Intelligent Customer Service Implementation Method, Apparatus, Device, and Storage Medium

Abstract

An embodiment of the present application provides a method, device, equipment, and storage medium for implementing an intelligent customer service. The method includes: obtaining a problem to be processed and a user identifier submitted by a client, inputting the problem to be processed into a set first language model to obtain a set of semantic keywords; calculating the average semantic similarity between the set of semantic keywords and a set of multiple candidate business tags respectively, and determining the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed; querying problem-related information from a set business database according to the target business tag, the set of semantic keywords, and the user identifier; generating a prompt text based on the problem-related information, the problem to be processed, and a set prompt word template, and inputting the prompt text into a set second language model to obtain a target reply message. This solution realizes the improvement of the reply ability of the intelligent customer service, adapts to the diverse question scenarios of different users, and optimizes the user experience.

Description

Technical Field

[0001] Embodiments of the present application relate to the field of computer technology, and in particular, to a method, apparatus, device, and storage medium for implementing an intelligent customer service. Background Art

[0002] With the continuous development of artificial intelligence technology, intelligent customer service has begun to be widely used in more and more industry fields, providing question-and-answer interactive services for customers. Relevant industry fields include e-commerce, finance, education, medical care, and games, etc. For example, in the e-commerce field, intelligent customer service can provide customers with support such as product consultation, order query, and after-sales service; in the game field, intelligent customer service can provide support such as gameplay introduction, operation instructions, and exception query. Based on the fact that intelligent customer service can automatically answer users' questions and replace the work of human customer service, the purpose of saving labor costs can be achieved.

[0003] However, in related technologies, intelligent customer service only replies based on the problem description provided by the user and uses predefined standard answers, resulting in poor problem-solving effects and being unable to adapt to the diverse question-asking scenarios of different users, which makes the user experience poor and needs to be improved. Summary of the Invention

[0004] Embodiments of the present application provide a method, apparatus, device, and storage medium for implementing an intelligent customer service, which solve the problems of poor problem-solving effects and being unable to adapt to the diverse question-asking scenarios of different users, resulting in poor user experience, and achieve accurately determining the relevant business field of the to-be-processed problem, reasonably expanding the background information of the to-be-processed problem, being beneficial to helping the language model more accurately identify the user's intention, giving an accurate reply, improving the reply ability of the intelligent customer service, adapting to the diverse question-asking scenarios of different users, and optimizing the user experience.

[0005] In a first aspect, an embodiment of the present application provides a method for implementing an intelligent customer service, and the method includes:

[0006] Obtain a to-be-processed problem and a user identifier submitted by a client, and input the to-be-processed problem into a set first language model to obtain a set of semantic keywords;

[0007] Calculate the average semantic similarity between the set of semantic keywords and a set of multiple candidate business tags respectively, and determine the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the to-be-processed problem;

[0008] Query problem-related information from a set business database according to the target business tag, the set of semantic keywords, and the user identifier;

[0009] Generate a prompt text based on the problem-related information, the problem to be processed, and the set prompt template, and input the prompt text into the set second language model to obtain the target response information.

[0010] In a second aspect, an intelligent customer service implementation device provided by an embodiment of the present application includes:

[0011] An acquisition module configured to acquire the problem to be processed submitted by the client and the user identifier;

[0012] A keyword generation module configured to input the problem to be processed into the set first language model to obtain a semantic keyword set;

[0013] A label determination module configured to calculate the average semantic similarity between the semantic keyword set and a set of multiple candidate business labels respectively, and determine the candidate business label with the highest average semantic similarity as the target business label corresponding to the problem to be processed;

[0014] An information determination module configured to query the problem-related information from the set business database according to the target business label, the semantic keyword set, and the user identifier;

[0015] A response determination module configured to generate a prompt text based on the problem-related information, the problem to be processed, and the set prompt template, and input the prompt text into the set second language model to obtain the target response information.

