Knowledge base updating method, intelligent customer service response system and readable storage medium

Abstract

The invention provides a knowledge base updating method, a knowledge base updating device, an intelligent customer service response system, a readable storage medium and a computer program product. The knowledge base updating method comprises the steps of determining a basic question and answer knowledge base; generating a question-answer pair data set based on a question set and an answer set included in the basic question-answer knowledge base; performing question and answer matching degree detection on question and answer pairs in the question and answer pair data set to obtain a detection result; matching answers in the answer set on the basis of keywords of questions for the question-answer pairs with unqualified detection results to obtain answers meeting matching degree detection requirements, generating corrected question-answer pairs on the basis of the answers meeting the matching degree detection requirements, and obtaining a corrected question-answer pair data set; and updating the basic question and answer knowledge base based on the corrected question and answer pair data set to obtain an updated question and answer knowledge base.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and more particularly to a knowledge base updating method, a knowledge base updating device, an intelligent customer service response system, a readable storage medium, and a computer program product. Background Art

[0002] In the existing inbound call system, the intelligent call system (i.e., intelligent customer service answering system) plays a key role, matching and processing incoming calls according to pre-configured rules. However, there are some problems with the existing configuration, such as: 1. It requires high-threshold configuration and high operational complexity: The configuration process of the intelligent call center system requires a certain level of technical skills and professional knowledge. The configuration operation is cumbersome and complex, and requires a lot of time and human resources to complete.

[0003] 2. The knowledge update of the intelligent call center system relies on manual intervention: As the business develops and changes, the knowledge configuration of the intelligent call center system needs to be continuously updated and adjusted. The existing update process relies on manual configuration modification, which is inefficient and slow.

[0004] 3. Limited knowledge coverage: The current knowledge configuration has a narrow coverage, resulting in a low problem-solving rate of the intelligent call center system. In many cases, the problem needs to be transferred to manual processing, which affects the efficiency of problem handling.

[0005] These problems limit the performance and scalability of intelligent call systems. Summary of the invention

[0006] In order to solve one of the above technical problems, the present disclosure provides a knowledge base updating method, a knowledge base updating device, an intelligent customer service response system, a readable storage medium and a computer program product.

[0007] According to one aspect of the present disclosure, a knowledge base updating method is provided, comprising: Determine the basic question-answering knowledge base; Generate a question-answer pair dataset based on the question set and answer set included in the basic question-answer knowledge base; Perform a matching test on the question and answer of the question and answer pairs in the question and answer pair data set to obtain the test results; For question-answer pairs with unqualified detection results, matching the answers in the answer set based on the keywords of the question to obtain answers that meet the matching test, and generating a revised question-answer pair based on the answers that meet the matching test to obtain a revised question-answer pair data set; The basic question-and-answer knowledge base is updated based on the revised question-and-answer pair data set to obtain an updated question-and-answer knowledge base.

[0008] According to at least one embodiment of the present disclosure, the knowledge base updating method further includes, between determining the basic question-answer knowledge base and generating the question-answer pair dataset based on the question set and the answer set included in the basic question-answer knowledge base: The question set and answer set of the basic question-answering knowledge base are preprocessed so that the questions in the question set and the answers in the answer set are texts that meet preset text character length specifications.

[0009] According to the knowledge base updating method of at least one embodiment of the present disclosure, the step of generating a question-answer pair dataset based on a question set and an answer set included in a basic question-answer knowledge base includes: Obtain keywords of questions in the question set and keywords of answers in the answer set; Match the keywords of the question and the keywords of the answer to obtain successfully matched questions and answers; Generate question-answer pairs based on successfully matched questions and answers.

[0010] According to the knowledge base updating method of at least one embodiment of the present disclosure, the step of generating a question-answer pair dataset based on a question set and an answer set included in a basic question-answer knowledge base includes: Receive user questions and extract keywords from user questions; Matching the keywords of the user's question with the keywords of the answers in the answer set to obtain successfully matched questions and answers; Generate question-answer pairs based on successfully matched questions and answers.

