Knowledge question-answering system, knowledge question-answering processing method and device thereof, medium and electronic equipment

Abstract

The invention discloses a knowledge question-answering system and a knowledge question-answering processing method and device thereof, a medium and electronic equipment. The knowledge question-answering system comprises a question-answering page, a vector database and a relational database, the vector database only stores vector data corresponding to all questions supported by the knowledge question-answering system, and the relational database stores answers to all the questions supported by the knowledge question-answering system and association relationships between the questions and the answers. Vector data corresponding to all questions supported by the knowledge question-answering system is not stored; the method comprises the steps that when an input question is received through the question and answer page, the input question is vectorized, and an input question vector corresponding to the input question is obtained; searching from the vector database by using the input problem vector to obtain N pieces of vector data with the highest similarity, N being a positive integer; and searching an answer corresponding to the input question from a relational database by using the N pieces of vector data, and displaying the answer. By adopting the scheme, the answering accuracy of the knowledge question-answering system supporting semantic question-answering can be improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent question answering technology, and in particular to a knowledge question answering system and a knowledge question answering processing method, device, medium, and electronic equipment thereof. Background Art

[0002] As enterprises continue to expand and their businesses continue to grow, more and more knowledge is accumulated within them, and the role of knowledge management is becoming increasingly important. Traditional methods that rely on manual knowledge inheritance and management suffer from significant drawbacks such as knowledge loss and inefficient knowledge dissemination, making them unable to meet the needs of modern enterprises.

[0003] The rapid development of information technology, especially artificial intelligence (AI) in recent years, has brought more possibilities to enterprise management. By introducing information-based and automated management tools, enterprises can effectively manage knowledge. This not only improves management efficiency but also reduces knowledge loss caused by human error.

[0004] The knowledge question and answer system is a typical knowledge management tool that can help companies effectively manage knowledge and help employees quickly search for problems, thereby improving the overall operating efficiency of the company.

[0005] However, existing knowledge question answering systems that support semantic question answering provide inaccurate answers. Summary of the Invention

[0006] The problem to be solved by the present invention is to improve the answer accuracy of a knowledge question answering system that supports semantic question answering.

[0007] To solve the above problems, an embodiment of the present invention provides a knowledge question answering processing method for a knowledge question answering system. The knowledge question answering system includes a question answering page, a vector database, and a relational database. The vector database stores vector data corresponding to all questions supported by the knowledge question answering system, and the relational database stores answers to all questions supported by the knowledge question answering system, as well as the associations between questions and answers.

[0008] The method comprises:

[0009] When an input question is received through the question-and-answer page, the input question is vectorized to obtain an input question vector corresponding to the input question;

[0010] Using the input question vector, search the vector database to obtain N vector data with the highest similarity, where N is a positive integer;

[0011] The N vector data are used to search the relational database for an answer corresponding to the input question and display the answer.

[0012] In a possible embodiment, the step of using the N vector data to search a preset relational database for an answer corresponding to the input question and displaying the answer includes:

[0013] Removing vector data that does not reach a similarity threshold from the N vector data;

[0014] The remaining vector data in the N vector data are used to search the relational database for an answer corresponding to the input question and display the answer.

[0015] In a possible embodiment, the using of the remaining vector data in the N vector data to search the relational database for an answer corresponding to the input question and displaying the answer includes:

[0016] When the number of remaining vector data in the N vector data is zero, a question-answering page of the knowledge question-answering system displays no relevant questions;

[0017] When the number of remaining vector data in the N vector data is greater than zero, the answer corresponding to the vector data with the highest similarity score is searched from the relational database, and the question and answer corresponding to the vector data with the highest similarity score are displayed on the question and answer page of the knowledge question and answer system.

[0018] In a possible embodiment, the using of the remaining vector data in the N vector data to search the relational database for an answer corresponding to the input question and displaying the answer further includes:

[0019] When the number of remaining vector data in the N vector data is greater than zero, in addition to displaying the vector data with the highest similarity score in the remaining vector data on the question and answer page of the knowledge question and answer system, questions corresponding to other vector data are also displayed.