[0016] In a third aspect, an intelligent customer service implementation device provided by an embodiment of the present application includes:

[0017] One or more processors;

[0018] A storage device configured to store one or more programs,

[0019] When the one or more programs are executed by the one or more processors, the one or more processors implement the intelligent customer service implementation method described in the embodiments of the present application.

[0020] In a fourth aspect, an embodiment of the present application further provides a non-volatile storage medium storing computer-executable instructions, and the computer-executable instructions are configured to execute the intelligent customer service implementation method described in the embodiments of the present application when executed by a computer processor.

[0021] In the embodiments of the present application, by obtaining the problem to be processed and the user identifier submitted by the client, inputting the problem to be processed into the set first language model to obtain a set of semantic keywords; calculating the average semantic similarity between the set of semantic keywords and multiple set candidate business tags respectively, and determining the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed; querying problem-related information from the set business database according to the target business tag, the set of semantic keywords and the user identifier; generating a prompt text based on the problem-related information, the problem to be processed and the set prompt word template, and inputting the prompt text into the set second language model to obtain the target reply information. In the above solution, by inputting the problem to be processed into the set first language model to obtain a set of semantic keywords, the key information of the problem to be processed can be effectively captured, and the user's question requirement can be accurately identified; by matching the corresponding target business tag for the problem to be processed based on the average semantic similarity and querying the problem-related information, the relevant business field of the problem to be processed can be accurately determined, the background information of the problem to be processed can be reasonably expanded, which is beneficial to helping the language model to more accurately identify the user's intention, give an accurate reply, improve the reply ability of the intelligent customer service, adapt to the diverse question scenarios of different users, and optimize the user experience. Description of the Drawings

[0022] Figure 1 It is a flowchart of a method for implementing an intelligent customer service provided by an embodiment of the present application;

[0023] Figure 2 It is a flowchart of a method for implementing an intelligent customer service including the process of calculating the average semantic similarity provided by an embodiment of the present application;

[0024] Figure 3 It is a flowchart of a method for implementing an intelligent customer service including the process of generating candidate business tags provided by an embodiment of the present application;

[0025] Figure 4 It is a flowchart of a method for implementing an intelligent customer service including the process of determining the problem-related information of the problem to be processed provided by an embodiment of the present application;

[0026] Figure 5 It is a flowchart of a method for implementing an intelligent customer service including the process of extracting problem-related information provided by an embodiment of the present application;

[0027] Figure 6 It is a flowchart of a method for implementing an intelligent customer service including the process of generating a prompt text provided by an embodiment of the present application;

[0028] Figure 7 It is a structural block diagram of an intelligent customer service implementation device provided by an embodiment of the present application;

[0029] Figure 8 A structural schematic diagram of an intelligent customer service implementation device provided by an embodiment of the present application. Detailed implementation manners

[0030] The following further elaborates on the embodiments of the present application in conjunction with the accompanying drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the embodiments of the present application, rather than limiting the embodiments of the present application. Additionally, it should be noted that for ease of description, only parts related to the embodiments of the present application are shown in the drawings, rather than all structures.

[0031] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, rather than to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein. Moreover, the objects distinguished by "first", "second", etc. are generally of the same category, and the number of objects is not limited. For example, the first object can be one or multiple. In addition, "and / or" in the specification and claims indicates at least one of the connected objects. The character " / " generally represents an "or" relationship between the associated objects before and after.

[0032] The intelligent customer service implementation method provided by the embodiments of the present application is used to identify relevant business tags for a to-be-processed problem provided by a user, then supplement the corresponding problem-related information, and finally generate a more accurate reply message through a set language model. Relevant industry fields include e-commerce, finance, education, medical care, and games, etc. Taking games as an example, the intelligent customer service can provide intelligent reply back-end support for gameplay introductions, operation instructions, exception queries, etc. The several application scenarios listed above are only exemplary and explanatory. In actual applications, this intelligent customer service implementation method can also be used in intelligent customer services in other scenarios. The embodiments of the present application do not limit this. The present application aims to provide an intelligent customer service implementation method, device, equipment, and storage medium to solve the problem of poor problem-solving effects, inability to adapt to the diverse question scenarios of different users, and poor user experience.