[0011] According to the knowledge base updating method of at least one embodiment of the present disclosure, the successfully matched questions and answers are questions and answers with complete keyword matches, or questions and answers with partial keyword matches; If an answer that fully matches the keywords of the question is not obtained, an answer that partially matches the keywords of the question is determined from the answer set.

[0012] According to the knowledge base updating method of at least one embodiment of the present disclosure, the step of performing a matching test on the question and answer of the question and answer pairs in the question and answer pair data set to obtain the test result includes: Extract and vectorize each keyword of the question and each keyword of the answer in the question-answer pair; Obtain the sum of similarities between each vectorized keyword of the question and each vectorized keyword of the answer; Questions and answers whose sum of similarities is greater than or equal to a preset vector similarity threshold are determined as qualified question-answer pairs, and questions and answers whose sum of similarities is less than or equal to or less than the preset vector similarity threshold are determined as unqualified question-answer pairs.

[0013] According to the knowledge base updating method of at least one embodiment of the present disclosure, the step of performing a matching test on the question and answer of the question and answer pairs in the question and answer pair data set to obtain the test result includes: Generate semantics for the question and answer of the question-answer pair through a large semantic model to obtain semantic information of the question and the answer; Whether the question and answer are a qualified question-answer pair is determined based on the degree of association between the semantic information of the question and the semantic information of the answer.

[0014] According to another aspect of the present disclosure, an intelligent customer service answering system is provided, including a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the following process: Receive user questions; Find answers in the question-and-answer knowledge base based on received user questions; Convert the found answer into answer voice and output it; The question-and-answer knowledge base is an updated question-and-answer knowledge base obtained by the knowledge base updating method according to any one of the embodiments of the present disclosure.

[0015] According to another aspect of the present disclosure, a readable storage medium is provided, wherein a computer program is stored in the readable storage medium, and when the computer program is executed by a processor, the computer program is used to implement the knowledge base updating method of any embodiment of the present disclosure.

[0016] According to another aspect of the present disclosure, a computer program product is provided. The computer program product includes a computer program. When the computer program is executed by a processor, it is used to implement the knowledge base updating method of any embodiment of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings illustrate exemplary embodiments of the present disclosure and together with the description serve to explain the principles of the present disclosure. These drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification.

[0018] Figure 1 It is a flowchart of a knowledge base updating method according to an embodiment of the present disclosure.

[0019] Figure 2 It is a flowchart of a knowledge base updating method according to another embodiment of the present invention.

[0020] Figure 3 It is a flowchart of generating a question-answer pair dataset based on a question set and an answer set included in a basic question-answer knowledge base according to an embodiment of the present disclosure.

[0021] Figure 4A schematic diagram of a process for generating a question-answer pair dataset based on a question set and an answer set included in a basic question-answer knowledge base according to another embodiment of the present disclosure is shown.

[0022] Figure 5 It is a flowchart of performing secondary matching detection of questions and answers on question-answer pairs in a question-answer pair data set according to an embodiment of the present disclosure.

[0023] Figure 6 The present invention discloses a knowledge base updating device based on a hardware architecture implemented by a processor and a memory in one embodiment of the present invention.

[0024] Figure 7 A schematic block diagram of the structure of an intelligent customer service response system 2000 according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0025] The present disclosure is further described in detail below in conjunction with the accompanying drawings and examples. It is understood that the specific examples described herein are only used to explain the relevant content, rather than to limit the present disclosure. It should also be noted that, for ease of description, only the parts related to the present disclosure are shown in the accompanying drawings.

[0026] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present disclosure can be combined with each other. The technical solution of the present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0027] Figure 1 It is a schematic diagram of the overall process of a knowledge base updating method according to an embodiment of the present invention.

