Information processing system, information processing method, and program

Abstract

To provide a mechanism to obtain more appropriate responses in continuous dialogue-type question-response processing.SOLUTION: An information processing system includes question acquisition means for acquiring a question, retrieval means for retrieving information using a word specified from the question acquired by the question acquisition means, and output means for outputting a response generated using the information retrieved by the retrieval means, and the acquired question, wherein when a question is further acquired after output of the response by the output means, the retrieval means retrieves information, using a word satisfying a predetermined condition, and a word specified from a newly acquired question, among words specified from the acquired questions.SELECTED DRAWING: Figure 1

Description

Technical Field

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

Background Art

[0002] In recent years, in companies and the like, the introduction of full-text search systems such as enterprise search has been increasing in order to find necessary information from a vast amount of accumulated electronic documents. In addition, by introducing a question-and-answer system that performs information search and answer through dialogue, it has become easier to reach the necessary information.

[0003] Patent Document 1 discloses a method of determining a search word set and a question type from an input question sentence, and generating an answer using the basis information obtained by searching a document set serving as a knowledge source according to this.

[0004] Patent Document 2 discloses a method of managing past question answers and obtained basis information, and performing an answer in consideration of the past question answers and basis information when the next question is made.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Patent Document 2

Disclosure of the Invention

Problems to be Solved by the Invention

[0006] However, Patent Document 1 has a problem that it cannot answer questions in continuous dialogue. For example, assuming that it can answer a question "What is the address of the Osaka office?", when the next question "What about the reception?" is asked, the search word is only "reception", and the information serving as the source of the answer cannot be limited to the "Osaka office".

[0007] In Patent Document 2, although it is possible to support continuous dialogue, it is necessary to continuously manage the documentary evidence (search results) regarding "Osaka Office + Address" and "Reception", and since the information to be managed becomes enormous, there is a problem that the memory area and subsequent processing become inefficient.

[0008] Therefore, it is desired to realize continuous dialogue and question-and-answer that can efficiently manage and process the memory area.

[0009] Therefore, an object of the present invention is to provide a mechanism for obtaining a more appropriate answer in a question-and-answer process in a continuous dialogue format.

Means for Solving the Problems

[0010] The present invention includes a question acquisition means for acquiring a question, a search means for performing an information search using words specified from the question acquired by the question acquisition means, an output means for outputting an answer generated using the information retrieved by the search means and the acquired question, and when a question is further acquired after the output of the answer by the output means, the search means uses words that satisfy a predetermined condition among the words specified from the questions acquired so far and words specified from the newly acquired question to perform an information search.

Effects of the Invention

[0011] According to the present invention, it is possible to provide a mechanism for obtaining a more appropriate answer in a question-and-answer process in a continuous dialogue format.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Mode for Carrying Out the Invention

[0013] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[0014] FIG. 1 is a diagram showing an example of the system configuration of a question-and-answer device according to an embodiment of the present invention.

[0015] As shown in FIG. 1, the question-and-answer device 100 is configured to be connected via the user terminal 110 and the network 120.

[0016] The question-and-answer device 100 presents a response sentence and the document that is the basis for the response to the question sentence acquired from the user terminal 110.

[0017] The user terminal 110 sends the question text input by the user to the question answering device 100 and displays the answer text and the basis information (including information such as the title of the document in which the basis information is described) returned by the question answering device 100. Specifically, the user terminal 110 uses a personal computer (such as a notebook PC or a desktop PC), a tablet terminal, a smartphone, etc., but is not limited thereto. Also, the configuration of various terminals or devices connected to the network 120 in FIG. 1 is an example, and it goes without saying that there are various configuration examples according to the use and purpose.

[0018] FIG. 2 is a block diagram showing an example of the hardware configuration of the question answering device 100 and the user terminal 110 in an embodiment of the present invention.