[0020] In a possible embodiment, the vector data includes: question content and question numbers supported by the knowledge question answering system.

[0021] In a possible embodiment, using the N vector data to search a preset relational database to obtain an answer corresponding to the input question includes:

[0022] Using the question numbers in the N vector data, the answer corresponding to the input question is searched from the relational database.

[0023] In a possible embodiment, the knowledge question and answer system further includes a maintenance page; and the method further includes:

[0024] When questions and answers are received through the maintenance page, the received questions are numbered, and the received questions and corresponding question numbers are vectorized and stored in the vector database, and the received answers are stored in the relational database, and the association between the question numbers and answers is updated in the relational database.

[0025] An embodiment of the present invention further provides a knowledge question-answering processing device, which is applied to a knowledge question-answering system; the knowledge question-answering system includes a question-answering page, a vector database, and a relational database, wherein the vector database stores vector data corresponding to all questions supported by the knowledge question-answering system, and the relational database stores answers to all questions supported by the knowledge question-answering system, as well as associations between questions and answers;

[0026] The device comprises:

[0027] a vectorization unit adapted to, when receiving an input question through the question-and-answer page, vectorize the input question to obtain an input question vector corresponding to the input question;

[0028] A first search unit is adapted to use the input question vector to search for N vector data with the highest similarity from the vector database;

[0029] The second search unit is adapted to use the N vector data to search the relational database for an answer corresponding to the input question and display the answer.

[0030] In a possible embodiment, the knowledge question and answer system further includes a maintenance page; and the knowledge question and answer processing device further includes:

[0031] The maintenance unit is adapted to, when questions and answers are received through the maintenance page, number the received questions, vectorize the received questions and corresponding question numbers together and store them in the vector database, store the received answers in the relational database, and update the association between the question numbers and the answers in the relational database.

[0032] An embodiment of the present invention further provides a knowledge question answering system, which includes any one of the above-mentioned knowledge question answering processing devices.

[0033] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of any of the above methods.

[0034] An embodiment of the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor executes the steps of any of the above methods when running the computer program.

[0035] Compared with the prior art, the technical solution of the embodiment of the present invention has the following advantages:

[0036] By applying the solution of the present invention, the knowledge question and answer system is not only provided with a relational database, but also with a vector database. The vector database stores vector data corresponding to all questions supported by the knowledge question and answer system. Subsequently, when an input question is received, the system can first search the vector database for the N vector data with the highest similarity based on the input question vector corresponding to the input question, and then use these N vector data to search the relational database for the answer corresponding to the input question and display it. Since all questions supported by the knowledge question and answer system are stored separately in the vector database, when the vector data is obtained based on the search for the input question, the interference of the answer on the question search can be eliminated, thereby improving the search accuracy of the vector data, thereby improving the search accuracy of the answer based on the vector data, and thus improving the answer accuracy of the knowledge question and answer system that supports semantic question and answer. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 1 is a schematic diagram of the structure of a knowledge question answering system according to an embodiment of the present invention;

[0038] Figure 2 is a flow chart of a knowledge question answering method according to an embodiment of the present invention;

[0039] Figure 3 is a flow chart of another knowledge question answering method in an embodiment of the present invention;

[0040] Figure 4 It is a structural diagram of a knowledge question and answer processing device in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] Currently, there are mainly two types of knowledge question answering systems:

[0042] 1) Knowledge question answering system based on relational database and search engine

[0043] Specifically, the knowledge maintainer maintains questions and answers in the system through the knowledge management module, and maintains the relationship between questions and answers. The knowledge question-answering system saves questions, answers, and the relationship between questions and answers in a relational database.

[0044] When receiving an input question, the knowledge question answering system can extract the keywords of the input question, search for relevant questions and answers from the relational database based on the keywords, and present them to the user for selection.

[0045] When the amount of knowledge is large, search performance may be affected. In this case, an enterprise-level search engine (such as Elasticsearch) can be introduced to store the knowledge and index the query content within the search engine. During a search, the search engine retrieves the question based on the keyword, then retrieves the relevant answers from the relational database, and finally presents them to the user for selection.