[0028] refer to Figure 1 The knowledge base updating method M100 of the present disclosure may include steps S102 to S110.

[0029] The knowledge base updating method disclosed in the present invention first executes step S102 to determine a basic question-and-answer knowledge base. The basic question-and-answer knowledge base determined in this step is a question-and-answer knowledge base that needs to be updated, and the basic question-and-answer knowledge base can be a question-and-answer knowledge base in a specific business field, such as a question-and-answer knowledge base in the field of commodity trading, real estate trading, or real estate leasing.

[0030] Exemplarily, the basic question-and-answer knowledge base determined in this step may be an existing question-and-answer knowledge base, such as a question-and-answer knowledge base in a specific business field used by human agents, wherein the questions may be consulting questions regarding real estate sales or real estate leasing, and the answers may be answers to consulting questions regarding real estate sales or real estate leasing.

[0031] In the basic question-and-answer knowledge base, questions can be configured in a question set, and answers can be configured in an answer set.

[0032] The intelligent customer service response system can be connected to the question and answer knowledge base of one or more business fields, such as the question and answer knowledge base in the field of real estate leasing, the question and answer knowledge base in the field of real estate sales, etc.

[0033] In the process of using the intelligent customer service response system, the user's voice question is generally converted into text first, and then a question matching the text is found in the question set of the question and answer knowledge base. Then, based on the question, a matching answer is found from the answer set of the question and answer knowledge base, and finally the matching answer is converted into voice and output.

[0034] Step S102 of the present disclosure determines the basic question and answer knowledge base, that is, determines the scope of business knowledge that the intelligent customer service response system needs to mount.

[0035] After determining the question-and-answer knowledge base that needs to be updated, namely the basic question-and-answer knowledge base, the knowledge base updating method of the present disclosure executes step S104 to generate a question-and-answer pair data set based on the question set and answer set included in the basic question-and-answer knowledge base.

[0036] In this step, the questions in the question set and the answers in the answer set are matched based on keyword similarity (first matching), and the matched questions and answers are taken as question-answer pairs to obtain a question-answer pair dataset.

[0037] The basic question-and-answer knowledge base may include multiple questions and multiple answers, wherein the multiple questions form a question set and the multiple answers form an answer set, wherein the questions and answers may be configured in the basic question-and-answer knowledge base in the form of text vectors.

[0038] In step S104 of the present disclosure, the questions in the question set and the answers in the answer set are matched to generate a question-answer pair data set.

[0039] The question and answer can be matched based on keywords to generate a question-answer pair, thereby generating a question-answer pair data set. The matching can be based on the vector similarity between the question keyword vector and the answer keyword vector. For example, if the vector similarity between a certain keyword vector in the question and a certain keyword vector in the answer is greater than or equal to a first vector similarity threshold, the question and answer are determined to be a question-answer pair, i.e., a question-answer pair.

[0040] After generating the question-answer pair data set of the basic question-answer knowledge base, the question-answer knowledge base updating method of the present disclosure executes step S106, performs a matching test on the question and answer of the question-answer pairs in the question-answer pair data set to obtain a test result.

[0041] Step S104 obtains preliminary matching (first matching) question-answer pairs, i.e., question-answer pairs. The matching degree, i.e., the matching degree, of these question-answer pairs may not necessarily meet the voice answer output requirements of the intelligent customer service response system.

[0042] Therefore, step S104 first determines a question-and-answer pair data set from the basic question-and-answer knowledge base based on the first vector similarity threshold, and step S106 further performs a matching detection based on a matching detection algorithm on the questions and answers of the question-and-answer pairs in the question-and-answer pair data set. Question-and-answer pairs that are less than or equal to the second vector similarity threshold (the second vector similarity threshold is greater than the first vector similarity threshold) may be determined as unqualified, and question-and-answer pairs that are greater than, equal to, or greater than the second vector similarity threshold may be determined as qualified. Those skilled in the art may pre-set or adjust the first vector similarity threshold and the second vector similarity threshold, all of which fall within the protection scope of the present disclosure.

