Information processing device and information processing program

Abstract

To allow for concretizing question text entered in natural text search.SOLUTION: A natural text search device 10 disclosed herein comprises a CPU 11. The CPU 11 is configured to receive input of question text for natural text search from a user, receive a natural text search execution instruction from the user, acquire information indicating the situation where the natural text search execution instruction was given, and use the acquired information to process the question text.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to an information processing apparatus and an information processing program.

Background Art

[0002] For example, Patent Document 1 describes a search device that performs a search considering the diversity of input content expressions. This search device includes a question input unit that receives a search query from a user, a distribution estimation unit that estimates the distribution of the search query in the semantic space of natural language, a storage unit that stores information for specifying the distribution of each of a plurality of predetermined text data in the semantic space, and a distribution search unit that searches for text data having a high similarity to the search query based on the information for specifying the distribution of the text data stored in the storage unit and the distribution estimated by the distribution estimation unit, and an output unit that outputs the text data searched by the distribution search unit.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, in a system for searching FAQs (Frequently Asked Questions), the content of the question sentence is often abstract, the words are insufficient, or the content expressions of the question sentences are different by the user. If the degree of abstraction of the question sentence is high, an appropriate answer may not be obtained.

[0005] The present disclosure aims to provide an information processing apparatus and an information processing program capable of concretizing a question sentence input in a natural language search.

Means for Solving the Problems

[0006] To achieve the above objective, the information processing device according to the first embodiment includes a processor, the processor receives input of a natural language search query from a user, receives an instruction to execute a natural language search from the user, obtains information indicating that the instruction to execute the natural language search has been given, and processes the query using the obtained information.

[0007] Furthermore, in the information processing device according to the second embodiment, the information representing the situation in which the execution instruction was given by the user is the content information that the user referred to when giving the execution instruction, and the processor extracts sentences or characteristic words from the content information that are to be reflected in the question sentence.

[0008] Furthermore, in the third embodiment, the information processing device is an information processing device according to the second embodiment in which the processing is to add the extracted text or characteristic words to the question text.

[0009] Furthermore, the information processing device according to the fourth embodiment is an information processing device according to the second embodiment in which the processing is performed to generate a new question sentence from the extracted text or characteristic words and the question sentence.

[0010] Furthermore, in the information processing device according to the fifth embodiment, the processor assigns a score to each of the feature words and extracts a predetermined number of the feature words in descending order of the assigned scores.

[0011] Furthermore, in the information processing device according to the sixth embodiment, the information representing the situation in which the execution instruction was given by the user is the workflow information that the user referred to when giving the execution instruction, and the processor extracts sentences or characteristic words from the workflow information that are to be reflected in the question.

[0012] Furthermore, in the information processing device according to the seventh embodiment, the text or characteristic words to be reflected in the question are extracted from the information representing the case and stage in the workflow information in the information processing device according to the sixth embodiment.

[0013] Furthermore, in the information processing device according to the eighth embodiment, the processor in the information processing device according to the first embodiment performs control to display the processed question text.

[0014] Furthermore, in the information processing device according to the ninth embodiment, the processor controls the display of multiple processed question sentences in the information processing device according to the eighth embodiment and accepts a selection from the user.

[0015] Furthermore, in the information processing device according to the tenth embodiment, the processor in the information processing device according to the first embodiment determines the level of abstraction of the question statement that has been received as input, and processes the question statement if the level of abstraction is equal to or greater than a threshold.

[0016] Furthermore, in the information processing device according to the 11th embodiment, the level of abstraction is determined in the information processing device according to the 10th embodiment by performing the natural language search based on the question and determining the number of search results obtained by the natural language search.

[0017] Furthermore, in the information processing device according to the 12th embodiment, the level of abstraction is determined in the information processing device according to the 11th embodiment based on the number of predetermined keywords included in the search results obtained by performing the natural language search based on the question statement.

[0018] Furthermore, in order to achieve the above objective, the information processing program according to the 13th embodiment receives input of a natural language search query from a user, receives an instruction to execute a natural language search from the user, obtains information indicating that the instruction to execute a natural language search has been given, and causes the computer to perform processing of the query using the obtained information. [Effects of the Invention]

[0019] According to the first aspect and the thirteenth aspect, there is an effect that the question sentence input in natural language search can be concretized.

[0020] According to the second aspect, there is an effect that the question sentence can be accurately concretized as compared with the case of using content information not referred to by the user at the time of the execution instruction.

[0021] According to the third aspect, there is an effect that the question sentence can be easily concretized as compared with the case of generating a new question sentence from a sentence or a keyword.

[0022] According to the fourth aspect, there is an effect that the question sentence can be arranged and concretized as compared with the case of adding a sentence or a keyword to the question sentence.

[0023] According to the fifth aspect, there is an effect that keywords can be extracted in descending order of score.

[0024] According to the sixth aspect, there is an effect that the question sentence can be accurately concretized as compared with the case of using workflow information not referred to by the user at the time of the execution instruction.

