CN117763099A

CN117763099A - Interaction method and device of intelligent customer service system

Info

Publication number: CN117763099A
Application number: CN202311502061.3A
Authority: CN
Inventors: 刘泓
Original assignee: Seashell Housing Beijing Technology Co Ltd
Current assignee: Seashell Housing Beijing Technology Co Ltd
Priority date: 2023-11-10
Filing date: 2023-11-10
Publication date: 2024-03-26

Abstract

The embodiment of the invention provides an interaction method and device of an intelligent customer service system, wherein the method comprises the following steps: acquiring first content of user consultation; generating a first embedded vector of the user consultation first content, acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in the vectorization database, acquiring a text block corresponding to the second embedded vector, inputting the user consultation first content and the text block as first context data into an intelligent customer service system, and generating an intelligent answer of the user consultation first content. The embodiment of the invention improves the reasoning capacity of the intelligent customer service system for the consultation content of the user and improves the accuracy of intelligent answers, thereby improving the quality of the intelligent answers.

Description

Interaction method and device of intelligent customer service system

Technical Field

The embodiment of the invention relates to the technical field of computers, in particular to an interaction method and device of an intelligent customer service system.

Background

The working mode of the existing intelligent customer service system is based on the following technology:

(1) Text analysis and semantic understanding: the system uses Natural Language Processing (NLP) techniques to analyze and understand text entered by the user. Including word segmentation, part-of-speech tagging, syntactic analysis, and semantic role tagging, to infer user intent and question type.

(2) Knowledge base and rules engine: the system typically builds a knowledge base containing predefined questions and answers, product information, common questions solutions, etc. The system will retrieve relevant information from the knowledge base based on the user's questions and generate answers based on predefined rules and logic.

(3) Template matching and rule matching: the system uses template matching and rule matching methods to match the user's questions to predefined templates or rules. These templates and rules may capture common patterns of questions and generate corresponding answers.

The answers and solutions of the existing intelligent customer service system are based on a pre-established knowledge base, rules or templates, answers are obtained by processing texts input by users, the intention and the demands of the users cannot be understood when facing complex questions, the reasoning capability is limited, the intelligent answers or the accuracy of the answers cannot be lowered, and the quality of the intelligent answers needs to be improved.

Disclosure of Invention

Aiming at the defects existing in the prior art, the embodiment of the invention provides an interaction method and device of an intelligent customer service system.

The embodiment of the invention provides an interaction method of an intelligent customer service system, which comprises the following steps: acquiring first content of user consultation; generating a first embedded vector of the user consultation first content, acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in a vectorization database, acquiring a text block corresponding to the second embedded vector, inputting the user consultation first content and the text block as first context data into an intelligent customer service system, and generating an intelligent answer of the user consultation first content.

According to the interaction method of the intelligent customer service system provided by the embodiment of the invention, the method further comprises the following steps: and responding to the first content of the user consultation to contain automatic flow processing information, and triggering a corresponding robot automatic processing flow according to the automatic flow processing information.

According to the interaction method of the intelligent customer service system provided by the embodiment of the invention, the method further comprises the following steps: collecting first data related to questions and answers, carrying out data processing on the first data and generating a formatted document; and carrying out blocking processing on the formatted document to obtain a text block, generating the second embedded vector of the text block, and storing the second embedded vector and the corresponding text block in the vectorization database in an associated manner.

According to the interaction method of the intelligent customer service system provided by the embodiment of the invention, the first data related to the acquisition question and answer comprises the following steps: and simulating manual operation, automatically logging in the enterprise intranet to capture second data related to questions and answers, and positioning and extracting the first data according to the second data.

According to the interaction method of the intelligent customer service system provided by the embodiment of the invention, the second data related to the automatic login enterprise intranet capturing question and answer comprises the following steps: and automatically logging in an enterprise intranet, and capturing relevant contents of a regulation system, a question-answer library and a knowledge base to obtain the second data.

According to the interaction method of the intelligent customer service system provided by the embodiment of the invention, the method further comprises the following steps: and responding to the second data update, updating the corresponding first data according to the updated second data, and updating the text block and the second embedded vector according to the updated first data.