[0043] After obtaining the matching test results of the questions and answers of the question and answer pairs in the question and answer pair data set, the question and answer knowledge base updating method disclosed in the present invention executes step S108, matches the answers in the answer set based on the keywords of the questions for the question and answer pairs with unqualified test results (i.e., the second matching), obtains the answer that satisfies the matching test, and generates a revised question and answer pair based on the answer that satisfies the matching test, thereby obtaining the revised question and answer pair data set.

[0044] Based on step S108, the knowledge base updating method of the present disclosure can determine a question-answer pair data set that satisfies the second vector similarity from the above-mentioned basic question-answer knowledge base.

[0045] After the above steps, the knowledge base updating method disclosed in the present invention corrects the questions and answers in the basic question-and-answer knowledge base based on the first vector similarity threshold and the second vector similarity threshold.

[0046] Therefore, the knowledge base updating method disclosed in the present invention updates the basic question and answer knowledge base based on the revised question and answer pair data set in step S110 to obtain an updated question and answer knowledge base.

[0047] The knowledge base updating method disclosed in the present invention can correct the question and answer pairs of the existing basic question and answer knowledge base by executing the above steps, and obtain an updated question and answer knowledge base. The intelligent customer service question and answer system is connected to the updated question and answer knowledge base to improve the accuracy of the output answers.

[0048] Figure 2 It is a flowchart of a knowledge base updating method according to another embodiment of the present invention.

[0049] refer to Figure 2 In some embodiments of the present disclosure, between step S102 and step S104, step S103 is also included, preprocessing the question set and answer set of the basic question and answer knowledge base, so that the questions in the question set and the answers in the answer set are texts that meet the preset text character length specifications.

[0050] Since the questions and answers in some basic question-and-answer knowledge bases are not all in text form, they may include pictures, texts, tables, links, etc.

[0051] The knowledge base updating method disclosed in the present invention can perform text recognition from pictures and tables in step S103 to obtain text, and convert questions in picture form, answers in picture form, questions in table form, answers in table form, etc. into text form, so as to facilitate the intelligent customer service response system to accurately and effectively output voice answers when using the question and answer knowledge base.

[0052] For pictures, tables, etc. whose text content cannot be extracted, they can be deleted from the basic question and answer knowledge base.

[0053] For answers in the form of links, they can be converted into text descriptions. For example, if the answer is "You can log in to www.abcedfg.com to check", and the URL is XXX official website, the answer can be converted into "You can log in to XXX official website to check" to make it more suitable for conversion into voice answers for output.

[0054] It should be noted that the text characters of the text content extracted from the image or table may be too short or too long, and are not suitable for conversion into voice content output.

[0055] In some embodiments of the present disclosure, text content that is shorter than a first preset text character length can be deleted, and text content that is longer than a second preset text character length can be aggregated based on keywords to remove redundant text to obtain text that is shorter than the second preset text character length.

[0056] Figure 3 It is a flowchart of generating a question-answer pair dataset based on a question set and an answer set included in a basic question-answer knowledge base according to an embodiment of the present disclosure.

[0057] refer to Figure 3 In the knowledge base updating method disclosed in the present invention, preferably, S104, generating a question-answer pair dataset based on a question set and an answer set included in the basic question-answer knowledge base, comprises: S1042, obtaining keywords of questions in the question set and keywords of answers in the answer set; S1044, matching the keywords of the question with the keywords of the answer to obtain successfully matched questions and answers; S1046: Generate question-answer pairs based on successfully matched questions and answers.

[0058] For example, the keywords of question A in the question set include "property area" and "property location", and the keywords of question B include "property unit price" and "property total price".

[0059] The keywords of answer A in the answer set include "120 square meters", "XX community in XXX street", and the keywords of answer B include "unit price XX million yuan", "total price XXX million yuan".