[0033] For the intelligent customer service implementation method provided by the embodiments of the present application, the execution subject of each step can be a computer device. The computer device refers to any electronic device with data calculation, processing, and storage capabilities, such as terminal devices like mobile phones, PCs (Personal Computers), and tablet computers, or devices such as servers. The embodiments of the present application do not limit this.

[0034] Figure 1The figure is a flowchart of an intelligent customer service implementation method provided by an embodiment of the present application. The intelligent customer service implementation method can be implemented with a background server as the execution entity. As Figure 1 shown, the intelligent customer service implementation method specifically includes the following steps:

[0035] Step S101: Obtain the problem to be processed and the user identifier submitted by the client, and input the problem to be processed into the set first language model to obtain a semantic keyword set.

[0036] Among them, the user can send a question request to the background server through the interaction interface of the client. The question request can include the problem to be processed and the user identifier. The background server can temporarily store the question request in the message queue. The background server can sequentially retrieve the currently required question request from the message queue according to the priority order or the time sequence, and obtain the problem to be processed and the user identifier therein. The problem to be processed can be the problem description text submitted by the user through the interaction interface of the client. Taking the game field as an example, the relevant scenarios of the problem to be processed can be account login anomalies, uncredited recharge, stolen game items, etc. For example, the user asks "I recharged but didn't get the items. I recharged a package worth X amount at time T, but didn't get the items. Please check. None of the Y game items have entered my game account!!!" The first language model can be a pre-trained BERT (Bidirectional Encoder Representations from Transformers), GPT (Generative Pre-trained Transformer), etc. Taking the game field as an example, the user problem description information and keyword annotation information of the collected historical records can be used as the training data set to train the first language model, calculate the loss function value of the preset loss function, and perform iterative update of the network parameters of the first language model through the backpropagation algorithm and related optimization algorithms until the loss converges, so that the first language model has the ability to identify the problem to be processed and output the semantic keyword set. The semantic keyword set can be the keyword information of the business field that can be used to identify its attribution in the problem to be processed. For example, for the problem to be processed: "I recharged but didn't get the items. I recharged a package worth X amount at time T, but didn't get the items. Please check", the semantic keyword set obtained after inputting it into the first language model can include "recharge", "amount", "package", and "time T".

[0037] Step S102: Calculate the average semantic similarity between the semantic keyword set and the set multiple candidate business tags respectively, and determine the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed.

[0038] Among them, the candidate service tags can be pre-generated by developers according to the service categories involved in the current application scenario. For example, taking the game field as an example, the candidate service tags can be "Login" set for the corresponding login service, "Payment" set for the corresponding payment service, and "Update" set for the corresponding update service, etc. The average semantic similarity can be used to measure the average similarity degree between each candidate service tag and all semantic keywords in the semantic keyword set. For example, the semantic keyword set includes "Recharge", "Amount", "Gift Package", and "T Time". Correspondingly, the average semantic similarity calculated with "Payment" is the highest. In one embodiment, each candidate service tag can calculate multiple semantic similarities with the semantic keywords in the semantic keyword set one by one, and take the average of the multiple semantic similarities to obtain the average semantic similarity. Thus, the average semantic similarity between each candidate service tag and the semantic keyword set can be obtained. In one embodiment, the multiple semantic similarities calculated between each candidate service tag and the semantic keywords in the semantic keyword set can be filtered for extreme values, and the corresponding average language similarity can be calculated based on the remaining semantic similarities. In one embodiment, the CBOW (Continuous Bag of Words) model, Skip-gram model, etc. can be used to convert the semantic keywords and candidate service tags into corresponding word vectors, and the similarity between the word vectors corresponding to the semantic keywords and candidate service tags can be determined by calculating the cosine similarity or Euclidean distance, etc., so as to calculate the average value of the multiple similarities corresponding to each candidate service tag and the semantic keyword set to obtain the average semantic similarity.

[0039] Step S103: Query the problem association information from the set service database according to the target service tag, the semantic keyword set, and the user identifier.

