corresponds to the element information, target information, processing information, and premise information extracted by the extraction unit 132. As already described, the selection unit 133 may select a template that matches the main subject structure and premise information extracted by the extraction unit 132, or may select a template whose degree of conformance with the main subject structure and premise information extracted by the extraction unit 132 is equal to or greater than a predetermined threshold.

[0036] By configuring the information processing device 1 to select a template based on premise information, it is possible to provide an answer that matches the intent of the user's inquiry.

[0037] The premise information may be configured to indicate a history of procedures, steps, work, etc., for the inquiry target. In this case, the storage unit 12 stores a template that associates the subject structure, premise information indicating the history, and answer text.

[0038] As an example, if the information processing system S is used to receive inquiries about malfunctions at a support center for electronic devices, the premise information may indicate a history of responses to the malfunctions. An example of the processing performed by the extraction unit 132 in this case is shown in Fig. 6. As shown in Fig. 6, the history of responses to the malfunctions is extracted as premise information. The template selection unit 133 selects a template based on the extracted premise information.

[0039] The template may be configured of a common template having an answer corresponding to the main topic structure and a supplemental template having an answer corresponding to the main topic structure and premise information. Specifically, the storage unit 12 further stores a plurality of supplemental templates each of which associates a template with reference premise information that is reference premise information, and a supplemental answer text corresponding to the template and the reference premise information.

[0040] The selection unit 133 further selects a supplemental template corresponding to the main topic structure and premise information from the plurality of supplemental templates. That is, the selection unit 133 selects a template corresponding to the main topic structure extracted by the extraction unit 132, and also selects a supplemental template corresponding to the main topic structure and premise information extracted by the extraction unit 132. Then, the response generation unit 134 generates a response text based on the template and the supplemental template.

[0041] By configuring the information processing device 1 in this way, it becomes possible to limit the influence of additions, changes, etc. to the answer content, and it becomes possible to provide flexible answers.

[0042] When the query text contains a subject structure or premise information in ambiguous terms, the information processing device 1 may be configured to extract the subject structure or premise information that has been matched. For example, when "what to submit when moving" is extracted as a candidate for the subject structure's target information, the extraction unit 132 may extract the target information by replacing "what to submit when moving" with "notification of moving out" or "notification of moving in." As an example, the storage unit 12 stores a table for name matching, and replaces information extracted as a candidate for the subject structure by referring to the table.

[0043] The storage unit 12 may also store a name matching model trained to output name-matched information corresponding to input information. In this case, the extraction unit 132 inputs information extracted as a candidate for the main structure into the name matching model, causing it to output the name-matched information. By configuring the information processing device 1 in this way, it is possible to prevent variations in expression from affecting answers, thereby improving the accuracy of answers provided to users.

[0044] If elements constituting the main subject structure cannot be extracted from the query text, a problem occurs in that a template cannot be selected and a reliable answer cannot be provided. Therefore, the information processing device 1 may be configured to send a confirmation message to the user to confirm the missing elements when elements constituting the main subject structure cannot be extracted from the query text.

[0045] The determination unit 136 determines whether any of the components of the main subject structure extracted by the extraction unit 132 are missing. As an example, the extraction model extracts candidates corresponding to each of the element information, target information, and processing information from the query text, and outputs the candidates for each element in association with a certainty factor, which is the likelihood of each element. If any of the elements output by the extraction model has a certainty factor equal to or less than a predetermined threshold, the determination unit 136 determines that any of the components of the main subject structure are missing. The determination unit 136 identifies elements with a certainty factor equal to or less than a predetermined threshold as missing components.

[0046] When the determination unit 136 determines that there are missing components, the message generation unit 137 generates a confirmation message for the user to confirm the missing components. As an example, the message generation unit 137 generates the confirmation message based on the elements extracted by the extraction unit 132 and the elements (missing elements) that the extraction unit 132 could not extract. As an example, the storage unit 12 stores a template for each missing component, and generates a confirmation message by inserting the identified element into the template corresponding to the missing element. For example, when the element information and processing information can be identified as "method" and "reissue" but the target information cannot be identified, the message generation unit 137 generates a confirmation message asking, "What reissue method are you looking for?" Furthermore, when none of the elements of the subject structure can be identified, the message generation unit 137 may generate a fixed phrase such as "Please tell me more details" as the confirmation message.