[0060] Then the "property area" in question A will be successfully matched with the "120 square meters" in answer A, the "property location" in question A will be successfully matched with the "XXX Street XX Community" in answer A, the "property unit price" in question B will be successfully matched with the "unit price XX million yuan" in answer B, and the "property total price" in question B will be successfully matched with the "total price XXX million yuan" in answer B.

[0061] The present disclosure regards successfully matched question A and answer A as a question-answer pair, regards question B and answer B as a question-answer pair, and generates a question-answer pair dataset.

[0062] Among them, when matching keywords, matching can be performed based on the vectorized representation of the keywords. Under the inspiration of the technical solution of the present disclosure, those skilled in the art can also adopt other keyword matching methods, all of which fall within the protection scope of the present disclosure.

[0063] Figure 4 A schematic diagram of a process for generating a question-answer pair dataset based on a question set and an answer set included in a basic question-answer knowledge base according to another embodiment of the present disclosure is shown.

[0064] refer to Figure 4 In some other embodiments of the present disclosure, in the knowledge base updating method of the present disclosure, preferably, S104, generating a question-answer pair dataset based on a question set and an answer set included in the basic question-answer knowledge base, comprises: S1042', receiving user questions and extracting keywords from the user questions; S1044', matching the keywords of the user's question with the keywords of the answers in the answer set to obtain successfully matched questions and answers; S1046', generate a question-answer pair based on the successfully matched question and answer.

[0065] Based on the question-answer pair generation method of this embodiment, the question-answer pair data set can be effectively expanded.

[0066] Among them, the "user question" described in this embodiment may be the "user question" received by the intelligent customer service response system during operation, which may be in the form of voice or text. If it is in voice form, it is converted into text form and then keywords are extracted.

[0067] For example, the intelligent customer service response system receives a user's question about the sale of a house, namely, the user's question is: What are the transportation facilities around community A?

[0068] The keywords of the user's question include "A community", "surroundings", and "transportation facilities". The answer A in the answer set is "There is a subway one kilometer away from A community". The keywords of answer A include "A community", "within one kilometer", and "subway". The present invention discloses answer A as the answer that successfully matches the above user question and generates a question-answer pair.

[0069] In some other embodiments of the present disclosure, if there is no answer in the answer set that completely matches the keywords of the user's question, an answer that partially matches the keywords of the user's question is searched in the answer set to generate a question-answer pair.

[0070] For example, there is answer B in the answer set, "A subway station will be built one kilometer away from community A in the next three years." The keywords of answer B are "community A," "one kilometer," "in the next three years," and "subway." The keywords of the above user question do not completely match the keywords of answer B, but are partially matched. In the absence of answer A, a question-answer pair is generated based on answer B and the above user question.

[0071] Figure 5 It is a flowchart of detecting the matching degree of questions and answers for question-answer pairs in a question-answer pair dataset according to an embodiment of the present disclosure.

[0072] refer to Figure 5 In some embodiments of the present disclosure, preferably, step S106, performing a matching test on the question and answer of the question and answer pairs in the question and answer pair data set to obtain a test result, includes: S1062, extracting each keyword of the question and each keyword of the answer in the question-answer pair and vectorizing them; S1064, obtaining the sum of similarities between each vectorized keyword of the question and each vectorized keyword of the answer; S1066. Questions and answers whose sum of similarities is greater than or equal to the second vector similarity threshold are determined as qualified question-answer pairs, and questions and answers whose sum of similarities is less than or equal to or less than the second vector similarity threshold are determined as unqualified question-answer pairs.

[0073] For example, the keywords of question one include "property A", "area" and "price", the keywords of answer one include "property A", "XX square meters", "XX million yuan", and the keywords of answer two include "property A", "next three years", and "subway".

[0074] Based on step S104 described above, question one and answer one may be determined as a question-answer pair and question one and answer two may also be determined as a question-answer pair. Based on the matching degree detection of this embodiment, question one and answer one may be determined as a qualified question-answer pair, and question one and answer two may be determined as an unqualified question-answer pair.