[0040] Among them, the business database can store user behavior information corresponding to different business tags for different users. Taking the game field as an example, for the payment business, it can store user order logs and error record logs, which are used to record the whole process of order operations performed by users in the platform or application, including but not limited to the time, status, operator, etc. of important information in each link such as order creation, modification, payment, delivery, receipt, refund, etc. For the login business, it can store user login logs and error record logs, which are used to record the login activities of users on the system or platform, including but not limited to login time, login location, login method, login result, and possible failure reasons such as incorrect password, account locked, etc. Of course, the business database can also have other business settings and information recording methods, which are not limited in this application. In one embodiment, the target business tag and the user identifier can be used as an index to search for relevant user information files, and the behavior information records matching the semantic keywords in the semantic keyword set can be extracted from the user information files as problem-related information. For example, the target business tag is "payment", the semantic keyword set includes "recharge", "amount", "gift package", and "T time", and the user identifier is "A1". Correspondingly, based on "payment" and "A1", relevant user order logs and error record logs can be searched. Based on the semantic keyword set, the following relevant first behavior information records can be extracted from the user order logs: {User: A1, Payment channel: WeChat, Payment amount: x yuan, Payment result: Paid, Payment time: T time, Shipped goods: Z gift package, Shipment result: Not shipped}, and the following relevant second behavior information records can be extracted from the error record logs: {User: A1, Shipped goods: Z gift package, Shipment result: Not shipped, Error reason: Shipment interface call timed out}. The first behavior information record and the second behavior information record can be used as the problem-related information of the to-be-processed problem. In one embodiment, the target business tag and the semantic keyword set can be used as an index to perform a full-database content search to obtain at least one behavior information record, and the target behavior information record corresponding to the user identifier can be filtered out from the at least one behavior information record as problem-related information.

[0041] Step S104: Generate a prompt text based on the problem-related information, the to-be-processed problem, and the set prompt word template, and input the prompt text into the set second language model to obtain the target reply information.

[0042] Among them, the second language model can be a pre-trained BERT model, GPT model, etc. Developers can collect different user questions in historical records and corresponding diverse artificial response expressions as a training data set, train the second language model, calculate the loss function value of a preset loss function, and perform iterative updates on the network parameters of the second language model through the backpropagation algorithm and related optimization algorithms until loss convergence is achieved, so that the second language model has the ability to output user-friendly reply information. The prompt template can be a structured prompt pre-constructed by developers to guide the second language model to generate text content that meets requirements. Among them, placeholders representing key information to be filled in can be set in the prompt template. In one embodiment, the question correlation information and the question to be processed can be filled into the placeholder positions preset in the prompt template respectively to obtain a prompt text. In one embodiment, the question correlation information can be split into multiple key information, the associated text positions of each key information in the question to be processed are determined according to semantic similarity, and the key information is added to the corresponding associated text positions to obtain an expanded question to be processed. Finally, the expanded question to be processed is filled into the placeholder positions preset in the prompt template to obtain a prompt text.

[0043] As described above, by obtaining the question to be processed and the user identifier submitted by the client, inputting the question to be processed into the set first language model to obtain a set of semantic keywords; calculating the average semantic similarity between the set of semantic keywords and each of the set of candidate business tags respectively, and determining the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the question to be processed; querying the question correlation information from the set business database according to the target business tag, the set of semantic keywords, and the user identifier; generating a prompt text based on the question correlation information, the question to be processed, and the set prompt template, and inputting the prompt text into the set second language model to obtain the target reply information. In the above solution, by inputting the question to be processed into the set first language model to obtain a set of semantic keywords, the key information of the question to be processed can be effectively captured, and the user's question needs can be accurately identified; by matching the corresponding target business tag for the question to be processed based on the average semantic similarity and querying the question correlation information, the relevant business field of the question to be processed can be accurately determined, and the background information of the question to be processed can be reasonably expanded, which is beneficial to helping the language model more accurately identify the user's intention, give an accurate reply, improve the reply ability of the intelligent customer service, adapt to the diverse question scenarios of different users, and optimize the user experience.