[0019] As shown in FIG. 2, an information processing device has a CPU (Central Processing Unit) 201, a RAM (Random Access Memory) 202, a ROM (Read Only Memory) 203, an input controller 205, a video controller 206, a memory controller 207, and a communication I / F controller 208 connected via a system bus 204.

[0020] The CPU 201 comprehensively controls each device and controller connected to the system bus 204.

[0021] The RAM 202 functions as the main memory, work area, etc. of the CPU 201. The CPU 201 loads programs and the like necessary for executing processing from the ROM 203 or the external memory 211 into the RAM 202 and realizes various operations by executing the loaded programs.

[0022] The ROM 203 or the external memory 211 stores a BIOS (Basic Input / Output System), an OS (Operating System), which are control programs executed by the CPU 201, a computer-readable executable program for implementing the present information processing method, and various necessary data (including data tables).

[0023] The input controller 205 controls inputs from input devices such as a keyboard 209 and a pointing device such as a mouse (not shown). When the input device is a touch panel, it is assumed that the user can give various instructions by pressing (touching with a finger or the like) in accordance with icons, cursors, or buttons displayed on the touch panel.

[0024] Also, the touch panel may be a touch panel such as a multi-touch screen that can detect positions touched by multiple fingers.

[0025] The video controller 206 controls the display to an external output device such as a display 210. The display includes the display of a notebook personal computer integrated with the main body. Note that the external output device is not limited to a display and may be, for example, a projector. Also, for a device capable of receiving the above-described touch operation, an input device is also provided.

[0026] Note that the video controller 206 can control a video memory (VRAM) for performing display control, and can use a part of the RAM 202 as a video memory area, or can separately provide a dedicated video memory.

[0027] The memory controller 207 controls access to the external memory 211. As the external memory, an external storage device (hard disk) that stores a boot program, various applications, font data, user files, edited files, and various data, a flexible disk (FD), or a compact flash (registered trademark) memory connected via an adapter to a PCMCIA card slot can be used.

[0028] The communication I / F controller 208 is connected to and communicates with an external device via a network, and executes communication control processing on the network. For example, communication using TCP / IP, a telephone line such as ISDN, and communication using a 3G line of a mobile phone are possible.

[0029] Furthermore, the CPU 201 enables display on the display 210 by executing an outline font expansion (rasterization) process on, for example, the display information area in the RAM 202. Also, the CPU 201 enables user instructions using a mouse cursor (not shown) on the display 210.

[0030] Next, FIG. 3 is a diagram showing an example of the functional configuration of the question-and-answer device 100 and the user terminal 110. The functions provided by each functional unit will be described in the description of the flowchart in FIG. 4 and the like.

[0031] First, the overall picture of the processing of the present invention will be described with reference to FIG. 13. In the present invention, for example, in a portal site for employees, a scenario is assumed where the employee searches for information (documents) within the site. For example, the user enters the first question "What is the address of the Osaka office?" in the question input form within the site. When search terms are extracted from the user's question sentence, for example, "Osaka", "office", "address", etc. are extracted in this question. Using the extracted search terms, domain knowledge (information within the employee-oriented site that serves as the basis for generating answer sentences (information on web pages and published document files, etc.)) is searched. A prompt including the searched domain knowledge and the question sentence entered by the user is given to the generation AI to create an answer sentence, the created answer sentence and the basis information referred to in creating the answer sentence are obtained, and the obtained information is notified to the user. As a result, the user can efficiently collect information within the site. Further, subject words (words that are the center of the conversation, which are "Osaka" and "address" in this example) are extracted from the user's question sentence, the answer sentence, and the domain knowledge, and the processing for the first question is completed. Subsequently, assuming that in the second question, for example, the user asks "What about the reception?" In the second question, in addition to the search term "reception" extracted from the question sentence, domain knowledge is searched including the subject words "Osaka" and "address" obtained in the first question. The subsequent processing is the same as that for the first question. By performing a search including these subject words, an answer to the question in a continuous conversation is realized. Details of these processes will be described later with reference to FIGS. 4 to 12.