[0075] In some other embodiments of the present disclosure, step S106, performing a matching test on the question and answer of the question and answer pairs in the question and answer pair data set to obtain a test result, may also include: Generate semantics for the question and answer of the question-answer pair through a large semantic model to obtain semantic information of the question and the answer; Whether the question and answer are a qualified question-answer pair is determined based on the degree of association between the semantic information of the question and the semantic information of the answer.

[0076] Among them, the semantic big model can be an existing semantic big model based on recurrent neural network, or a semantic big model based on transformer architecture, such as the GPT (Generative Pre-Trained Transformer) series model, and the present disclosure does not make any special limitations on this.

[0077] In some embodiments of the present disclosure, a semantic big model can be used to perform intent recognition on the semantic information of the question and the semantic information of the answer to obtain the intent information of the question and the intent information of the answer respectively, and the degree of correlation between the question and the answer is judged based on the intent information of the question and the intent information of the answer, and question-answer pairs with a degree of correlation greater than or greater than equal to a correlation threshold are determined as qualified question-answer pairs, and question-answer pairs with a degree of correlation less than, equal to or less than the correlation threshold are determined as unqualified question-answer pairs.

[0078] Based on the knowledge base updating method described above in the present disclosure, the question and answer pairs in the updated question and answer knowledge base are more accurate.

[0079] Based on the knowledge base updating method described above, the present disclosure also provides a knowledge base updating device.

[0080] Figure 6 The present invention discloses a knowledge base updating device based on a hardware architecture implemented by a processor and a memory in one embodiment of the present invention.

[0081] In some embodiments of the present disclosure, reference Figure 6 , the knowledge base updating device 1000 comprises: A basic question and answer knowledge base determination module 1002, the basic question and answer knowledge base determination module 1002 determines a basic question and answer knowledge base; A question-answer pair data set generation module 1004, which generates a question-answer pair data set based on a question set and an answer set included in a basic question-answer knowledge base; A matching degree detection module 1006, which performs matching degree detection of questions and answers on the question-answer pairs in the question-answer pair data set to obtain a detection result; The question-answer pair correction module 1008 matches the answers in the answer set based on the keywords of the question for the question-answer pair whose detection result is unqualified, obtains an answer that satisfies the matching degree detection, and generates a corrected question-answer pair based on the answer that satisfies the matching degree detection; The question and answer knowledge base updating module 1010 updates the basic question and answer knowledge base based on the revised question and answer pair data set to obtain an updated question and answer knowledge base.

[0082] The knowledge base updating device 1000 of the present disclosure may include corresponding modules for executing each or several steps in the above flowchart. Therefore, each step or several steps in the above flowchart may be executed by the corresponding module, and the device may include one or more modules in these modules. The module may be one or more hardware modules specially configured to execute the corresponding steps, or implemented by a processor configured to execute the corresponding steps, or stored in a computer-readable medium for implementation by a processor, or implemented by some combination.

[0083] The hardware structure of the knowledge base updating device of the present disclosure can be realized by bus architecture. The bus architecture may include any number of interconnected buses and bridges, depending on the specific application and overall design constraints of the hardware. The bus 1100 connects various circuits including one or more processors 1200, memory 1300 and / or hardware modules together. The bus 1100 may also connect various other circuits 1400 such as peripherals, voltage regulators, power management circuits, external antennas, etc.

[0084] The bus 1100 may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Component (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the figure only uses one connecting line, but does not mean that there is only one bus or one type of bus.