According to the interaction method of the intelligent customer service system provided by the embodiment of the invention, the method further comprises the following steps: obtaining user consultation second content in batches, generating a third embedded vector of the user consultation second content, obtaining at least one second embedded vector meeting a preset similarity requirement with the third embedded vector in the vectorization database, and obtaining the text block corresponding to the second embedded vector; and inputting the second content and the text block consulted by the user as second context data into a language model for automatically generating intelligent answers, and training the language model to obtain the intelligent customer service system.

According to the interaction method of the intelligent customer service system provided by the embodiment of the invention, the method further comprises the following steps: acquiring a correct answer corresponding to the intelligent answer output by the language model when the intelligent answer is a wrong answer; and inputting the second context data into the language model, and continuing training the language model by taking the correct answer as an output label.

The embodiment of the invention also provides electronic equipment, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the steps of the interaction method of any one of the intelligent customer service systems are realized when the processor executes the program.

The embodiment of the invention also provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the interaction method of the intelligent customer service system as described in any one of the above.

The embodiment of the invention also provides a computer program product, which comprises a computer program, wherein the computer program realizes the steps of the interaction method of the intelligent customer service system when being executed by a processor.

According to the interaction method and device for the intelligent customer service system, the first content is consulted by the user, the first embedded vector of the first content is generated, at least one second embedded vector meeting the preset similarity requirement with the first embedded vector in the vectorization database is obtained, the text block corresponding to the second embedded vector is obtained, the first content and the text block are consulted by the user and serve as first context data to be input into the intelligent customer service system, the intelligent answer of the first content is generated, the reasoning capacity of the intelligent customer service system for the consultation content of the user is improved, and the accuracy of the intelligent answer is improved, so that the quality of the intelligent answer is improved.

Drawings

In order to more clearly illustrate the technical solutions of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of an interaction method of an intelligent customer service system according to an embodiment of the present invention;

FIG. 2 is a second flow chart of an interaction method of the intelligent customer service system according to the embodiment of the invention;

FIG. 3 is a schematic structural diagram of a device for constructing an intelligent customer service system according to an embodiment of the present invention;

fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Fig. 1 is a schematic flow chart of an interaction method of an intelligent customer service system according to an embodiment of the present invention. As shown in fig. 1, the method includes:

step S1, acquiring first content consulted by a user.

The first content of the user consultation is the user consultation content in the using process of the intelligent customer service system.

Step S2, generating a first embedded vector of the user consultation first content, acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in a vectorization database, acquiring a text block corresponding to the second embedded vector, inputting the user consultation first content and the text block as first context data into an intelligent customer service system, and generating an intelligent answer of the user consultation first content.

Generating a first embedded vector of the user consultation first content through vectorization processing, and acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in a vectorization database; the text block and the second embedded vector are stored in the vectorization database in an associated mode; and acquiring text blocks corresponding to at least one second embedded vector of which the first embedded vector meets the preset similarity requirement in the vectorization database, and inputting the first content and the text blocks consulted by the user as first context data into the intelligent customer service system to generate intelligent answers of the first content consulted by the user. The predetermined similarity requirement may be that the similarity value is greater than a predetermined similarity threshold, or that the similarity value is a predetermined number of values preceding the high-to-low ranking of similarity values.

According to the interaction method of the intelligent customer service system, the first content is consulted by the user, the first embedded vector of the first content is generated, at least one second embedded vector meeting the preset similarity requirement with the first embedded vector in the vectorization database is obtained, the text block corresponding to the second embedded vector is obtained, the first content and the text block are consulted by the user and serve as the first context data to be input into the intelligent customer service system, the intelligent answer of the first content is generated, the reasoning capacity of the intelligent customer service system for the first content is improved, the accuracy of the intelligent answer is improved, and therefore the quality of the intelligent answer is improved.

The existing intelligent customer service system cannot automatically trigger related processes, manual intervention or additional manual operation is needed, and processing efficiency and user experience are reduced.

In order to improve processing efficiency and enhance user experience, the intelligent customer service system provided by the embodiment of the invention responds to the fact that the first content contains the automatic flow processing information, and triggers the corresponding robot automatic processing flow according to the automatic flow processing information, for example, triggers the RPA robot to perform automatic flow processing, combines the intelligent customer service system with the RPA robot, and can automatically wake up the related RPA robot to perform operations such as form filling, instruction execution and the like when the intelligent customer service system generates a solution.