[0046] 2) Knowledge question answering system based on large model + local knowledge base

[0047] Knowledge maintainers organize knowledge content into documents according to a specific format. The knowledge question-answering system automatically segments the documents and then vectorizes them into a local vector database. This knowledge content may include questions and answers, or just answers. When both are included, the questions and answers are not clearly distinguished, but rather the content is intermingled.

[0048] When receiving an input question, the knowledge question answering system first vectorizes the input question, then matches several semantically closest knowledge contents from the local vector database, and then sends the question and the matched knowledge contents to the big model in a certain prompt word format. Finally, the big model returns the question result.

[0049] Regarding solution 1), regardless of whether or not the enhanced search engine solution is used, the main disadvantage is that this solution is based on keyword processing and search, and cannot achieve semantic-level question-answering effects, that is, it does not support semantic search or question-answering issues.

[0050] Solution 2 is the current mainstream AI-based approach for building local knowledge bases. By incorporating a vector database, this solution supports semantic search and question-answering. Furthermore, since the knowledge question-answering system doesn't distinguish between questions and answers, it directly vectorizes the knowledge content into the local vector database, improving the efficiency of building the local knowledge base while ensuring a certain level of answer accuracy.

[0051] However, because the knowledge Q&A system automatically segments knowledge and then vectorizes and stores the segmented content, questions and answers in the local vector database are processed and stored identically. This interferes with the answer content during question matching, resulting in reduced accuracy. This solution performs poorly, especially in knowledge Q&A scenarios with clear questions and answers.

[0052] To address this issue, the present invention provides a knowledge question-answering system. This system includes a vector database that stores only the vector data corresponding to all questions supported by the knowledge question-answering system. The answers to all supported questions are still stored in a relational database. Subsequent searches for matching questions in the vector database can eliminate interference from answer content in matching question content, thereby improving the accuracy of the knowledge question-answering system's responses.

[0053] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0054] First, before describing the knowledge question answering method in the embodiment of the present invention, a brief description of the knowledge question answering system to which the knowledge question answering method is applied is given:

[0055] Reference Figure 1 The knowledge question answering system 10 includes: a question and answer page 11, a vector database 12, a relational database 13, and a knowledge question answering processing device 14. The vector database 12 stores vector data corresponding to all questions supported by the knowledge question answering system, and the relational database 23 stores answers to all questions supported by the knowledge question answering system, as well as the association between questions and answers.

[0056] In a specific implementation, the vector database 12 stores vector data corresponding to all questions supported by the knowledge question answering system, but does not store answers to all questions supported by the knowledge question answering system. The vector data may include question content and question number, and each question content and the question number corresponding to the question content are vectorized together to form a vector data.

[0057] In a specific implementation, the relational database 13 may store only the answers to all questions supported by the knowledge question answering system and the association between questions and answers, or it may store both the question content and question numbers of all questions supported by the knowledge question answering system. The association between questions and answers may be the association between question content and answers, the association between question numbers and answers, or the association between question content, question numbers, and answers.

[0058] For example, if the knowledge question answering system supports 1000 questions, the vector database 12 may store the question contents and question numbers of the 1000 questions, but not the answers to the 1000 questions. The answers to the 1000 questions may be stored in the relational database 13.

[0059] In a specific implementation, when constructing the vector database 12, the question content and number of each question can be read from the relational database 13 that stores the question content and question numbers of all questions supported by the knowledge question-answering system, and the question content and number of the same question can be vectorized as a string to form a vector data. In this way, the vector data corresponding to each question can be obtained and stored in the vector database.

[0060] In a specific implementation, the user can input questions through the question and answer page 11 , and the knowledge question and answer processing device 14 can search for answers based on the questions input by the user and display the search results on the question and answer page 11 .

[0061] In some embodiments, the knowledge question and answer system 10 may further include a maintenance page 15 , which is used by a knowledge maintainer to maintain questions, answers, and relationships between the questions and answers supported by the knowledge question and answer system.