[0044] Figure 2 The flowchart of an intelligent customer service implementation method including the process of calculating the average semantic similarity provided by an embodiment of the present application. As Figure 2 shown, the intelligent customer service implementation method specifically includes the following steps:

[0045] Step S201: Obtain the problem to be processed and the user identifier submitted by the client, and input the problem to be processed into the set first language model to obtain a set of semantic keywords.

[0046] Step S202: Query the first vocabulary description corresponding to each semantic keyword in the set of semantic keywords from the set vocabulary library, and the second vocabulary description corresponding to each of the set of candidate business tags.

[0047] Among them, the vocabulary library can record multiple words related to the target business field and the vocabulary description corresponding to each word. The vocabulary description is used to provide associated information such as the definition and context background of the word, which is beneficial to more accurately calculating the semantic similarity between the semantic keyword and the candidate business tag subsequently.

[0048] Step S203: Convert the first vocabulary description and the second vocabulary description into corresponding first text vectors and second text vectors respectively.

[0049] Among them, the first vocabulary description and the second vocabulary description can be converted into text vector representations through models such as the CBOW model, Skip-gram model, or GloVe (Global Vectors for Word Representation) model. This application does not make limitations here.

[0050] Step S204: Calculate the similarity between each of the multiple first text vectors corresponding to the set of semantic keywords and the second text vector corresponding to each candidate business tag, and take the average of the obtained similarity results to obtain the average semantic similarity corresponding to each candidate business tag.

[0051] Among them, the similarity between the vectors can be determined by calculating the cosine value of the angle between the first text vector and the second text vector, the Euclidean distance, or the Pearson correlation coefficient. This application does not make limitations here. Multiple similarities can be calculated respectively between the second text vector corresponding to each candidate business tag and the multiple first text vectors corresponding to the set of semantic keywords, and the average of the obtained multiple similarities can be taken to obtain the average semantic similarity between each candidate business tag and the set of semantic keywords, which is used to characterize the similarity degree between each candidate business tag and the problem to be processed.

[0052] Step S205: Determine the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed.

[0053] Step S206: Query the problem-related information from the set business database according to the target business tag, the set of semantic keywords, and the user identifier.

[0054] Step S207: Generate a prompt text based on the problem association information, the problem to be processed, and the set prompt word template, and input the prompt text into the set second language model to obtain the target response information.

[0055] As described above, by querying the first vocabulary description and the second vocabulary description corresponding to the semantic keyword and the candidate business label respectively, more comprehensive text description information can be effectively provided, and the average semantic similarity between the semantic keyword set and the candidate business label can be calculated more accurately based on the vector representations of the first vocabulary description and the second vocabulary description.

[0056] Figure 3 This is a flowchart of an intelligent customer service implementation method provided by an embodiment of the present application, which includes a process of generating candidate business labels. As Figure 3 shown, the intelligent customer service implementation method specifically includes the following steps:

[0057] Step S301: Obtain the problem to be processed submitted by the client and the user identifier, and input the problem to be processed into the set first language model to obtain a semantic keyword set.

[0058] Step S302: Obtain multiple historical conversations within a preset time range, and extract the first preset number of keywords with the highest occurrence frequency from the multiple historical conversations.

[0059] Among them, the preset time range can be used to determine the reference range of valid historical conversation data, and the historical conversation can be the recorded Q&A information between the user and the intelligent customer service or the manual customer service. In one embodiment, the keyword extraction can be obtained by identifying nouns, verbs, etc. that are the main components of the sentence through grammar rules. In one embodiment, the keyword extraction can be to search for the vocabulary recorded in the set keyword library from the historical conversation. By counting the occurrence frequency of each keyword that appears in the historical conversation, the keywords are sorted in ascending or descending order according to the occurrence frequency, and the first preset number of keywords with the highest occurrence frequency are selected.

[0060] Step S303: Calculate the semantic similarity between the keywords and multiple preset business labels respectively, and add the keywords to the keyword set associated with the preset business label with the highest semantic similarity.