[0025] According to the seventh aspect, there is an effect that the question sentence can be accurately concretized as compared with the case of not considering the project and stage in the workflow information.

[0026] According to the eighth aspect, there is an effect that the user can grasp the processed question sentence.

[0027] According to the ninth aspect, there is an effect that the user can select the desired question sentence from among a plurality of processed question sentences.

[0028] According to the tenth aspect, there is an effect that only the question sentence with a high degree of abstraction can be selectively processed.

[0029] According to the eleventh embodiment, the level of abstraction can be determined by the number of search results.

[0030] According to the twelfth embodiment, the level of abstraction can be determined by the number of keywords. [Brief explanation of the drawing]

[0031] [Figure 1] This is a diagram showing an example of the configuration of a natural text search system according to the first embodiment. [Figure 2] This is a block diagram showing an example of the electrical configuration of a natural text search device according to the first embodiment. [Figure 3] This is a block diagram showing an example of the functional configuration of a natural text search device according to the first embodiment. [Figure 4] This figure shows an example of an additional text table according to the first embodiment. [Figure 5] This figure illustrates the question processing process according to the first embodiment. [Figure 6] This figure shows an example of a score assigned to a characteristic word according to the first embodiment. [Figure 7] This flowchart shows an example of the processing flow by the natural text search program according to the first embodiment. [Figure 8] This is a block diagram showing an example of the functional configuration of a natural text search device according to the second embodiment. [Figure 9] This figure shows an example of an additional text table according to the second embodiment. [Figure 10] This figure illustrates the question processing process according to the second embodiment. [Figure 11] This flowchart shows an example of the processing flow by the natural text search program according to the second embodiment. [Figure 12] This figure shows an example of screen transitions in the natural language search screen according to this embodiment. [Figure 13] This figure shows another example of screen transitions in the natural language search screen according to the embodiment. [Figure 14] This is a block diagram showing an example of the functional configuration of a natural text search device according to the third embodiment. [Figure 15] This flowchart shows an example of the processing flow by the natural text search program according to the third embodiment. [Modes for carrying out the invention]

[0032] Hereinafter, an example of an embodiment for carrying out the technology of this disclosure will be described in detail with reference to the drawings. Components and processes that perform the same operation, action, or function are given the same reference numerals throughout the drawings, and redundant explanations may be omitted as appropriate. Each drawing is only a schematic representation to the extent that the technology of this disclosure can be fully understood. Therefore, the technology of this disclosure is not limited to the illustrated examples. Furthermore, in this embodiment, explanations of configurations not directly related to the technology of this disclosure or well-known configurations may be omitted.

[0033] [First Embodiment] Figure 1 shows an example of the configuration of the natural text search system 100 according to the first embodiment.

[0034] As shown in Figure 1, the natural language search system 100 according to this embodiment comprises a natural language search device 10 and a terminal device 30. In the example in Figure 1, one terminal device is shown, but the number is arbitrary. The natural language search device 10 is an example of an information processing device.

[0035] Terminal device 30 is a terminal device used by users of the natural language search service, and is an information terminal such as a smartphone, tablet, or personal computer (PC). Users operate terminal device 30 to access the natural language search device 10 via the network N and obtain search results for their natural language search query from the natural language search device 10.

[0036] The natural language search device 10 is connected to the terminal device 30 via a network N. Network N can be, for example, the Internet, a LAN (Local Area Network), or a WAN (Wide Area Network). The natural language search device 10 is, for example, a server computer located on the cloud, and receives natural language search queries entered by the user from the terminal device 30, and outputs the search results for those queries to the terminal device 30.

[0037] In the example shown in Figure 1, the natural language search device 10 receives a search query, which is a question sentence entered by the user from the terminal device 30. However, it is not limited to this, and the device may also accept search queries entered directly by the user from its own control panel without going through the terminal device 30.

[0038] Figure 2 is a block diagram showing an example of the electrical configuration of the natural text search device 10 according to the first embodiment.

[0039] As shown in Figure 2, the natural language search device 10 according to this embodiment includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, an input / output interface (I / O) 14, a storage unit 15, a display unit 16, an operation unit 17, and a communication unit 18.

[0040] The CPU 11, ROM 12, RAM 13, and I / O 14 are connected to each other via a bus. The I / O 14 is connected to various functional units, including a storage unit 15, a display unit 16, an operation unit 17, and a communication unit 18. These functional units are capable of communicating with the CPU 11 via the I / O 14.

[0041] The control unit is comprised of a CPU 11, ROM 12, RAM 13, and I / O 14. The control unit may be configured as a sub-control unit that controls the operation of a part of the natural language search device 10, or as part of the main control unit that controls the operation of the entire natural language search device 10. Integrated circuits or IC chipsets, such as LSIs (Large Scale Integrations), are used for some or all of each block of the control unit. Individual circuits may be used for each of the above blocks, or circuits that integrate some or all of them may be used. The above blocks may be provided as a single unit, or some of the blocks may be provided separately. Furthermore, parts of each of the above blocks may be provided separately. For the integration of the control unit, dedicated circuits or general-purpose processors may be used, not limited to LSIs.