[0062] Specifically, the knowledge maintainer can add or remove questions, or adjust the relationship between questions and answers through the maintenance page 15. The knowledge question and answer processing device 14 can adjust the contents of the vector database 12 and the relational database 13 based on the input of the maintenance page 15.

[0063] For example, when new questions and answers are received through the maintenance page 15, the knowledge question and answer processing device 14 can number the received question content, and vectorize the received question content and the corresponding question number and store them in the vector database, and store the received answer in the relational database, and add the association between the new question number and the answer in the relational database.

[0064] For example, when deleting questions and answers through the maintenance page 15, the knowledge question and answer processing device 14 can access the vector database and the relational database, and delete the corresponding questions, answers, and the association between the question numbers and answers from the vector database and the relational database.

[0065] In order to enable those skilled in the art to better understand and implement the present invention, the methods, structures, computer-readable storage media and electronic devices corresponding to the above-mentioned knowledge question and answer processing device are described in detail below.

[0066] Figure 2 Flowchart of a knowledge question answering method according to an embodiment of the present invention. Figure 2 , the method may include the following steps:

[0067] Step 21: When an input question is received through the question-and-answer page, the input question is vectorized to obtain an input question vector corresponding to the input question.

[0068] In a specific implementation, the input question can be a statement or a question. For example, the input question can be "Which foods are good for health?" or "I want to know what foods are good for health."

[0069] In a specific implementation, the knowledge question and answer processing device can vectorize the input question to obtain an input question vector corresponding to the input question.

[0070] Step 22: Using the input question vector, search the vector database to obtain N vector data with the highest similarity.

[0071] In a specific implementation, the vector database only stores vector data corresponding to all questions supported by the knowledge question answering system, that is, a number of vector data formed by the question content and question numbers of all questions supported by the knowledge question answering system.

[0072] After obtaining the input question vector, the input question vector can be matched with the question content in the vector data in the vector database to obtain the N vector data with the highest similarity, where N is a positive integer. Afterwards, the N vector data can be used to search the relational database for the answer corresponding to the input question and display it.

[0073] In a specific implementation, the value of N can be set according to actual needs, for example, N=2, 3 or 5.

[0074] Step 23: Using the N vector data, search the relational database to obtain the answer corresponding to the input question and display it.

[0075] In a specific implementation, after obtaining the N vector data with the highest similarity, the N vector data can be directly matched with the association between questions and answers in the relational database to obtain the answers corresponding to the N vector data.

[0076] In one embodiment of the present invention, a similarity threshold can also be set. Using this similarity threshold, N vector data can be further screened to eliminate vector data with lower similarity, and the remaining vector data can be used to match answers in the relational database, thereby further improving the accuracy of the match.

[0077] In a specific implementation, the question number in the vector data can be matched with the association between questions and answers in the relational database to obtain the answer corresponding to the same question number as the answer corresponding to the vector data. Among them, the answer corresponding to the vector data with the highest semantic similarity to the input question is the answer corresponding to the input question.

[0078] Figure 3This is a flow chart of a knowledge question answering method in another embodiment of the present invention. Figure 3 , the method may include the following steps:

[0079] Step 31: Receive input questions.

[0080] Step 32: vectorize the character string containing the input question to obtain an input question vector.

[0081] Regarding step 31 and step 32, please refer to the above description of step 21 for details, which will not be repeated here.

[0082] Step 33: Search the vector database to obtain N vector data with the highest similarity.

[0083] In a specific implementation, each vector data may include question content and question number supported by the knowledge question answering system. Each question content and the question number corresponding to the question content are vectorized together to form a vector data.

[0084] Step 34 : Compare the similarity scores of the N vector data with the similarity threshold to obtain the remaining C vector data.

[0085] Assume that the similarity scores of N vector data are compared with a similarity threshold, and after removing the vector data with a similarity score less than the similarity threshold, C vector data remain, where N≥C≥0.

[0086] Step 35: determine whether C is greater than 0.

[0087] When C=0, execute step 36. When C>0, execute step 37.

[0088] Step 36: Displaying that there are no relevant questions on the question and answer page of the knowledge question and answer system.