[0085] Any process or method description in the flowchart or otherwise described herein can be understood as a module, fragment or portion of a code representing one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present disclosure includes other implementations, in which the functions may not be performed in the order shown or discussed, including performing the functions in a substantially simultaneous manner or in a reverse order according to the functions involved, which should be understood by a person skilled in the art of the present disclosure. The processor performs the various methods and processes described above. For example, the method implementation in the present disclosure can be implemented as a software program, which is tangibly contained in a machine-readable medium, such as a memory. In some embodiments, part or all of the software program can be loaded and / or installed via a memory and / or a communication interface. When the software program is loaded into the memory and executed by the processor, one or more steps in the method described above can be performed. Alternatively, in other embodiments, the processor can be configured to perform one of the above methods in any other appropriate manner (e.g., by means of firmware).

[0086] The logic and / or steps represented in the flowchart or otherwise described herein may be embodied in any readable storage medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other system that can fetch instructions from and execute instructions on an instruction execution system, apparatus, or device), or in combination with such instruction execution systems, apparatuses, or devices.

[0087] The present disclosure also provides an intelligent customer service response system, which includes a processor and a memory, wherein the memory stores a computer program. When the computer program is executed by the processor, the processor performs the following process: Receive user questions; Find answers in the question-and-answer knowledge base based on received user questions; The found answer is converted into answering speech and output.

[0088] Among them, the question and answer knowledge base is an updated question and answer knowledge base obtained based on the knowledge base updating method of any implementation method of the present disclosure.

[0089] In some embodiments of the present disclosure, if the user question is in voice form, it is converted into text form to perform the search for an answer in the question-and-answer knowledge base based on the received user question.

[0090] Figure 7 A schematic block diagram of the structure of an intelligent customer service response system 2000 according to an embodiment of the present disclosure is shown.

[0091] refer to Figure 7 , the intelligent customer service answering system 2000 includes: A voice receiving module 2002 receives user questions in voice form; A first voice conversion module 2004, which converts the user's question in voice form into the user's question in text form; Question and answer knowledge base 2006, storing question and answer pairs in text form; A search module 2008, the search module 2008 searches for answers in the question-answer knowledge base 2006 based on the user's question in text form; The second voice conversion module 2010 converts the found answer into a response voice and outputs it.

[0092] Among them, the question and answer knowledge base is an updated question and answer knowledge base obtained based on the knowledge base updating method of any embodiment of the present disclosure.

[0093] In some other embodiments of the present disclosure, the first voice conversion module 2004 and the second voice conversion module 2010 may be one voice conversion module.

[0094] It should be noted that in some embodiments of the present disclosure, the knowledge base updating device of the present disclosure can be configured as a part of the intelligent customer service response system within the intelligent customer service response system, and the two use a set of hardware processing devices (the memory, processor, other circuits, bus, etc. described above). In other embodiments of the present disclosure, the knowledge base updating device of the present disclosure is independent of the intelligent customer service response system and is connected to the intelligent customer service response system in a data access manner.

[0095] The present disclosure also provides a readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, it is used to implement the knowledge base updating method of any embodiment of the present disclosure.

[0096] "Readable storage medium" can be any device that can contain, store, communicate, propagate or transmit a program for use with an instruction execution system, device or apparatus or in conjunction with such instruction execution system, device or apparatus. More specific examples of readable storage media include the following: an electrical connection portion having one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable read-only memory (CDROM), etc.

[0097] The present disclosure further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it is used to implement at least the knowledge base updating method of any embodiment of the present disclosure.

[0098] The computer program may be stored in a readable storage medium or transferred from one readable storage medium to another readable storage medium, for example, the computer program may be transferred from one website, computer, server or data center to another website, computer, server or data center by wired or wireless means. The readable storage medium may be any available medium that can be accessed or a data storage device such as a server, data center, etc. that integrates one or more available media. The available medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; it may also be an optical medium, such as a digital video disk; it may also be a semiconductor medium, such as a solid state drive. The computer readable storage medium may be a volatile or non-volatile storage medium, or may include both volatile and non-volatile types of storage media.