[0042] For example, the storage unit 15 may be an HDD (Hard Disk Drive), an SSD (Solid State Drive), or flash memory. The storage unit 15 stores a natural language search program 15A for executing the natural language search service according to this embodiment. This natural language search program 15A may also be stored in the ROM 12.

[0043] The natural language search program 15A may, for example, be pre-installed on the natural language search device 10. The natural language search program 15A may also be implemented by storing it on a non-volatile storage medium or distributing it via a network N and installing it on the natural language search device 10 as appropriate. Examples of non-volatile storage mediums include CD-ROMs (Compact Disc Read Only Memory), magneto-optical disks, HDDs, DVD-ROMs (Digital Versatile Disc Read Only Memory), flash memory, and memory cards.

[0044] The display unit 16 may include, for example, a liquid crystal display (LCD), an organic EL (electroluminescence) display, etc. The display unit 16 may also have an integrated touch panel. The operation unit 17 is equipped with, for example, a keyboard, mouse, or other device for operation input. The display unit 16 and the operation unit 17 receive various instructions from the user of the natural language search device 10. The display unit 16 displays various information such as the results of processing performed in response to the instructions received from the user, and notifications regarding the processing.

[0045] The communication unit 18 is connected to a network N, such as the Internet, LAN, or WAN, and can communicate with the terminal device 30 via the network N.

[0046] By the way, as mentioned above, in the natural language search system 100, the content of the question often varies depending on the user, such as being abstract, lacking words, or having different wording. If the question is highly abstract, it may not be possible to obtain an appropriate answer.

[0047] In contrast, the natural language search device 10 according to this embodiment receives input of a natural language search query from the user, receives an instruction from the user to execute the natural language search, obtains information indicating that the instruction to execute the natural language search has been given, and processes the query using the obtained information.

[0048] Specifically, the CPU 11 of the natural language search device 10 according to this embodiment functions as the various parts shown in Figure 3 by writing the natural language search program 15A stored in the memory unit 15 to the RAM 13 and executing it. Note that the CPU 11 is an example of a processor.

[0049] Figure 3 is a block diagram showing an example of the functional configuration of the natural language search device 10 according to the first embodiment.

[0050] As shown in Figure 3, the CPU 11 of the natural language search device 10 according to this embodiment functions as a reception unit 11A, an acquisition unit 11B, a processing unit 11C, and a search unit 11D.

[0051] The reception unit 11A receives input of a natural language search query from the user via the terminal device 30 and receives instructions from the user to execute the natural language search.

[0052] When the acquisition unit 11B receives an execution instruction for natural language search from the reception unit 11A, it acquires information indicating the status of the execution instruction. Here, "information indicating the status of the execution instruction" is, for example, the content information that the user was referring to when giving the execution instruction. The content information is associated with, for example, a URL (Uniform Resource Locator) that represents the source of the content information.

[0053] The processing unit 11C processes the question text using the information acquired by the acquisition unit 11B. Specifically, the processing unit 11C extracts sentences or characteristic words to be reflected in the question text from the content information that the user was referring to when issuing the execution instruction, and processes the question text using the extracted sentences or characteristic words. This processing may involve, for example, adding the extracted sentences or characteristic words to the question text, or generating a new question text from the extracted sentences or characteristic words and the question text.

[0054] The processing unit 11C may also control the terminal device 30 to display the processed question text and prompt the user to confirm the processed question text. Alternatively, the processing unit 11C may control the terminal device 30 to display multiple processed question texts and accept selection from the user for the multiple processed question texts.

[0055] The search unit 11D performs a natural language search on the processed query text obtained by the processing unit 11C, and outputs the search results to the terminal device 30.

[0056] Here, the memory unit 15 stores the supplementary text table 151. The extraction of sentences or characteristic words to be reflected in the question is performed, for example, using the supplementary text table 151.

[0057] Figure 4 shows an example of an additional text table 151 according to the first embodiment.

[0058] In the supplementary text table 151 shown in Figure 4, the source (URL) of the content information accessible to the user and the supplementary text are defined in association. The supplementary text is defined as a characteristic sentence that represents the content information. Alternatively, characteristic words to be added may be defined in association with the supplementary text. The characteristic words to be added are defined as characteristic words that represent the content information.

[0059] Figure 5 is a diagram illustrating the question processing process according to the first embodiment.

[0060] In Figure 5 (S1), the user operates the terminal device 30 to access content information (for example, manual: opening a savings account) from the terminal device 30. This content information is associated with a referral source (URL) "http: / / manual / yokin / kouza / ".

[0061] In (S2), if the user wants to ask a question while viewing content information, the terminal device 30 displays the natural language search screen 40. The user enters a question, for example, "I want to know how to verify my identity," into the natural language search screen 40 and presses the "Execute Search" button.