[0089] Specifically, after removing vector data with a similarity score less than a similarity threshold, if there is no remaining vector data, it can be directly displayed on the question and answer page that there are no relevant questions, that is, there are no questions in the knowledge question and answer system that match the input question.

[0090] Step 37: Search the relational database to obtain the answer corresponding to the vector data with the highest similarity score, and display the question corresponding to the vector data with the highest similarity score on the question-and-answer page of the knowledge question-and-answer system.

[0091] Specifically, after removing vector data with a similarity score below a similarity threshold, the corresponding answer can be retrieved from the relational database based on the question number in the vector data with the highest similarity score. The question and answer content are then displayed on the Q&A page to inform the user that this is the closest question and answer in the system. At the same time, C-1 other questions can also be displayed on the Q&A page to inform the user that the system has C-1 similar questions. The user can select these questions to continue asking questions, which helps the user quickly and accurately find the questions and answers in the system, improving the efficiency of knowledge Q&A.

[0092] The knowledge question-answering method of the present invention vectorizes and stores only the question content and number in a vector database, eliminating interference from the answer content in matching the question content, thereby improving the accuracy of the knowledge question-answering process. Furthermore, in addition to providing the user with the most similar question and answer in the system, the user is also provided with a list of other highly similar questions in the system, enabling users to quickly and accurately find questions and answers in the system.

[0093] Reference Figure 4 The embodiment of the present invention further provides a knowledge question answering processing device 14, which may include: a vectorization unit 141, a first search unit 142, and a second search unit 143.

[0094] The vectorization unit 141 is adapted to, when receiving an input question through the question-answering page, vectorize the input question to obtain an input question vector corresponding to the input question;

[0095] The first search unit 142 is adapted to use the input question vector to search for N vector data with the highest similarity from the vector database;

[0096] The second search unit 143 is adapted to use the N vector data to search the relational database for an answer corresponding to the input question and display the answer.

[0097] In some embodiments, the knowledge question answering system further includes a maintenance page. Accordingly, the knowledge question answering processing device 14 may further include a maintenance unit 144. The maintenance unit 144 is adapted to update the vector database and the relational database based on the data received from the maintenance page.

[0098] For example, when questions and answers are received through the maintenance page, the maintenance unit 144 can number the received questions, vectorize the received questions and the corresponding question numbers and store them in the vector database, store the received answers in the relational database, and update the association between the question numbers and the answers in the relational database.

[0099] Regarding the vectorization unit 141, the first search unit 142 and the second search unit 143, they can be implemented with reference to the description of the corresponding steps in the above-mentioned knowledge question and answer processing method, and will not be repeated here.

[0100] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of any of the above methods.

[0101] In a specific implementation, the computer-readable storage medium may include: ROM, RAM, magnetic disk or optical disk, etc.

[0102] An embodiment of the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor executes the steps of any of the above methods when running the computer program.

[0103] Regarding the various modules / units included in the various devices and products described in the above embodiments, they can be software modules / units, hardware modules / units, or partly software modules / units and partly hardware modules / units. For example, for various devices and products applied to or integrated into a chip, the various modules / units included therein can all be implemented in the form of hardware such as circuits, or at least part of the modules / units can be implemented in the form of software programs, which run on the processor integrated inside the chip, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits; for various devices and products applied to or integrated into a chip module, the various modules / units included therein can all be implemented in the form of hardware such as circuits, and different modules / units can be located in the same component (such as a chip, circuit module, etc.) or different components of the chip module, or at least part of the modules / units can be implemented in the form of hardware such as circuits. The element can be implemented in the form of a software program, which runs on the processor integrated inside the chip module, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits; for various devices and products applied to or integrated in the terminal, the various modules / units contained therein can all be implemented in the form of hardware such as circuits, and different modules / units can be located in the same component (for example, chip, circuit module, etc.) or different components in the terminal, or, at least some modules / units can be implemented in the form of a software program, which runs on the processor integrated inside the terminal, and the remaining (if any) modules / units can be implemented in the form of hardware such as circuits.

[0104] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be based on the scope defined by the claims.