Implementation effect example:

the user: "I want to go to Guangzhou business trip, please help I submit a business trip application. "

And (3) model: "when please ask to go? "

The user: "No. 22 to No. 25 for 6 months". "

And (3) model: "good, submitted, please log in to travel system to view. "

The user: "what is the travel criteria in Guangzhou? "

And (3) model: "Guangzhou is a first-line city, accommodation cost is 600 yuan/person/day, and meal cost is 150 yuan/person/day. "

According to the interaction method of the intelligent customer service system, provided by the embodiment of the invention, the corresponding robot automatic processing flow is triggered according to the automatic flow processing information by responding to the fact that the user consultation content contains the automatic flow processing information, so that the automatic flow processing capability of the intelligent customer service system is improved.

The first data is related data which is assisted in question and answer, and specific content can be set according to requirements. The first data is data processed and a formatted document is generated. Wherein the data processing may comprise: cleaning the collected first data, removing unnecessary labels, blank characters and the like, and ensuring the consistency and accuracy of the data; text analysis and processing, such as word segmentation, stop word removal, part-of-speech tagging, etc., is performed to better understand and utilize the data. In formatted document generation, data is structured and categorized, and based on the processed data, a formatted document is generated for use in subsequent data processing.

The blocking process of the formatted document may use a LangChain tool to divide the formatted document into appropriate blocks (e.g., paragraphs, chapters, etc.) for subsequent processing.

And further carrying out vectorization processing on the obtained text block to obtain a corresponding second embedded vector. For example, text2vec-large-Chinese may be used to perform vectorization, each text block is converted into a corresponding second embedded vector, and word-level or subword-level embedded vectors may be selected as desired. The vectorized representation (second embedded vector) is associated with the original text block and stored to the vectorized database for retrieval and matching in the vectorized database.

According to the interaction method of the intelligent customer service system, the first data related to questions and answers are collected, the first data are subjected to data processing and formatted documents are generated, the formatted documents are subjected to block processing to obtain text blocks, second embedded vectors of the text blocks are generated, the second embedded vectors and the corresponding text blocks are stored in a vectorization database in an associated mode, generation of the vectorization database is achieved, and context information is provided when intelligent answers are generated for the intelligent customer service system.

When first data related to questions and answers are collected, manual operation can be simulated by using the RPA technology, the enterprise intranet is automatically logged in, and second data related to questions and answers are browsed and grabbed. And the RPA robot can accurately position and extract the required first data by analyzing the webpage structure, keywords and the like.

According to the interaction method of the intelligent customer service system, provided by the embodiment of the invention, the second data related to the questions and answers are automatically captured by logging in the enterprise intranet through simulating manual operation, and the first data is positioned and extracted according to the second data, so that the automation level of the construction of the intelligent customer service system is improved, and the processing efficiency is improved.

And when the enterprise intranet is automatically logged in to capture the second data related to the questions and answers, the second data can be obtained by browsing and capturing the related contents of the regulation system, the question and answer library (Q & A) and the knowledge base.

According to the interaction method of the intelligent customer service system, the second data is obtained by automatically logging in the enterprise intranet and capturing the relevant contents of the regulation system, the question-answer library and the knowledge base, and the data quality of the second data is improved.

The reliability and timeliness of the data sources also have an impact on the performance of the system. In order to improve the data quality and timeliness, the embodiment of the invention rapidly reflects the update of the second data to the intelligent customer service system, so that the intelligent customer service system answers according to the latest context data, and the quality of intelligent answer is further improved.

The update of the second data may be monitored according to a set time interval. If the updated information of the second data is obtained, the corresponding first data is updated according to the updated second data, for example, the first data is extracted according to the updated second data and the original data is replaced, or the updated content of the second data is obtained and the corresponding content in the first data is replaced.

And carrying out data processing according to the updated first data, generating a formatted document, carrying out block division processing on the formatted document to obtain updated text blocks, generating a second embedded vector of the updated text blocks, and storing the updated second embedded vector and the corresponding text blocks in a vectorization database in an associated manner to replace the original second embedded vector and the text blocks.

According to the interaction method of the intelligent customer service system, the corresponding first data is updated according to the updated second data by responding to the second data update, and the text block and the second embedded vector are updated according to the updated first data, so that the data quality and timeliness are improved, and the quality of intelligent answers is improved.

The user consultation second content for model training is acquired, for example, the user consultation second content can be grasped in batches from a work order system, and the quality and the diversity of problem data are ensured so as to cover the problems of different fields and types. Vectorizing the second content consulted by the user, generating a third embedded vector consulting the second content by the user, performing similarity matching on the third embedded vector and the second embedded vector in the vectorization database, and obtaining at least one second embedded vector meeting the preset similarity requirement with the third embedded vector in the vectorization database.