[0061] In one embodiment, the CBOW model, the Skip-gram model, etc. can be used to convert the keywords and the preset business labels into corresponding vectors respectively, and the similarity between the keywords and the preset business labels can be determined by calculating the cosine value of the vector angle, the Euclidean distance, or the Pearson correlation coefficient. By adding the keywords to the keyword set associated with the preset business label with the highest semantic similarity, it is beneficial to count the number of keywords that appear in the historical conversation associated with each preset business label.

[0062] Step S304: Screen out a second preset number of candidate service tags with the largest number of associated keywords from multiple preset service tags.

[0063] Among them, the number of keywords in the keyword set corresponding to each preset service tag can be counted, sorted in ascending or descending order according to the number of associated keywords, and a second preset number of candidate service tags with the largest number of associated keywords are screened out.

[0064] Step S305: Calculate the average semantic similarity between the semantic keyword set and each of the set of candidate service tags, and determine the candidate service tag with the highest average semantic similarity as the target service tag corresponding to the problem to be processed.

[0065] Step S306: Query problem-related information from the set business database according to the target service tag, the semantic keyword set, and the user identifier.

[0066] Step S307: Generate a prompt text based on the problem-related information, the problem to be processed, and the set prompt template, and input the prompt text into the set second language model to obtain the target reply information.

[0067] As described above, by extracting keywords based on historical conversations, the key information of the questions that users often ask can be effectively captured. By screening tags by counting the number of keywords associated with each preset service tag, preset service tags that users basically do not pay attention to or have low relevance can be effectively filtered out, which is beneficial to improving the efficiency of subsequent matching of the target service tag for the problem to be processed.

[0068] Figure 4 This is a flowchart of an intelligent customer service implementation method provided by an embodiment of the present application, which includes a process of determining problem-related information of a problem to be processed. As Figure 4 shown, the intelligent customer service implementation method specifically includes the following steps:

[0069] Step S401: Obtain the problem to be processed submitted by the client and the user identifier, and input the problem to be processed into the set first language model to obtain a semantic keyword set;

[0070] Step S402: Calculate the average semantic similarity between the semantic keyword set and each of the set of candidate service tags, and determine the candidate service tag with the highest average semantic similarity as the target service tag corresponding to the problem to be processed;

[0071] Step S403: Query the target file identifier and the target storage path from the set association information table according to the target service tag.

[0072] Among them, the association information table can be a pre-generated one that records the target file identifier and the target storage path of the target log file corresponding to each service label. The number of the target file identifier and the target storage path can be one or more, which are used to point to one or more target log files that meet the service label. These target log files can be log files including multiple users, or log files of the same user that record different types of information. The target log file can independently record the historical behavior information of the same user, and different types of information log files can be set for each user. For example, taking the login service as an example, user login logs and error record logs can be set for each user.

[0073] In one embodiment, the association information table can be pre-generated, and the specific process is as follows:

[0074] Obtain multiple candidate service labels set, query the storage paths of the log files associated with each candidate service label from the set business database, and construct an association information table based on each candidate service label, the file identifier of the associated log file, and the storage path.

[0075] Step S404: Query at least one target log file from the set business database according to the target file identifier and the target storage path.

[0076] Among them, the target file identifier and the target storage path can be used as index information to query at least one target log file pointed to by them from the business database.

[0077] Step S405: Extract problem association information that matches the user identifier and the semantic keyword set from at least one target log file.

[0078] Among them, the log files corresponding to the current user can be filtered out from at least one target log file based on the user identifier, and the behavior information records containing the semantic keyword set can be extracted from the filtered log files as the problem association information.

[0079] Step S406: Generate a prompt text based on the problem association information, the problem to be processed, and the set prompt word template, and input the prompt text into the set second language model to obtain the target reply information.

[0080] As described above, by pre-setting the association information table that records the file identifier and the storage path corresponding to the log file, it can be used to quickly locate and obtain the target log file required, improving the query efficiency of the target log file.