[0047] The display control unit 135 displays the confirmation message generated by the message generation unit 137 on the information terminal 2. The message generation unit 137 may transmit the generated confirmation message to the information terminal 2 as a text message. The acquisition unit 131 acquires a response to the confirmation message. The extraction unit 132 further extracts a main subject structure based on the response to the confirmation message acquired by the acquisition unit 131. For example, the extraction unit 132 inputs the query text, the confirmation message, and the response to the confirmation message into an extraction model, and outputs the main subject structure. The extraction unit 132 may extract the main subject structure and premise information based further on the response to the confirmation message.

[0048] The display control unit 135 may display the confirmation message along with answer candidates for the confirmation message. As an example, the message generation unit 137 identifies answer candidates based on the main subject structure extracted by the extraction unit 132. The answer candidates are, for example, information included in other elements associated with the same element extracted by the extraction unit 132 among the elements constituting the main subject structure in the template stored in the storage unit 12. That is, when element information and processing information are extracted as “method” and “reissue” and target information cannot be identified, and “health insurance card,” “long-term care insurance card,” and “My Number card” are pre-registered as target information for a record whose element information and processing information are “method” and “reissue” in the template information, the message generation unit 137 identifies “health insurance card,” “long-term care insurance card,” and “My Number card” as answer candidates. The message generation unit 137 generates a confirmation message configured to allow selection from the identified answer candidates. The display control unit 135 displays the confirmation message and the selectable answer candidates on the information terminal 2. The acquisition unit 131 acquires the answer candidate selected by the user from among the answer candidates as a response to the confirmation message.

[0049] Even if elements constituting the main subject structure are extracted, there may be cases where the number of templates cannot be narrowed down due to a lack of premise information. In such cases, the message generation unit 137 may be configured to generate a confirmation message for confirming the premise information. In other words, if multiple templates correspond based on the main subject structure, the message generation unit 137 may generate a confirmation message. Furthermore, as an example, the storage unit 12 may store a necessary information table. The necessary information table associates element information, target information, and processing information included in the main subject structure with premise information that needs to be acquired in addition to the element information, target information, and processing information. The determination unit 136 refers to the necessary information table based on the extracted main subject structure and premise information, and determines whether the necessary premise information has been acquired. If the determination unit 136 determines that the necessary premise information has not been acquired, the message generation unit 137 may generate a confirmation message.

[0050] By configuring the information processing device 1 to generate a confirmation message when there are missing elements in the query text to identify the answer, it becomes possible to accurately grasp the user's intention and provide an answer that is in line with the user's intention.

[0051] However, if an appropriate template corresponding to the extracted main subject structure cannot be selected, a problem may arise in which an appropriate answer cannot be provided to the user. Therefore, the information processing device 1 may be configured to expand the template based on information acquired by the search.

[0052] When the selection unit 133 cannot select a template corresponding to the main topic structure, the search unit 138 searches for candidate text that will be an answer candidate responding to the text, based on the main topic structure. The selection unit 133 cannot select a template corresponding to the main topic structure when there is no template corresponding to the extracted main topic structure or when there is no template corresponding to the extracted main topic structure and premise information. In this case, the search unit 138 performs a web search, a database search, or the like, using the extracted main topic structure and premise information as keys to acquire candidate text.

[0053] The display control unit 135 displays the candidate texts searched by the search unit 138 on a terminal used by the administrator of the information processing device, and also displays a screen (hereinafter referred to as an approval screen) on the terminal to prompt the administrator for approval. As an example, the approval screen displays the inquiry text, the candidate texts, and a button for selecting whether to approve or deny the approval.