[0062] In (S3), when the reception unit 11A of the natural language search device 10 receives a search execution instruction along with the question entered via the natural language search screen 40, the acquisition unit 11B acquires the source (URL) of the content information that the user is currently viewing from the terminal device 30. Then, the processing unit 11C, based on the acquired source (URL) of the content information, refers to the additional sentence table 151 shown in Figure 4 above as an example and extracts sentences to be added. In the example in Figure 5, the sentence to be added is extracted as "In opening a regular savings account,". Then, the processing unit 11C adds the extracted sentence to the question, for example, "I want to know the method of identity verification in opening a regular savings account." In other words, the original abstract question, "I want to know the method of identity verification," is automatically processed into the more specific question, "I want to know the method of identity verification in opening a regular savings account."

[0063] In (S4), the search unit 11D of the natural language search device 10 performs a natural language search on the processed question sentence obtained in (S3) and displays the search results on the natural language search screen 40 of the terminal device 30.

[0064] Furthermore, the extraction of sentences or characteristic words to be reflected in the question can be performed without using the supplementary sentence table 151 described above. In this case, it is conceivable to obtain the content information itself from the source (URL) of the content information and extract sentences or characteristic words to be reflected in the question from, for example, the title, summary, header, etc., of the content information.

[0065] Furthermore, although the above describes the case where text is added to the original question, characteristic words may also be added to the original question. In this case, the processing unit 11C, for example, assigns a score to each characteristic word and extracts and assigns a predetermined number of characteristic words in descending order of their assigned scores.

[0066] Figure 6 shows an example of the score assigned to the characteristic word according to the first embodiment.

[0067] As shown in Figure 6, a score is assigned to the feature words extracted from the content information. For example, TF-IDF (Term Frequency-Inverse Document Frequency) is used to assign this score. TF-IDF is a statistical measure (numerical value) intended to reflect how important a particular word is. In this case, the processing unit 11C extracts one or two feature words in descending order of their TF-IDF scores. Alternatively, a threshold may be set in advance for the scores, and feature words whose scores are above the threshold may be extracted. In the example in Figure 6, "ordinary deposit account" and "comprehensive account" are extracted as feature words to be added. The processing unit 11C then adds the extracted feature words to the question sentence, for example, "I want to know how to verify my identity for an ordinary deposit account and a comprehensive account." In other words, the initial abstract question sentence, "I want to know how to verify my identity," is automatically processed into the more specific question sentence, "I want to know how to verify my identity for an ordinary deposit account and a comprehensive account."

[0068] Next, with reference to Figure 7, the operation of the natural language search device 10 according to the first embodiment will be described.

[0069] Figure 7 is a flowchart showing an example of the processing flow by the natural language search program 15A according to the first embodiment.

[0070] First, the natural language search program 15A is started by the CPU 11 of the natural language search device 10 and executes the following steps. In this example, as shown in Figure 5 above, it is assumed that the user operates the terminal device 30 to access content information (for example, manual: opening a regular savings account) from the terminal device 30 and refers to the content information.

[0071] In step S101 of Figure 7, the CPU 11 displays the natural language search screen 40 shown in Figure 5 above on the terminal device 30, as an example, in accordance with the user's operation of the terminal device 30, and accepts the input of a question sentence from the natural language search screen 40.

[0072] In step S102, the CPU 11 receives a search execution command, for example, by pressing the "Execute Search" button on the natural language search screen 40 shown in Figure 5 above.

[0073] In step S103, the CPU 11 obtains the source (URL) of the content information that the user is currently viewing from the terminal device 30.

[0074] In step S104, the CPU 11, based on the source (URL) of the content information obtained in step S103, refers to the supplementary text table 151 shown in Figure 4 above as an example, extracts the text to be added, and processes the question by adding the extracted text. As a result, the initial abstract question, "I want to know how to verify my identity," is automatically processed into a more specific question, "I want to know how to verify my identity when opening a regular savings account."

[0075] In step S105, the CPU 11 performs a natural language search on the processed question text that was prepared in step S104.

[0076] In step S106, the CPU 11 outputs the search results of the natural language search performed in step S105 to the terminal device 30, and the series of processes by the natural language search program 15A is completed.

[0077] Thus, according to this embodiment, sentences or characteristic words to be reflected in the question are extracted from the content information that the user was referring to when issuing a search command. As a result, the question entered in natural language search is made more concrete, and more accurate search results are obtained.

[0078] [Second Embodiment] The first embodiment described above describes a method of processing question text using content information. The second embodiment describes a method of processing question text using workflow information.

[0079] Figure 8 is a block diagram showing an example of the functional configuration of the natural language search device 10A according to the second embodiment.

[0080] As shown in Figure 8, the CPU 11 of the natural language search device 10A according to this embodiment functions as a receiving unit 11A, an acquisition unit 11E, a processing unit 11F, and a search unit 11D. The same reference numerals are used for components that are the same as those of the natural language search device 10 described in the first embodiment, and repeated explanations are omitted.