[0032] (First Embodiment) Next, with reference to the flowchart of FIG. 4, the processing executed by the question-and-answer processing unit 302 in the first embodiment of the present invention will be described.

[0033] The flowchart of FIG. 4 is a process executed by the CPU 201 of the question-and-answer device 100 by reading a predetermined control program, and is a flowchart showing a process in which the question-and-answer processing unit 302 generates an answer sentence for a question sentence in response to a request from the dialogue management processing unit 301.

[0034] The dialogue management processing unit 301 realizes the dialogue (question and answer) with the user by repeating the processing of the flowchart in FIG. 4.

[0035] The dialogue management processing unit 301 manages by associating the question sentence, the domain knowledge acquired in step S405 described later, the answer sentence generated in step S406, and the topic words to be generated in step S407 as the history of the dialogue (question sentence and answer sentence). The number of dialogues to be managed is not particularly specified, and in this embodiment, it is sufficient if there is a previous dialogue.

[0036] In step S401, the question and answer processing unit 302 receives the question sentence input by the user and the list of previous topic words from the dialogue management processing unit 301. In the case of the first question, the list of topic words in the dialogue is empty.

[0037] In step S402, the question and answer processing unit 302 extracts a list of search words from the question sentence.

[0038] In step S403, if there are topic words in the list of previous topic words acquired in step S401, the question and answer processing unit 302 transfers the processing to step S404. If there are no topic words, the processing is transferred to step S405.

[0039] In step S404, the question and answer processing unit 302 adds the topic words to the list of search words extracted in step S402.

[0040] In step S405, the question and answer processing unit 302 uses the search processing unit 303 to acquire the domain knowledge related to the search word list from the domain knowledge storage table 501 in the domain knowledge storage area 304. The method for acquiring the related domain knowledge is not particularly specified, but in this embodiment, full-text search is used. Also, when there are multiple related domain knowledge, the number of items to be acquired may be limited by judging based on the score of the full-text search or the like.

[0041] Domain knowledge, in the context of the present invention, refers to information that serves as the basis for answering questions from users. For example, when applying the present invention to a search system for an intranet site for company employees (including a question-and-answer system in an interactive (chat) format), the information posted on the intranet site for employees (such as information on web pages and posted document files) is the domain knowledge. Domain knowledge is registered and managed, for example, by registering information on sites obtained by performing a forward match search on URLs and updating it at regular intervals, or by registering data in advance by an administrator.

[0042] FIG. 7 is a diagram showing the content of domain knowledge when information on an intranet site and a corporate site is used as domain knowledge.

[0043] In step S406, the question-and-answer processing unit 302 uses the answer generation processing unit 305 to generate an answer sentence using the domain knowledge acquired in step S405 and the question sentence. Although the method for generating the answer sentence is not particularly specified, in this embodiment, an AI for generation is used.

[0044] In step S407, the question-and-answer processing unit 302 extracts a subject word from the question sentence acquired in step S401, the domain knowledge acquired in step S405, and the answer sentence generated in step S406. The subject word extraction process will be described later.

[0045] In step S408, the answer sentence for the question generated in S406 is displayed to the user. It is also possible to display the basis information (the domain knowledge acquired in step S405) together with the answer sentence.

[0046] Next, the subject word extraction process in step S407 will be described using the flowchart of FIG. 6.

[0047] In step S601, the question-and-answer processing unit 302 initializes a temporary area and a subject word list.

[0048] In step S602, the question-and-answer processing unit 302 saves the list of search terms (if there was no previous topic term, the search terms obtained in S402) obtained in step S404 in a temporary area.

[0049] In step S603, the question-and-answer processing unit 302 saves the list of words extracted from the domain knowledge obtained in step S405 in a temporary area. Further, words are also extracted from the answer sentence generated in step S406 and saved in the temporary area.

[0050] In step S604, the question-and-answer processing unit 302 totals and evaluates the words in the temporary area based on the frequency of occurrence.