[0054] When the administrator approves the candidate text, the template registration unit 139 associates the main topic structure with the candidate text and stores them in the storage unit 12. As an example, when the administrator presses an approval button on the approval screen, the template registration unit 139 associates the main topic structure extracted by the extraction unit 132 with the candidate text and stores them in the storage unit 12. Furthermore, the template registration unit 139 may associate the main topic structure and premise information extracted by the extraction unit 132 with the candidate text and store them in the storage unit 12.

[0055] The timing at which a template is stored in the storage unit 12 is not limited to after an approval operation by the administrator. That is, the template registration unit 139 may be configured to store the subject structure, premise information, and searched candidate text in the storage unit 12 in advance, and to activate the template so that it can be used for an answer when an operation is performed by the administrator.

[0056] The approval screen may be configured so that the administrator can edit the main subject structure, premise information, and candidate text. In this case, the template registration unit 139 stores the edited main subject structure, premise information, and candidate text in the storage unit 12.

[0057] If the selection unit 133 cannot select a template corresponding to the main subject structure, the display control unit 135 may display a screen indicating that an answer cannot be provided (for example, a screen urging the user to inquire at a counter) on the information terminal 2. Furthermore, if the approval button is pressed, the response generation unit 134 may generate a response text based on the candidate text. Then, the display control unit 135 may display the generated response text on the information terminal 2 of the user who sent the query text.

[0058] If an appropriate template corresponding to the extracted main subject structure cannot be selected, the information processing device 1 is configured to expand the template based on information acquired by the search, thereby providing a highly flexible response.

[0059] [Processing flow in information processing device 1] Fig. 7 is a flowchart showing the flow of processing in the information processing device 1. The flowchart shown in Fig. 7 starts from the point when the information processing device 1 is started up and becomes capable of acquiring the query text.

[0060] The acquisition unit 131 acquires a query text (S01). The extraction unit 132 extracts the main topic structure and premise information contained in the query text (S02). The determination unit 136 determines whether the extracted information is missing any elements (S03).

[0061] If there are no missing elements in the extracted information (NO in S03), the selection unit 133 determines whether a template is selectable (S04). As an example, the selection unit 133 determines that a template is selectable if the template can be uniquely identified. If it is determined that the template is selectable (YES in S04), the selection unit 133 selects a template based on the extracted main subject structure and premise information (S05). The response generation unit 134 generates response text based on the selected template (S06). The display control unit 135 displays the response text on the information terminal 2 (S07). Then, the information processing device 1 ends the process.

[0062] If there are missing elements (YES in S03), the message generation unit 137 generates a confirmation message (S08). The display control unit 135 displays the generated confirmation message on the information terminal 2 (S09). The acquisition unit 131 acquires a response to the confirmation message as inquiry text (S01), and executes the processes from S02 onwards. Note that the information processing device 1 may cause the extraction unit 132 to further extract a main structure or a premise structure from the acquired response, and execute the processes from S04 onwards. Note that, as described above, the message generation unit 137 may generate a confirmation message including answer options, and the acquisition unit 131 may acquire the option selected by the user from the answer options.

[0063] If the selection unit 133 determines that a template cannot be selected (NO in S04), the search unit 138 searches for candidate text based on the extracted main topic structure and premise information (S10). The display control unit 135 displays an approval screen on the administrator's terminal to prompt approval of the candidate text (S11). When the administrator performs an approval operation, the template registration unit 139 registers the main topic structure, premise information, and candidate text as a template (S12). The response generation unit 134 generates response text based on the approved candidate text (S13). Then, the information processing device 1 ends the processing.

[0064] [Effects of information processing device 1] The information processing device 1 is configured to extract a subject structure that indicates the intent of the inquiry and select an answer template associated with an answer that corresponds to the subject structure, thereby making it possible to provide an answer that is in line with the intent of the questioner.

[0065] <Modification> In the above description, an example has been described in which the information processing system S is configured as a client-server system, but the functions of the information processing device 1 may be configured to be executed on a tablet terminal or a reception terminal installed at a counter. In this case, the information processing device 1 according to the modified example may be configured to include a voice input unit that accepts voice input from a user, a voice recognition unit that recognizes the input voice, a touch panel, a display unit that displays a response text, a speaker that outputs the response text as voice, etc.