[0081] When the acquisition unit 11E receives an execution instruction for natural language search from the reception unit 11A, it acquires information representing the status of the execution instruction for natural language search. Here, "information representing the status of the execution instruction" is, for example, the workflow information that the user referred to when issuing the execution instruction. The workflow information is associated with, for example, the case and stage (phase) of the workflow information.

[0082] The processing unit 11F processes the question text using the information acquired by the acquisition unit 11E. Specifically, the processing unit 11F extracts sentences or characteristic words to be reflected in the question text from the workflow information that the user referred to when issuing an execution instruction, and processes the question text using the extracted sentences or characteristic words. The processing here may, as described above, involve, for example, adding the extracted sentences or characteristic words to the question text, or generating a new question text from the extracted sentences or characteristic words and the question text.

[0083] The search unit 11D performs a natural language search on the processed query text obtained by the processing unit 11F, and outputs the search results to the terminal device 30.

[0084] Here, the memory unit 15 stores the supplementary text table 152. The extraction of sentences or characteristic words to be reflected in the question is performed, for example, using the supplementary text table 152.

[0085] Figure 9 shows an example of an additional text table 152 according to the second embodiment.

[0086] In the supplementary text table 152 shown in Figure 9, the cases, stages (phases), and supplementary texts of the workflow information that users can refer to are defined in association. The supplementary text is defined as a characteristic sentence that represents the workflow information. Alternatively, characteristic words to be added may be defined in association with the supplementary text. The characteristic words to be added are defined as characteristic words that represent the workflow information.

[0087] Figure 10 is a diagram illustrating the question processing process according to the second embodiment.

[0088] In (S11) of Figure 10, the user operates the terminal device 30 to access workflow information (for example, the case management system: opening a savings account) from the terminal device 30. This workflow information is associated with a case called "savings account" and a stage (phase) called "customer verification".

[0089] In (S12), if the user wants to ask a question while viewing workflow information, the terminal device 30 displays the natural language search screen 40. The user enters a question, for example, "I want to know how to verify my identity," into the natural language search screen 40 and presses the "Execute Search" button.

[0090] In (S13), when the reception unit 11A of the natural language search device 10A receives a search execution instruction along with the question entered via the natural language search screen 40, the acquisition unit 11E acquires the case and stage (phase) of the workflow information that the user is currently viewing from the terminal device 30. Then, the processing unit 11F, based on the acquired case and stage (phase) of the workflow information, refers to the additional sentence table 152 shown in Figure 9 above as an example and extracts sentences to be added. In the example in Figure 10, the sentence to be added is extracted as "Regular deposit, in customer verification,." Then, the processing unit 11F adds the extracted sentence to the question, for example, "Regular deposit, in customer verification, I want to know the method of identity verification." In other words, the initial abstract question, "I want to know the method of identity verification," is automatically processed into the more specific question, "Regular deposit, in customer verification, I want to know the method of identity verification."

[0091] In (S14), the search unit 11D of the natural language search device 10A performs a natural language search on the processed question text obtained in (S13) and displays the search results on the natural language search screen 40 of the terminal device 30.

[0092] In addition, the above describes the case where text is added to the original question, but characteristic words may also be added to the original question. In this case, the processing unit 11F, similar to the content information described above, assigns a score to each characteristic word, and extracts and assigns a predetermined number of characteristic words in descending order of their assigned scores.

[0093] Next, with reference to Figure 11, the operation of the natural language search device 10A according to the second embodiment will be described.

[0094] Figure 11 is a flowchart showing an example of the processing flow by the natural language search program 15A according to the second embodiment.

[0095] First, the natural language search program 15A is launched by the CPU 11 of the natural language search device 10A and executes the following steps. In this example, as shown in Figure 10 above, it is assumed that the user operates the terminal device 30 to access workflow information (for example, case management system: ordinary deposit account opening) from the terminal device 30 and refers to the workflow information.

[0096] In step S111 of Figure 11, the CPU 11, in accordance with the user's operation of the terminal device 30, displays the natural language search screen 40 shown in Figure 10 above as an example on the terminal device 30, and accepts the input of a question sentence from the natural language search screen 40.

[0097] In step S112, the CPU 11 receives a search execution command, for example, by pressing the "Execute Search" button on the natural language search screen 40 shown in Figure 10 above.

[0098] In step S113, the CPU 11 retrieves the case and stage (phase) of the workflow information that the user is currently viewing from the terminal device 30.

[0099] In step S114, the CPU 11, based on the case and stage (phase) of the workflow information obtained in step S113, refers to the supplementary text table 152 shown in Figure 9 above as an example, extracts the text to be added, and processes the question by adding the extracted text. As a result, for example, the initial abstract question "I want to know how to verify my identity" is automatically processed into a more specific question "I want to know how to verify the identity of a customer in a regular savings account."

[0100] In step S115, the CPU 11 performs a natural language search on the processed question text that was processed in step S114.

[0101] In step S116, the CPU 11 outputs the search results of the natural language search performed in step S115 to the terminal device 30, and the series of processes by the natural language search program 15A is completed.