[0051] In step S605, for each word totaled and evaluated in step S604, the question-and-answer processing unit 302 starts the loop processing until step S608.

[0052] In step S606, the question-and-answer processing unit 302 determines whether the evaluation value of the word is greater than or equal to a reference value. If it is greater than or equal to the reference value, the process proceeds to step S607. If it is less than the reference value, the process proceeds to step S608.

[0053] In step S607, the question-and-answer processing unit 302 adds the words that meet the reference value to the topic word list.

[0054] In step S608, if there are still words to be processed, the question-and-answer processing unit 302 performs the loop processing from step S605. If there are no words to be processed, the process ends.

[0055] Next, as a specific example of the question-and-answer processing shown in the flowchart of FIG. 4, the case where the question-and-answer processing is performed on the domain knowledge table 501 of FIG. 7 will be described.

[0056] In step S401, the question-and-answer processing unit 302 receives the question sentence "What is the address of the Osaka office?" input by the user from the dialogue management processing unit 301 and the list of previous topic words. Since this is the first question, the list of topic words is empty.

[0057] In step S402, the question-and-answer processing unit 302 extracts a list of search terms from the question sentence. In this embodiment, the acquisition of search terms is realized by extracting only content words from the result of morphological analysis of the question sentence. For the question sentence "What is the address of the Osaka office?", a list consisting of the three words "Osaka", "office", and "address" is obtained.

[0058] In step S403, since there is no subject term in the subject term list acquired in step S401, the question-and-answer processing unit 302 transfers the process to step S405.

[0059] In step S405, the question-and-answer processing unit 302 uses the search processing unit 303 to obtain, from the domain knowledge table 501 in the domain knowledge storage area 304, as domain knowledge related to the search term list ("Osaka", "office", "address"), the top three domain knowledge 701 to 703 of the full-text search results.

[0060] In step S406, the question-and-answer processing unit 302 provides the domain knowledge 701 to 703 acquired in step S405 and the question sentence to the answer generation processing unit 305 to generate an answer sentence. FIG. 8 shows an overview of the answer sentence generation process using the generation AI in this embodiment.

[0061] The answer sentence generation process using the generation AI will be described with reference to FIG. 8.

[0062] The text of the top three domain knowledge 701 to 703 of the full-text search results acquired in step S405 (C in FIG. 8) and the question sentence input by the user (Q in FIG. 8) are input into a template to create a prompt as shown in FIG. 8 and given to the generation AI. In S406, the generation AI generates an answer sentence 801 from this prompt.

[0063] In step S407, the question-and-answer processing unit 302 calls the subject term extraction process for the question sentence acquired in step S401, the domain knowledge acquired in step S405, and the answer sentence generated in step S406.

[0064] The details of the process of step S407 will be described with reference to the flowchart of FIG. 6.

[0065] In step S601, the question-and-answer processing unit 302 initializes by emptying the temporary area and the subject word list. FIG. 9 shows the state of the temporary area in a series of processes.

[0066] In step S602, the question-and-answer processing unit 302 saves the list of search words (the search words obtained in step S404; if there was no previous subject word, the search words obtained in step S402) in the temporary area (901).

[0067] In step S603, the question-and-answer processing unit 302 saves the list of words extracted from the domain knowledge obtained in step S405 in the temporary area (902 - 904). Further, words are also extracted from the answer sentence generated in step S406 and saved in the temporary area (905).

[0068] In step S604, the question-and-answer processing unit 302 totals and evaluates the words in the temporary area based on the frequency of occurrence. In the present embodiment, the frequency of occurrence itself in the temporary area is used as the evaluation value, but the evaluation method is not particularly specified. It may also use tf·idf etc. in the domain knowledge storage table. FIG. 10 shows the result of the total and evaluation.

[0069] In step S605, the question-and-answer processing unit 302 starts the loop process until step S608 for the word 1001 "Osaka".