[0066] The present invention has been described above using embodiments, but the technical scope of the present invention is not limited to the scope described in the above embodiments, and various modifications and changes are possible within the scope of the gist of the present invention. For example, all or part of the device can be configured by functionally or physically distributing or integrating any unit. Furthermore, new embodiments resulting from any combination of multiple embodiments are also included in the embodiments of the present invention. The effects of the new embodiments resulting from the combination also have the effects of the original embodiments. [Explanation of symbols]

[0067] 1. Information processing equipment 2. Information terminal 11 Communications Department 12 Storage section 13 Control Unit 131 Acquisition Department 132 Extraction part 133 Selection Section 134 Response Generation Unit 135 Display control unit 136 Judgment section 137 Message Generation Unit 138 Search Department 139 Template Registration Section

Claims

1. an acquisition unit that acquires text indicating an inquiry from a user; an extraction unit that extracts a main structure included in the text, the main structure having elements of element information indicating an element to be queried in the text, object information indicating a target of the query in the text, and processing information indicating the content of the processing to be queried in the text; a selection unit that selects a template corresponding to the subject structure extracted by the extraction unit; a response generation unit that generates a response text indicating a response to the text based on the template selected by the selection unit; An information processing device having the above.

2. a storage unit that stores a plurality of templates in which reference element information, which is reference element information, reference object information, which is reference object information, reference process information, which is reference process information, and answer texts corresponding to the reference element information, the reference object information, and the reference process information are associated with each other; the selection unit selects a template corresponding to the element information, the target information, and the processing information extracted by the extraction unit from among a plurality of templates stored in the storage unit. The information processing device according to claim 1 .

3. The extraction unit further extracts premise information that is an element that modifies any of the components that make up the main structure included in the text, the selection unit selects a template corresponding to the main subject structure further based on the premise information extracted by the extraction unit. The information processing device according to claim 1 .

4. the storage unit stores a plurality of templates each of which is premise information that modifies any of the constituent elements that make up the main structure, and which further associates reference premise information that is reference premise information; the extraction unit further extracts the premise information about the main topic structure included in the text acquired by the acquisition unit; the selection unit selects, from the plurality of templates, a template corresponding to the element information, the target information, the processing information, and the premise information extracted by the extraction unit. The information processing device according to claim 2 .

5. the storage unit further stores a plurality of supplemental templates each associated with the template, reference premise information that is reference premise information, and a supplemental answer text corresponding to the template and the reference premise information; the selection unit further selects a supplemental template corresponding to the subject structure and the premise information from among a plurality of supplemental templates; the response generator generates the response text based on the template and the supplemental template. The information processing device according to claim 2 .

6. a determination unit that determines whether or not there are any missing components among the components of the main structure extracted by the extraction unit; a message generating unit configured to generate a message for the user to confirm the missing component when the determining unit determines that the missing component is present; The information processing device according to claim 1 .

7. a search unit that searches for candidate texts that are answer candidates responding to the text based on the subject structure when the selection unit cannot select a template that corresponds to the subject structure; a display control unit that displays the candidate text searched by the search unit on a terminal used by an administrator of the information processing device and displays a screen on the terminal that prompts approval from the administrator; a template registration unit that, when the administrator approves the candidate text, associates the subject structure with the candidate text and stores them in the storage unit; The information processing device according to claim 2 , further comprising:

8. The computer executes an acquisition step of acquiring text indicating an inquiry from a user; an extraction step of extracting a subject structure included in the text, the subject structure having as its components element information indicating an element to be queried in the text, object information indicating a target of the query in the text, and processing information indicating the content of the processing to be queried in the text; a selection step of selecting a template corresponding to the subject structure extracted in the extraction step; a response generation step of generating a response text indicating a response to the text based on the template selected in the selection step; An information processing method comprising:

Images