[0102] Thus, according to this embodiment, sentences or characteristic words to be reflected in the query are extracted from the workflow information that the user referred to when issuing a search command. As a result, the query entered in natural language search is made more concrete, and more accurate search results are obtained.

[0103] Next, with reference to Figures 12 and 13, the screen transitions of the natural language search screen 40 displayed on the terminal device 30 will be explained.

[0104] Figure 12 shows an example of the screen transitions of the natural language search screen 40 according to this embodiment.

[0105] In (S21) of Figure 12, the user enters a question, for example, "I want to know how to verify my identity," into the natural language search screen 40 and presses the "Execute Search" button.

[0106] In (S22), a natural language search is performed on the question "I want to know how to verify my identity" entered above, and the search results are displayed on the natural language search screen 40. Here, "Identity verification when obtaining a loan," "Identity verification when creating a card," and "Identity verification for a housing loan" are obtained as search results. If the desired information is found in the search results, the user presses the "Yes" button; if the desired information is not found in the search results, the user presses the "No" button. In the example in Figure 12, the "No" button is pressed.

[0107] In (S23), the original question is modified with additional text and a new search is performed. The new search results are then displayed on the natural language search screen 40. For example, the search results might read, "The question has been modified to 'I want to know the method of identity verification when opening a savings account,' and a new search has been performed." The search results might then show "Identity verification when opening a savings account."

[0108] Figure 13 shows another example of the screen transitions in the natural language search screen 40 according to this embodiment.

[0109] In Figure 13 (S31), the user enters a question, for example, "I want to know how to verify my identity," into the natural language search screen 40 and presses the "Execute Search" button.

[0110] In (S32), candidate questions (query candidates) obtained by adding text to the original question are displayed on the natural language search screen 40. As candidate questions, for example, "I want to know how to verify my identity when opening a regular savings account," "I want to know how to verify my identity when opening a time deposit account," and "Search with the original question" are displayed for the user to select. In the example in Figure 13, "I want to know how to verify my identity when opening a regular savings account" is selected.

[0111] In (S33), a natural language search is performed using the question selected above, and the search results are displayed on the natural language search screen 40. For example, with the question, "I want to know how to verify my identity when opening a savings account," the search result will be "Identity verification when opening a savings account."

[0112] [Third Embodiment] In the third embodiment, we will describe a method in which the level of abstraction of the initial question is determined, and the question is modified only if the level of abstraction is high.

[0113] Figure 14 is a block diagram showing an example of the functional configuration of the natural language search device 10B according to the third embodiment.

[0114] As shown in Figure 14, the CPU 11 of the natural language search device 10B according to this embodiment functions as a reception unit 11A, a determination unit 11G, an acquisition unit 11H, a processing unit 11J, and a search unit 11D. The same reference numerals are used for components that are the same as those of the natural language search device 10 described in the first embodiment, and repeated explanations are omitted.

[0115] The determination unit 11G determines whether the level of abstraction of the question received by the reception unit 11A is above a threshold. Here, the level of abstraction is determined, for example, by performing a natural language search based on the question and determining the number of search results obtained by the natural language search. A large number of search results suggests that the question is abstract. Therefore, it can be said that the more search results there are, the higher the level of abstraction. Alternatively, the level of abstraction may be determined by performing a natural language search based on the question and determining the number of predetermined keywords included in the search results obtained by the natural language search. Here, keywords are words that are highly relevant to the content information that the user is referring to. A large number of keywords suggests that the question is specific. Therefore, it can be said that the fewer keywords there are, the higher the level of abstraction. If the level of abstraction is expressed on a scale of 1 to 10, for example, the threshold can be set in advance as "7".

[0116] The acquisition unit 11H acquires information representing the status of the natural language search execution command when the determination unit 11G determines that the level of abstraction is above a threshold, that is, that the level of abstraction is relatively high. Here, "information representing the status of the execution command" is, for example, the content information that the user was referring to when issuing the execution command. The content information is associated with, for example, a URL representing the source of the content information.

[0117] The processing unit 11J processes the question text using the information acquired by the acquisition unit 11H. Specifically, the processing unit 11J extracts sentences or characteristic words to be reflected in the question text from the content information that the user referred to when issuing the execution instruction, and processes the question text using the extracted sentences or characteristic words. This processing may involve, for example, adding the extracted sentences or characteristic words to the question text, or generating a new question text from the extracted sentences or characteristic words and the original question text. In other words, the processing unit 11J processes the question text using content information only when the initial question text is relatively abstract.

[0118] Next, with reference to Figure 15, the operation of the natural language search device 10B according to the third embodiment will be described.

[0119] Figure 15 is a flowchart showing an example of the processing flow by the natural language search program 15A according to the third embodiment.

[0120] First, the natural language search program 15A is launched by the CPU 11 of the natural language search device 10B and executes the following steps. In this example, as shown in Figure 5 above, it is assumed that the user operates the terminal device 30 to access content information (for example, manual: opening a regular savings account) from the terminal device 30 and refers to the content information.