[0070] Here, the reference value in step S606 is set to 5, which is the maximum value of the evaluation values of the total and evaluation results in the present embodiment. The reference value is not particularly specified. It may be calculated dynamically as in the present embodiment or may be a fixed value.

[0071] In step S606, since the evaluation value 5 of the word 1001 "Osaka" is greater than or equal to the reference value, the question-and-answer processing unit 302 transfers the process to step S607.

[0072] In step S607, the question-and-answer processing unit 302 adds the word 1001 "Osaka" to the subject word list.

[0073] In step S608, since there is a word 1002 "office" to be processed, the repetition process from step S605 is performed.

[0074] In step S606, since the evaluation value 5 of the word 1002 "office" is greater than or equal to the reference value, the process proceeds to step S607.

[0075] In step S607, the question-and-answer processing unit 302 adds the word 1002 "office" to the subject word list.

[0076] In step S608, since there is a word 1003 "address" to be processed, the repetition process from step S605 is performed.

[0077] In step S606, since the evaluation value 4 of the word 1003 "address" is less than the reference value, the process proceeds to step S608.

[0078] In step S608, since there are still words to be processed, the process returns to the repetition process from step S605. As a result of performing the same process for all the remaining words, the words 1001 "Osaka" and 1002 "office" are obtained as the subject word list and the process ends.

[0079] When the process of the flowchart in FIG. 6 ends, the process proceeds to step S408 in FIG. 4.

[0080] In step S408, the answer sentence to the question generated in S406 is displayed to the user. Also, the acquired domain knowledge may be displayed simultaneously as the basis information.

[0081] Fig. 11 shows an example of a screen presenting the response sentence and the basis information (domain knowledge obtained in step S405) to the user. In the question form 1101, the user's question sentence "What is the address of the Osaka office?" is entered. In the response form 1102, the response sentence generated from the domain knowledge generated in S406 and the question sentence is displayed. Along with the response sentence, the domain knowledge obtained in step S405 is displayed as the basis information 1103. The basis information 1103 is in the form of a link, enabling the display of the content of the basis information upon an instruction such as a click. The link may display the content of the domain storage table 501 or may be configured to refer to external information. Also, there may be a function that allows the user to evaluate the response sentence generated from the question sentence. In the additional question form 1104, the user can input the content for which they want to continue asking questions.

[0082] Next, assume that the question sentence "What about reception?" is subsequently entered (question input field 1104 in Fig. 11).

[0083] The dialogue management processing unit 301 calls the question response processing of the question response processing unit 302 for the question sentence "What about reception?" and the subject word list ("Osaka", "office").

[0084] In step S401, the question response processing unit 302 receives the question sentence "What about reception?" input by the user from the dialogue management processing unit 301 and the list of previous subject words ("Osaka", "office").

[0085] In step S402, the question response processing unit 302 extracts a list of search words ("reception") from the question sentence "What about reception?".

[0086] In step S403, since there are subject words in the list of previous subject words obtained in step S401, the question response processing unit 302 transfers the processing to step S404.

[0087] In step S404, the question-and-answer processing unit 302 adds the subject words ("Osaka" "office") to the list of search terms ("reception") extracted in step S402. As a result, the search term list becomes ("reception" "Osaka" "office").

[0088] In step S405, the question-and-answer processing unit 302 uses the search processing unit 303 to obtain, from the domain knowledge table 501 in the domain knowledge storage area 304, the domain knowledge 702, 701, 704 that are the top 3 full-text search results as the domain knowledge related to the search term list ("reception" "Osaka" "office").