And because the second embedded vector and the corresponding text block are stored in an associated mode, according to the associated storage relation, acquiring a text block corresponding to at least one second embedded vector of which the third embedded vector meets the preset similarity requirement, inputting the acquired text block and the second content consulted by the user as second context data into a language model for automatically generating intelligent answers, such as an LLM model, and generating intelligent answers of the second content consulted by the user, namely reasoning the language model to synthesize the text block and the second context information reflected by the second content consulted by the user, and generating intelligent answers of the second content consulted by the user. And repeatedly executing the process of inputting different user consultation second contents and corresponding text blocks as second context data into the language model to generate intelligent answers of the second user consultation contents, training the language model, and completing training to obtain the intelligent customer service system.

Wherein the LLM model can use a ChatGLM-6B model.

According to the interaction method of the intelligent customer service system, the user consultation second content is obtained in batches, the third embedded vector of the user consultation second content is generated, at least one second embedded vector meeting the preset similarity requirement with the third embedded vector in the vectorization database is obtained, the text block corresponding to the second embedded vector is obtained, the user consultation second content and the text block are used as second context data to be input into a language model for automatically generating intelligent answers, the language model is trained, the intelligent customer service system is obtained, the reasoning capacity of the intelligent customer service system for the user consultation content is improved, the accuracy of the intelligent answers is improved, and therefore the quality of the intelligent answers is improved.

Traditional intelligent customer service systems lack the ability to learn and optimize continuously. The system is usually operated based on a fixed model and knowledge, and is difficult to flexibly adapt to new scenes and requirements. Thus, the system requires manual intervention and periodic manual updates to maintain its accuracy and effectiveness.

In order to further improve accuracy of intelligent answers and continuously optimize an intelligent customer service system, the embodiment of the invention generates the intelligent answers of the user consultation second contents by taking the user consultation second contents and corresponding text blocks as second context data input language models, further judges the quality of the answers according to predefined standards and indexes, and marks the quality of the answers. If the answer is correct, the next labeling is carried out; if the answer is wrong, a correct answer is given. And for the answers marked as errors, the user consults the second content and the corresponding text blocks as context data to input the language model, and the language model is further trained by taking the correct answers as output labels, so that the language model carries out correction learning according to the correct answers and the context, and the accuracy and the specialty of the model on similar questions are improved.

And through continuous manual labeling and model updating, the loop iteration and continuous improvement of the answer quality of the intelligent customer service system are realized.

According to the interactive method of the intelligent customer service system, the second context data is input into the language model by acquiring the correct answer corresponding to the intelligent answer output by the language model as the wrong answer, and the correct answer is used as the output label, so that the language model is continuously trained, and the quality of the intelligent answer is further improved.

Fig. 2 is a second flowchart of an interaction method of the intelligent customer service system according to an embodiment of the present invention. As shown in FIG. 2, the intelligent customer service system comprises a complete set of processes from construction to use, including data acquisition, data processing, model training, manual labeling, question-answering and RPA, and the specific implementation of each step is described in detail below.

1. Data acquisition

Data acquisition of the RPA robot:

the RPA technology is used for simulating manual operation, automatically logging in the enterprise intranet, and browsing and grabbing relevant contents of the regulation system, Q & A and the knowledge base.

The robot can accurately position and extract the required data by analyzing the webpage structure, keywords and the like.

And (3) data processing:

and cleaning the acquired data, removing unnecessary labels, blank characters and the like, and ensuring the consistency and accuracy of the data.

Text analysis and processing, such as word segmentation, stop word removal, part-of-speech tagging, etc., is performed to better understand and utilize the data.

Generating a formatted document:

the data is structured and classified and, based on the processed data, a formatted document (e.g., structured text or other data format) is generated for use in subsequent data processing.

2. Data processing

V LangChain tandem task:

using LangChain tool, defining flow and logic of whole task, using formatted document generated in last step as input, using LangChain to make text block and dividing document into proper block (such as paragraph and chapter) for subsequent processing. The LangChain tool may involve data preprocessing such as data cleaning and word segmentation in data blocking processing.