Claims

1. A knowledge question answering processing method for a knowledge question answering system, characterized in that: The knowledge question and answer system includes a question and answer page, a vector database, and a relational database. The vector database only stores vector data corresponding to all questions supported by the knowledge question and answer system. The relational database stores answers to all questions supported by the knowledge question and answer system and the relationships between questions and answers, but does not store vector data corresponding to all questions supported by the knowledge question and answer system. The method comprises: When an input question is received through the question-and-answer page, the input question is vectorized to obtain an input question vector corresponding to the input question; Using the input question vector, search the vector database to obtain N vector data with the highest similarity, where N is a positive integer; The N vector data are used to search the relational database for an answer corresponding to the input question and display the answer.

2. The knowledge question answering method according to claim 1, wherein: The method of using the N vector data to search a preset relational database to obtain an answer corresponding to the input question and displaying the answer includes: Removing vector data that does not reach a similarity threshold from the N vector data; The remaining vector data in the N vector data are used to search the relational database for an answer corresponding to the input question and display the answer.

3. The knowledge question answering method according to claim 2, wherein: The method of using the remaining vector data in the N vector data to search the relational database for an answer corresponding to the input question and displaying the answer includes: When the number of remaining vector data in the N vector data is zero, a question-answering page of the knowledge question-answering system displays no relevant questions; When the number of remaining vector data in the N vector data is greater than zero, the answer corresponding to the vector data with the highest similarity score is searched from the relational database, and the question and answer corresponding to the vector data with the highest similarity score are displayed on the question and answer page of the knowledge question and answer system.

4. The knowledge question answering method according to claim 3, wherein: The method further comprises searching the relational database for an answer corresponding to the input question using the remaining vector data in the N vector data and displaying the answer. When the number of remaining vector data in the N vector data is greater than zero, in addition to displaying the vector data with the highest similarity score in the remaining vector data on the question and answer page of the knowledge question and answer system, questions corresponding to other vector data are also displayed.

5. The knowledge question answering method according to claim 1, wherein: The vector data includes: question content and question numbers supported by the knowledge question answering system.

6. The knowledge question answering method according to claim 5, wherein: The method of using the N vector data to search a preset relational database to obtain an answer corresponding to the input question includes: Using the question numbers in the N vector data, the answer corresponding to the input question is searched from the relational database.

7. The knowledge question answering method according to claim 1, wherein: The knowledge question answering system further includes a maintenance page; and the method further includes: When questions and answers are received through the maintenance page, the received questions are numbered, and the received questions and corresponding question numbers are vectorized and stored in the vector database, and the received answers are stored in the relational database, and the association between the question numbers and answers is updated in the relational database.

8. A knowledge question answering processing device, characterized in that: Applied to a knowledge question-answering system; the knowledge question-answering system includes a question-answering page, a vector database, and a relational database, wherein the vector database stores only vector data corresponding to all questions supported by the knowledge question-answering system; the relational database stores answers to all questions supported by the knowledge question-answering system and the relationships between questions and answers, but does not store vector data corresponding to all questions supported by the knowledge question-answering system; The device comprises: a vectorization unit adapted to, when receiving an input question through the question-and-answer page, vectorize the input question to obtain an input question vector corresponding to the input question; A first search unit is adapted to use the input question vector to search for N vector data with the highest similarity from the vector database; The second search unit is adapted to use the N vector data to search the relational database for an answer corresponding to the input question and display the answer.

9. The knowledge question answering device according to claim 8, wherein: The knowledge question answering system further includes a maintenance page; the knowledge question answering processing device further includes: The maintenance unit is adapted to, when questions and answers are received through the maintenance page, number the received questions, vectorize the received questions and corresponding question numbers together and store them in the vector database, store the received answers in the relational database, and update the association between the question numbers and the answers in the relational database.

10. A knowledge question answering system, characterized in that: Including the knowledge question and answer processing device described in claim 8 or 9.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the steps of the method according to any one of claims 1 to 7.

12. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein: When the processor runs the computer program, the steps of the method according to any one of claims 1 to 7 are performed.

Images