[0089] Detailed description of the subsequent processing will be omitted. In this processing, if the search for domain knowledge is executed without including the subject words in the questions after the second time, there is a possibility that the desired information cannot be obtained. For example, suppose the first question is "What is the address of the Osaka office?" and then the second question is "What is the reception?" In the second question, since the search is executed without including the subject words "Osaka" and "office", it becomes a search with only the search term "reception", and there is a possibility that the reception of the "head office" or "Nagoya office" is generated as an answer. On the other hand, by including "Osaka" and "office" from the subject word list in the search term list as in the present invention, it becomes possible to include the domain knowledge 702 necessary for an appropriate answer in the top search results. As described above, since the searched domain knowledge is given to the generation AI as a prompt to generate an answer sentence, the prompt given to the generation AI becomes appropriate because appropriate domain knowledge can be searched. As a result, correct question-and-answer is possible even for a question with the subject omitted (the question "What is the reception?" in this embodiment).

[0090] (Second Embodiment) Next, a second embodiment of the present invention will be described.

[0091] The second embodiment is different from the first embodiment in the answer sentence generation process of step S406 and the domain knowledge given to the subject word extraction process of step S407 in the question-and-answer processing.

[0092] In step S406 of the second embodiment, the question-and-answer processing unit 302 uses the answer generation processing unit 305 to generate an answer sentence using the domain knowledge and the question sentence obtained in step S405, and at the same time selects answer basis information (knowledge used for answer sentence generation among the domain knowledge obtained in step S405).

[0093] FIG. 12 shows an example of answer sentence generation involving selection of basis information. A prompt including the texts of the top three domain knowledge items in the full-text search results obtained in step S405 (C1 to C3 in FIG. 12) and the question sentence input by the user (Q in FIG. 12) is created and given to the generation AI. In S406, the generation AI generates an answer sentence from this prompt. At this time, the domain knowledge used by the generation AI for generating the answer to the question is selected as the answer basis information 1201 and displayed together with the answer sentence.

[0094] Then, in step S407 of the second embodiment, the subject extraction process is performed using only the selected answer basis information as the domain knowledge. In the first embodiment, in S602, the list of search terms obtained in step S402 and the list of words extracted from the domain knowledge obtained in step S405 were stored in a temporary area (902 to 904). Words were extracted and evaluated from these domain knowledge items and the answer sentence to determine the subject. On the other hand, in the second embodiment, only the selected answer basis information is extracted from the domain knowledge obtained in S405, and the subject is determined from the list of words obtained therefrom and the search terms only.

[0095] In this way, by further limiting the domain knowledge obtained by full-text search or the like to only the domain knowledge required for answer sentence generation, more accurate question-and-answer becomes possible.

[0096] As described above, the embodiments have been shown, but the present invention can be implemented, for example, as an embodiment such as a system, apparatus, method, program, or recording medium. Specifically, it may be applied to a system composed of a plurality of devices, or may be applied to an apparatus composed of a single device.

[0097] In addition, the program in the present invention is a program that can be executed by a computer according to the processing methods of the flowcharts shown in FIGS. 4 and 6, and the storage medium of the present invention stores a program that can be executed by a computer according to the processing methods of FIGS. 4 and 6. Note that the program in the present invention may be a program for each processing method of each device in FIGS. 4 and 6.

[0098] As described above, it goes without saying that the object of the present invention can also be achieved by supplying a recording medium recording a program that realizes the functions of the above-described embodiments to a system or device, and causing a computer (or CPU or MPU) of the system or device to read and execute the program stored in the recording medium.

[0099] In this case, the program itself read from the recording medium realizes the novel functions of the present invention, and the recording medium recording the program constitutes the present invention.

[0100] As the recording medium for supplying the program, for example, a flexible disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a DVD-ROM, a magnetic tape, a non-volatile memory card, a ROM, an EEPROM, a silicon disk, etc. can be used.

[0101] In addition, by executing the program read by the computer, not only the functions of the above-described embodiments are realized, but also based on the instructions of the program, an OS (operating system) or the like running on the computer performs part or all of the actual processing, and it goes without saying that the case where the functions of the above-described embodiments are realized by the processing is also included.