[0099] Those skilled in the art should understand that the above embodiments are only for the purpose of clearly illustrating the present disclosure, and are not intended to limit the scope of the present disclosure. For those skilled in the art, other changes or modifications may be made based on the above disclosure, and these changes or modifications are still within the scope of the present disclosure.

Claims

1. A knowledge base updating method, characterized in that: include: Determine the basic question-answering knowledge base; Generate a question-answer pair dataset based on the question set and answer set included in the basic question-answer knowledge base; Perform a matching test on the question and answer of the question and answer pairs in the question and answer pair data set to obtain the test results; For question-answer pairs with unqualified detection results, matching the answers in the answer set based on the keywords of the question to obtain answers that meet the matching test, and generating a revised question-answer pair based on the answers that meet the matching test to obtain a revised question-answer pair data set; as well as The basic question-and-answer knowledge base is updated based on the revised question-and-answer pair data set to obtain an updated question-and-answer knowledge base.

2. The knowledge base updating method according to claim 1, characterized in that: Between the step of determining the basic question-answer knowledge base and the step of generating a question-answer pair dataset based on the question set and the answer set included in the basic question-answer knowledge base, the step further includes: The question set and answer set of the basic question-answering knowledge base are preprocessed so that the questions in the question set and the answers in the answer set are texts that meet preset text character length specifications.

3. The knowledge base updating method according to claim 1, characterized in that: The step of generating a question-answer pair dataset based on a question set and an answer set included in a basic question-answer knowledge base includes: Obtain keywords of questions in the question set and keywords of answers in the answer set; Matching the keywords of the question with the keywords of the answer to obtain successfully matched questions and answers; and Generate question-answer pairs based on successfully matched questions and answers.

4. The knowledge base updating method according to claim 1 or 3, characterized in that: The step of generating a question-answer pair dataset based on a question set and an answer set included in a basic question-answer knowledge base includes: Receive user questions and extract keywords from user questions; Matching the keywords of the user's question with the keywords of the answers in the answer set to obtain successfully matched questions and answers; and Generate question-answer pairs based on successfully matched questions and answers.

5. The knowledge base updating method according to claim 4, characterized in that: The successfully matched questions and answers are questions and answers with complete keyword matches, or questions and answers with partial keyword matches; If an answer that fully matches the keywords of the question is not obtained, an answer that partially matches the keywords of the question is determined from the answer set.

6. The knowledge base updating method according to claim 1, characterized in that: The matching degree detection of questions and answers is performed on the question-answer pairs in the question-answer pair data set to obtain the detection result, including: Extract and vectorize each keyword of the question and each keyword of the answer in the question-answer pair; Obtaining the sum of similarities between each vectorized keyword of the question and each vectorized keyword of the answer; and Questions and answers whose sum of similarities is greater than or equal to a preset vector similarity threshold are determined as qualified question-answer pairs, and questions and answers whose sum of similarities is less than or equal to or less than the preset vector similarity threshold are determined as unqualified question-answer pairs.

7. The knowledge base updating method according to claim 1 or 6, characterized in that: The matching degree detection of questions and answers is performed on the question-answer pairs in the question-answer pair data set to obtain the detection result, including: Generate semantics for the question and answer of the question-answer pair through a semantic big model to obtain semantic information of the question and semantic information of the answer; and Whether the question and answer are a qualified question-answer pair is determined based on the degree of association between the semantic information of the question and the semantic information of the answer.

8. An intelligent customer service response system, characterized in that: The system comprises a processor and a memory, wherein the memory stores a computer program. When the computer program is executed by the processor, the processor performs the following processes: Receive user questions; Find answers in the question-and-answer knowledge base based on received user questions; as well as Convert the found answer into answer voice and output it; The question-and-answer knowledge base is an updated question-and-answer knowledge base obtained based on the knowledge base updating method according to any one of claims 1 to 7.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the computer program is used to implement the knowledge base updating method according to any one of claims 1 to 7.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the computer program is used to at least implement the knowledge base updating method according to any one of claims 1 to 7.

Images