Vectorization of v.Embedding:

and carrying out vectorization operation by using text2vec-large-Chinese, converting each text block into a corresponding embedded vector, selecting word level or sub word level embedded vectors according to requirements, and associating the vectorized representation with the original text block so as to search and match in a vectorization database.

The vectorization database stores:

creating a vectorization database for storing vectorized text blocks and their associated information, and storing the vectorized text blocks and their associated information into the database for subsequent retrieval and query operations. Wherein the associated information includes original text blocks, stored time information, and the like.

3. Model training

User consultation acquisition:

user consultation content is obtained from the work order system in batches, and quality and diversity of problem data are ensured so as to cover problems in different fields and types.

Vectorization of v.Embedding:

and vectorizing the user consultation by using text2vec-large-Chinese, and matching the vectorized user consultation with the vector representation of the text block in the vectorization database constructed before in similarity.

And (3) data retrieval:

according to the similarity matching result, a text block corresponding to the similarity vector is acquired for acquiring related data (such as an answer, an explanation, etc.).

Question answering:

and taking the user consultation content and the retrieved text blocks as contexts, inputting the contexts to the ChatGLM-6B for reasoning, and generating answers related to the user consultation content.

4. Manual labeling

Manually annotating the quality of the answer:

and delivering the answers generated by the model in the previous step to a manual labeling team for evaluation, and judging the quality of the answers by the labeling team according to the pre-defined standard and index. If the answer is correct, proceed to the next annotation, if the answer is wrong, then give the correct answer.

Retraining:

for answers marked as wrong, the previous context is re-input to the ChatGLM-6B model, and supervised learning is performed with the correct answer as an output label. Thus, the ChatGLM-6B can perform correction learning according to correct answers and context, and accuracy and specialty of the model on similar questions are improved.

Loop iteration and persistence improvement:

repeating the steps, and gradually improving the answer quality of ChatGLM-6B through continuous manual labeling and model updating.

5. Question and answer

A user dialogue:

the user carries out multiple rounds of conversations with the intelligent customer service system, questions are presented, requirements are expressed or solutions are sought, the intelligent customer service system receives user input and calls the ChatGLM-6B to reply, and the ChatGLM-6B carries out reasoning and generates accurate and professional answers and solutions based on the context information.

Flow initiation:

if some processes (such as application, approval, inquiry and the like) need to be initiated in the conversation process, the intelligent customer service system can automatically wake up the corresponding RPA robot for processing.

The preferred embodiments of the present embodiment may be freely combined on the premise that the logic or structure does not conflict with each other, and the present invention is not limited to this.

The following describes the interaction device of the intelligent customer service system provided by the embodiment of the present invention, and the interaction device of the intelligent customer service system described below and the interaction method of the intelligent customer service system described above can be referred to correspondingly.

Fig. 3 is a schematic structural diagram of an interaction device of an intelligent customer service system according to an embodiment of the present invention. As shown in fig. 3, the apparatus includes an acquisition module 10 and a generation module 20, where: the acquisition module 10 is configured to: acquiring first content of user consultation; the generating module 20 is configured to: generating a first embedded vector of the user consultation first content, acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in a vectorization database, acquiring a text block corresponding to the second embedded vector, inputting the user consultation first content and the text block as first context data into an intelligent customer service system, and generating an intelligent answer of the user consultation first content.

According to the interaction device of the intelligent customer service system, provided by the embodiment of the invention, the first embedded vector of the first content is generated by acquiring the first content consulted by the user, at least one second embedded vector meeting the preset similarity requirement with the first embedded vector in the vectorization database is acquired, the text block corresponding to the second embedded vector is acquired, the first content consulted by the user and the text block are used as the first context data to be input into the intelligent customer service system, the intelligent answer of the first content consulted by the user is generated, the reasoning capacity of the intelligent customer service system on the content consulted by the user is improved, and the accuracy of the intelligent answer is improved, so that the quality of the intelligent answer is improved.

According to the interaction device of the intelligent customer service system provided by the embodiment of the invention, the device further comprises an automatic flow processing module, which is used for: and responding to the first content of the user consultation to contain automatic flow processing information, and triggering a corresponding robot automatic processing flow according to the automatic flow processing information.

According to the interaction device of the intelligent customer service system, provided by the embodiment of the invention, the corresponding robot automatic processing flow is triggered according to the automatic flow processing information by responding to the fact that the user consultation content contains the automatic flow processing information, so that the automatic flow processing capability of the intelligent customer service system is improved.