[0102] Furthermore, after the program read from the recording medium is written into the memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, based on the instructions of the program code, a CPU or the like provided in the function expansion board or function expansion unit performs part or all of the actual processing, and it goes without saying that cases where the functions of the above-described embodiments are realized by this processing are also included.

[0103] In addition, the present invention may be applied to a system composed of a plurality of devices or to an apparatus composed of a single device. Needless to say, the present invention is also applicable when achieved by supplying a program to a system or an apparatus. In this case, by reading out the recording medium storing the program for achieving the present invention to the system or the apparatus, the system or the apparatus can enjoy the effects of the present invention.

[0104] Furthermore, by downloading and reading out the program for achieving the present invention from a server, a database, etc. on the network by a communication program, the system or the apparatus can enjoy the effects of the present invention. Note that all configurations combining the above-described embodiments and their modifications are also included in the present invention.

Explanation of Reference Numerals

[0105] 100 Question-and-Answer Device 110 User Terminal 120 Network

Claims

1. Question acquisition means for acquiring questions, Search means for performing an information search using words specified from the questions acquired by the question acquisition means, Output means for outputting an answer generated using the information retrieved by the search means and the acquired questions, Comprising: When a question is further acquired after the output of the answer by the output means, the search means uses, among the words specified from the questions acquired so far, the words that satisfy a predetermined condition and the words specified from the newly acquired question to perform an information search. An information processing apparatus characterized by this.

2. When a question is further acquired after the output of the answer by the output means, the search means uses, among the words specified from the questions acquired so far and the words specified from the retrieved information, the words that satisfy a predetermined condition and the words specified from the newly acquired question to perform an information search. The information processing apparatus according to claim 1, characterized by this.

3. When a question is further acquired after the output of the answer by the output means, the search means uses, among the words specified from the questions acquired so far, the words specified from the retrieved information, and the words specified from the output answer, the words that satisfy a predetermined condition and the words specified from the newly acquired question to perform an information search. The information processing apparatus according to claim 1, characterized by this.

4. The output means outputs the information used for generating the answer among the retrieved information in an identifiable manner. The information processing apparatus according to any one of claims 1 to 3, characterized by this.

5. The words specified from the retrieved information used in the information search when a question is further acquired after the output of the answer by the output means are the words specified from the information used for generating the answer among the retrieved information. The information processing apparatus according to claim 4, characterized by this.

6. The predetermined condition is a condition based on the frequency of occurrence of words. The information processing apparatus according to claim 1, characterized by this.

7. The words specified from the questions are independent words. The information processing apparatus according to claim 1, characterized by this.

8. The answer output by the output means is created using a generative AI. The information processing apparatus according to claim 1, characterized by this.

9. The information processing apparatus according to claim 8, wherein the output means outputs an answer created by inputting the question acquired by the question acquisition means and the information retrieved by the retrieval means into a generation AI.

10. The information processing apparatus according to claim 1, wherein the information to be retrieved by the retrieval means is information specified by a URL that satisfies a predetermined condition.

11. A question acquisition means for acquiring a question, A retrieval means for performing information retrieval using words specified from the question acquired by the question acquisition means, An output means for outputting an answer generated using the information retrieved by the retrieval means and the acquired question, Comprising: When a question is further acquired after the output of the answer by the output means, the retrieval means performs information retrieval using words that satisfy a predetermined condition among the words specified from the questions acquired so far and words specified from the newly acquired question. An information processing system characterized by this.

12. A question acquisition step in which a question acquisition means acquires a question, A retrieval step in which a retrieval means performs information retrieval using words specified from the question acquired in the question acquisition step, An output step in which an output means outputs an answer generated using the information retrieved in the retrieval step and the acquired question, Comprising: When a question is further acquired after the output of the answer in the output step, the retrieval step uses words that satisfy a predetermined condition among the words specified from the questions acquired so far and words specified from the newly acquired question. A control method for an information processing apparatus, characterized by performing information retrieval.

13. A program for causing a computer to function as each means according to any one of claims 1 to 10.

Images