[0121] In step S121 of Figure 15, the CPU 11 displays the natural language search screen 40 shown in Figure 5 above on the terminal device 30, as an example, in accordance with the user's operation of the terminal device 30, and accepts the input of a question sentence from the natural language search screen 40.

[0122] In step S122, the CPU 11 receives a search execution instruction, for example, by pressing the "Execute Search" button on the natural language search screen 40 shown in Figure 5 above.

[0123] In step S123, the CPU 11 determines whether the level of abstraction of the question is above a threshold. If it determines that the level of abstraction of the question is above the threshold (positive determination), it proceeds to step S124. If it determines that the level of abstraction of the question is below the threshold (negative determination), it proceeds to step S127.

[0124] In step S124, the CPU 11 obtains the source (URL) of the content information that the user is currently viewing from the terminal device 30.

[0125] In step S125, the CPU 11, based on the source (URL) of the content information obtained in step S124, refers to the supplementary text table 151 shown in Figure 4 above as an example, extracts the text to be added, and processes the question by adding the extracted text. As a result, the initial abstract question, "I want to know how to verify my identity," is automatically processed into a more specific question, "I want to know how to verify my identity when opening a regular savings account."

[0126] In step S126, the CPU 11 performs a natural language search on the processed question text prepared in step S125, and then proceeds to step S128.

[0127] Meanwhile, in step S127, the CPU 11 performs a natural language search on the initial question received in step S121, and then proceeds to step S128.

[0128] In step S128, the CPU 11 outputs the search results of the natural language search performed in step S126 or step S127 to the terminal device 30, and the series of processes by the natural language search program 15A ends.

[0129] Thus, according to this embodiment, the level of abstraction of the initial question is determined, and the question is only modified if the level of abstraction is high. Therefore, only questions with a high level of abstraction are selectively made more concrete.

[0130] In each of the above embodiments, the term "processor" refers to a processor in a broad sense, and includes general-purpose processors (e.g., CPU: Central Processing Unit, etc.) and dedicated processors (e.g., GPU: Graphics Processing Unit, ASIC: Application Specific Integrated Circuit, FPGA: Field Programmable Gate Array, programmable logic device, etc.).

[0131] Furthermore, the processor operations in each of the above embodiments may not be performed by a single processor, but may also be performed by multiple processors located in physically separate locations working together. Also, the order of the processor operations is not limited to the order described in each of the above embodiments, but may be changed as appropriate.

[0132] The above description illustrates a natural language search device as an example of an information processing device according to the embodiment. The embodiment may take the form of a program that causes a computer to execute the functions of each part of the information processing device. The embodiment may also take the form of a non-temporary storage medium that is readable by a computer and stores these programs.

[0133] Furthermore, the configuration of the information processing device described in the above embodiment is merely an example, and may be modified as needed without departing from the main purpose.

[0134] Furthermore, the program processing flow described in the above embodiment is just one example, and unnecessary steps may be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0135] Furthermore, although the above embodiment describes a case in which the process according to the embodiment is realized by a software configuration using a computer by executing a program, the embodiment is not limited to this. The embodiment may also be realized by a hardware configuration or a combination of a hardware configuration and a software configuration.

[0136] The following is further disclosed regarding the embodiments described above.

[0137] The information processing device relating to (((1))) includes a processor, the processor receiving input of a natural language search query from a user, receiving an instruction to execute a natural language search from the user, acquiring information indicating the status of the natural language search execution instruction, and processing the query using the acquired information.

[0138] The information processing device relating to (((2))) is such that, in the information processing device relating to (((1))) the information representing the situation in which the execution instruction was given by the user is the content information that the user referred to when giving the execution instruction, and the processor extracts from the content information the sentence or characteristic words to be reflected in the question sentence.

[0139] The information processing device relating to (((3))) is an information processing device relating to (((2))) in which the processing is to add the extracted text or characteristic words to the question text.

[0140] The information processing device relating to (((4))) is an information processing device relating to (((2))) in which the processing is to generate a new question sentence from the extracted text or characteristic words and the question sentence.

[0141] The information processing device relating to (((5))) is an information processing device relating to any one of (((2))) to (((4))) in which the processor assigns a score to each of the feature words and extracts a predetermined number of the feature words in descending order of the assigned scores.

[0142] The information processing device relating to (((6))) is such that, in the information processing device relating to (((1))) the information representing the situation in which the execution instruction was given by the user is the workflow information that the user referred to when giving the execution instruction, and the processor extracts from the workflow information the sentences or characteristic words to be reflected in the question sentence.

[0143] The information processing device related to (((7))) extracts the text or characteristic words to be reflected in the question from the information representing the case and stage in the workflow information in the information processing device related to (((6))).

[0144] The information processing device relating to (((8))) is an information processing device relating to any one of (((1))) to (((7))) in which the processor performs control to display the processed question text.

[0145] The information processing device relating to (((9))) is an information processing device relating to (((8))) in which the processor controls the display of multiple processed question sentences and accepts a selection from the user.