According to the interaction device of the intelligent customer service system provided by the embodiment of the invention, the device further comprises a preprocessing module for: collecting first data related to questions and answers, carrying out data processing on the first data and generating a formatted document; and carrying out blocking processing on the formatted document to obtain a text block, generating the second embedded vector of the text block, and storing the second embedded vector and the corresponding text block in the vectorization database in an associated manner.

According to the interaction device of the intelligent customer service system, provided by the embodiment of the invention, the first data related to questions and answers are collected, the first data are subjected to data processing and formatted documents are generated, the formatted documents are subjected to block processing to obtain text blocks, the second embedded vectors of the text blocks are generated, the second embedded vectors and the corresponding text blocks are associated and stored in the vectorization database, the vectorization database is generated, and the context information is provided when the intelligent customer service system generates intelligent answers.

According to the interaction device of the intelligent customer service system provided by the embodiment of the invention, the preprocessing module is specifically used for acquiring first data related to questions and answers when being used for acquiring the first data related to questions and answers: and simulating manual operation, automatically logging in the enterprise intranet to capture second data related to questions and answers, and positioning and extracting the first data according to the second data.

According to the interaction device of the intelligent customer service system, provided by the embodiment of the invention, the second data related to the questions and answers are automatically captured by logging in the enterprise intranet through simulating manual operation, and the first data is positioned and extracted according to the second data, so that the automation level of the construction of the intelligent customer service system is improved, and the processing efficiency is improved.

According to the interaction device of the intelligent customer service system provided by the embodiment of the invention, when the preprocessing module is used for automatically logging in the second data related to the question and answer captured by the enterprise intranet, the preprocessing module is specifically used for: and automatically logging in an enterprise intranet, and capturing relevant contents of a regulation system, a question-answer library and a knowledge base to obtain the second data.

According to the interaction device of the intelligent customer service system, the second data is obtained by automatically logging in the enterprise intranet and capturing the relevant contents of the regulation system, the question-answer library and the knowledge base, and the data quality of the second data is improved.

According to the interaction device of the intelligent customer service system provided by the embodiment of the invention, the preprocessing module is further used for: and responding to the second data update, updating the corresponding first data according to the updated second data, and updating the text block and the second embedded vector according to the updated first data.

According to the interactive device of the intelligent customer service system, the corresponding first data is updated according to the updated second data by responding to the second data update, and the text block and the second embedded vector are updated according to the updated first data, so that the data quality and timeliness are improved, and the quality of intelligent answers is improved.

According to the interaction device of the intelligent customer service system provided by the embodiment of the invention, the device further comprises a training module for: obtaining user consultation second content in batches, generating a third embedded vector of the user consultation second content, obtaining at least one second embedded vector meeting a preset similarity requirement with the third embedded vector in the vectorization database, and obtaining the text block corresponding to the second embedded vector; and inputting the second content and the text block consulted by the user as second context data into a language model for automatically generating intelligent answers, and training the language model to obtain the intelligent customer service system.

According to the interaction device of the intelligent customer service system, provided by the embodiment of the invention, the user consultation second content is obtained in batches, the third embedded vector of the user consultation second content is generated, at least one second embedded vector meeting the preset similarity requirement with the third embedded vector in the vectorization database is obtained, the text block corresponding to the second embedded vector is obtained, the user consultation second content and the text block are used as second context data to be input into a language model for automatically generating intelligent answers, the language model is trained, the intelligent customer service system is obtained, the reasoning capacity of the intelligent customer service system for the user consultation content is improved, and the accuracy of the intelligent answers is improved, so that the quality of the intelligent answers is improved.

According to the interaction device of the intelligent customer service system provided by the embodiment of the invention, the training module is further used for: acquiring a correct answer corresponding to the intelligent answer output by the language model when the intelligent answer is a wrong answer; and inputting the second context data into the language model, and continuing training the language model by taking the correct answer as an output label.

According to the interactive device of the intelligent customer service system, the second context data is input into the language model by acquiring the correct answer corresponding to the intelligent answer output by the language model as the wrong answer, and the correct answer is used as the output label, so that the language model is continuously trained, and the quality of the intelligent answer is further improved.

Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, as shown in fig. 4, the electronic device may include: processor 510, communication interface (Communications Interface) 520, memory 530, and communication bus 540, wherein processor 510, communication interface 520, memory 530 complete communication with each other through communication bus 540. Processor 510 may invoke logic instructions in memory 530 to perform an interaction method for an intelligent customer service system, the method comprising: acquiring first content of user consultation; generating a first embedded vector of the user consultation first content, acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in a vectorization database, acquiring a text block corresponding to the second embedded vector, inputting the user consultation first content and the text block as first context data into an intelligent customer service system, and generating an intelligent answer of the user consultation first content.

Further, the logic instructions in the memory 530 described above may be implemented in the form of software functional units and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

In another aspect, an embodiment of the present invention further provides a computer program product, where the computer program product includes a computer program, where the computer program may be stored on a non-transitory computer readable storage medium, where the computer program when executed by a processor, can perform an interaction method of an intelligent customer service system provided by the foregoing methods, where the method includes: acquiring first content of user consultation; generating a first embedded vector of the user consultation first content, acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in a vectorization database, acquiring a text block corresponding to the second embedded vector, inputting the user consultation first content and the text block as first context data into an intelligent customer service system, and generating an intelligent answer of the user consultation first content.

In yet another aspect, an embodiment of the present invention further provides a non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor, is implemented to perform the method for interaction of an intelligent customer service system provided by the above methods, the method comprising: acquiring first content of user consultation; generating a first embedded vector of the user consultation first content, acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in a vectorization database, acquiring a text block corresponding to the second embedded vector, inputting the user consultation first content and the text block as first context data into an intelligent customer service system, and generating an intelligent answer of the user consultation first content.

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An interaction method of an intelligent customer service system is characterized by comprising the following steps:

acquiring first content of user consultation;

generating a first embedded vector of the user consultation first content, acquiring at least one second embedded vector meeting a preset similarity requirement with the first embedded vector in a vectorization database, acquiring a text block corresponding to the second embedded vector, inputting the user consultation first content and the text block as first context data into an intelligent customer service system, and generating an intelligent answer of the user consultation first content.

2. The interaction method of an intelligent customer service system according to claim 1, wherein the method further comprises:

and responding to the first content of the user consultation to contain automatic flow processing information, and triggering a corresponding robot automatic processing flow according to the automatic flow processing information.

3. The interaction method of an intelligent customer service system according to claim 1, wherein the method further comprises:

collecting first data related to questions and answers, carrying out data processing on the first data and generating a formatted document;

and carrying out blocking processing on the formatted document to obtain a text block, generating the second embedded vector of the text block, and storing the second embedded vector and the corresponding text block in the vectorization database in an associated manner.

4. An interaction method for an intelligent customer service system according to claim 3, wherein the collecting first data related to questions and answers comprises:

and simulating manual operation, automatically logging in the enterprise intranet to capture second data related to questions and answers, and positioning and extracting the first data according to the second data.

5. The interaction method of the intelligent customer service system according to claim 4, wherein the automatically logging in the intranet to capture the second data related to the question and answer comprises:

and automatically logging in an enterprise intranet, and capturing relevant contents of a regulation system, a question-answer library and a knowledge base to obtain the second data.

6. The intelligent customer service system interaction method according to claim 5, wherein the method further comprises:

and responding to the second data update, updating the corresponding first data according to the updated second data, and updating the text block and the second embedded vector according to the updated first data.

7. An interaction method for an intelligent customer service system according to claim 3, wherein the method further comprises:

obtaining user consultation second content in batches, generating a third embedded vector of the user consultation second content, obtaining at least one second embedded vector meeting a preset similarity requirement with the third embedded vector in the vectorization database, and obtaining the text block corresponding to the second embedded vector;

and inputting the second content and the text block consulted by the user as second context data into a language model for automatically generating intelligent answers, and training the language model to obtain the intelligent customer service system.

8. The intelligent customer service system interaction method according to claim 7, wherein the method further comprises:

acquiring a correct answer corresponding to the intelligent answer output by the language model when the intelligent answer is a wrong answer;

and inputting the second context data into the language model, and continuing training the language model by taking the correct answer as an output label.

9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the interaction method of the intelligent customer service system according to any of claims 1 to 8 when the program is executed.

10. A non-transitory computer readable storage medium, having stored thereon a computer program, which when executed by a processor, implements the steps of the interaction method of the intelligent customer service system according to any of claims 1 to 8.

11. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the interaction method of the intelligent customer service system according to any of claims 1 to 8.