[0146] The information processing device relating to (((10))) is an information processing device relating to any one of (((1))) to (((9))) in which the processor determines the level of abstraction of the question statement that has been received as input, and if the level of abstraction is equal to or greater than a threshold, it processes the question statement.

[0147] The information processing device relating to (((11))) is determined in the information processing device relating to (((10))) by performing the natural language search based on the question and determining the level of abstraction based on the number of search results obtained by the natural language search.

[0148] The information processing device relating to (((12))) is determined in the information processing device relating to (((10))) by performing the natural language search based on the question statement and determining the level of abstraction based on the number of predetermined keywords included in the search results obtained by the natural language search.

[0149] The information processing program related to (((13))) receives input of a natural language search query from the user, receives an instruction to execute the natural language search from the user, obtains information indicating that the instruction to execute the natural language search has been given, and causes the computer to perform processing of the query using the obtained information.

[0150] According to (((1))) and (((13))), this has the effect of making the question entered in natural language search more concrete.

[0151] According to (((2))), this has the effect of making the question more specific and accurate compared to when the user uses content information that they have not referred to when giving the execution instruction.

[0152] According to (((3))), this method has the effect of making it easier to concretize the question compared to generating a new question from text or characteristic words.

[0153] According to (((4))), this method has the effect of making the question more structured and specific compared to adding sentences or characteristic words to the question.

[0154] According to (((5))), this has the effect of being able to extract characteristic words in order of highest score.

[0155] According to (((6))), this has the effect of making the question more specific and precise compared to when the user uses workflow information that they have not referred to when issuing an execution instruction.

[0156] According to (((7))), this has the effect of allowing for more precise and specific question wording compared to not considering the case and stage in the workflow information.

[0157] According to (((8))), this has the effect of allowing the user to understand the processed question text.

[0158] According to (((9))), this has the effect of allowing the user to select the desired question from among several processed question sentences.

[0159] According to (((10))), this has the effect of selectively processing only highly abstract question sentences.

[0160] According to (((11))), this has the effect of allowing us to determine the level of abstraction based on the number of search results.

[0161] According to (((12))), this has the effect of allowing us to determine the level of abstraction based on the number of keywords. [Explanation of symbols]

[0162] 10, 10A, 10B Natural Text Search Device 11 CPU 11A Reception Desk 11B, 11E, 11H Acquisition Department 11C, 11F, 11J processing section 11D Search Unit 11G judgment section 12 ROM 13 RAM 14 I / O 15 Storage section 15A Natural Text Search Program 16 Display 17 Control section 18 Communications Department 30 Terminal devices 40 Natural Text Search Screen 100 Natural Text Search Systems

Claims

1. Equipped with a processor, The aforementioned processor, The system accepts natural language search query input from the user. The user has given an instruction to perform a natural language search. Obtain information representing the status of the execution instruction for the aforementioned natural language search, Using the information obtained above, the question sentence is processed as follows: The information representing the circumstances under which the execution instruction was given by the user is the workflow information that the user referenced when giving the execution instruction. The aforementioned processor, From the aforementioned workflow information, extract the sentences or characteristic words to be reflected in the aforementioned question. Information processing device.

2. The sentences or characteristic words to be reflected in the aforementioned question are extracted from the information representing the case and stage in the workflow information. The information processing apparatus according to claim 1.

3. The processor performs control to display the processed question text. The information processing apparatus according to claim 1.

4. The processor controls the display of multiple processed question sentences and accepts a selection from the user. The information processing apparatus according to claim 3.

5. comprising a processor, The aforementioned processor, The system accepts natural language search query input from the user. The user has given an instruction to perform a natural language search. Obtain information representing the status of the execution instruction for the aforementioned natural language search, Using the information obtained above, the question sentence is processed as follows: The aforementioned processor, The abstraction level of the question received as input is determined, and if the abstraction level is above a threshold, the question is processed. The aforementioned level of abstraction is determined by performing a natural language search based on the aforementioned question and by determining the number of predetermined keywords included in the search results obtained by the natural language search. Information processing device.

6. The system accepts natural language search query input from the user. The user has given an instruction to perform a natural language search. Obtain information representing the status of the execution instruction for the aforementioned natural language search, Using the information obtained above, the question sentence is processed as follows: The information representing the circumstances under which the execution instruction was given by the user is the workflow information that the user referenced when giving the execution instruction. From the aforementioned workflow information, extract the sentences or characteristic words to be reflected in the aforementioned question. An information processing program designed to be executed by a computer.

7. The system accepts input of a natural language search query from the user, The user has given an instruction to perform a natural language search. Obtain information representing the status of the execution instruction for the aforementioned natural language search, Using the information obtained above, the question sentence is processed as follows: The abstraction level of the question received as input is determined, and if the abstraction level is above a threshold, the question is processed. The aforementioned level of abstraction is determined based on the number of predetermined keywords included in the search results obtained by performing the aforementioned natural language search based on the aforementioned question statement. An information processing program designed to be executed by a computer.

Images