[0081] Figure 5 This is a flowchart of an intelligent customer service implementation method provided by an embodiment of the present application, which includes the process of extracting problem association information. AsFigure 5 As shown in the figure, the implementation method of the intelligent customer service specifically includes the following steps:

[0082] Step S501: Obtain the problem to be processed submitted by the client and the user identifier, and input the problem to be processed into the set first language model to obtain a set of semantic keywords;

[0083] Step S502: Calculate the average semantic similarity between the set of semantic keywords and multiple set candidate business tags respectively, and determine the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed;

[0084] Step S503: Query the target file identifier and the target storage path from the set association information table according to the target business tag.

[0085] Step S504: Query at least one target log file from the set business database according to the target file identifier and the target storage path.

[0086] Step S505: Extract the associated log file corresponding to the user identifier from at least one target log file.

[0087] Among them, the associated log file can be one or more, and can record the behavior information of the target user corresponding to the user identifier. Taking the payment business as an example, the associated log file can include the user order log and the error record log. Taking the update business as an example, the associated log file can include the user update log and the error record log.

[0088] Step S506: Retrieve at least one target record item from the associated log file based on the set of semantic keywords.

[0089] Among them, the target record item containing the semantic keywords in the set of semantic keywords can be retrieved from the associated log file to supplement the context information of the problem to be processed. For example, the set of semantic keywords includes "recharge", "amount", "gift package", and "T time". Then, the relevant behavior information records can be extracted from the user order log: {User: A1, Payment channel: WeChat, Payment amount: x yuan, Payment result: Paid, Payment time: T time, Shipped goods: Z gift package, Shipped result: Not shipped}.

[0090] Step S507: Integrate at least one target record item to obtain problem-related information.

[0091] In one embodiment, if there are multiple target record items, the multiple target record items can be concatenated to obtain problem-related information. In one embodiment, if there are multiple target record items, the duplicate information in the multiple target record items can be removed and merged to obtain problem-related information.

[0092] Step S508: Generate a prompt text based on the problem-related information, the problem to be processed, and the set prompt word template, and input the prompt text into the set second language model to obtain the target response information.

[0093] As described above, by retrieving and extracting problem-related information from the associated log file based on the semantic keyword set, accurate context information can be provided for the problem to be processed, effectively expanding the information background of the problem to be processed, and improving the answering effect of the subsequent generated response information.

[0094] Figure 6 It is a flowchart of an intelligent customer service implementation method provided by an embodiment of the present application, which includes a process of generating a prompt text. As Figure 6 shown, the intelligent customer service implementation method specifically includes the following steps:

[0095] Step S601: Obtain the problem to be processed submitted by the client and the user identifier, and input the problem to be processed into the set first language model to obtain a semantic keyword set.

[0096] Step S602: Calculate the average semantic similarity between the semantic keyword set and each of the set multiple candidate business tags, and determine the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed.

[0097] Step S603: Query the problem-related information from the set business database according to the target business tag, the semantic keyword set, and the user identifier.

[0098] Step S604: Integrate and process the problem-related information and the problem to be processed to obtain problem description information.

[0099] Among them, the integration process can be to remove duplicates and fuse relevant information in the problem-related information and the problem to be processed. For example, the problem to be processed is: "I recharged but didn't get the items. I recharged a package worth X yuan at time T, but didn't get the items. Please check. None of the Y game items have entered my game account!!!", the first line of information record is: {User: A1, Payment channel: WeChat, Payment amount: x yuan, Payment result: Paid, Payment time: T time, Shipped goods: Z package, Shipment result: Not shipped}, the second line of information record is: {User: A1, Shipped goods: Z package, Shipment result: Not shipped, Error reason: Shipment interface call timed out}, and the problem description information obtained by the integration process can be: {User: A1, Payment channel: WeChat, Payment amount: x yuan, Payment result: Paid, Payment time: T time, Shipped goods: Z package, Shipment result: Not shipped, Error reason: Shipment interface call timed out}.

[0100] Step S605: Query the reference answer information that matches the problem description information from the set answer database.

[0101] Among them, the answer database can store the standardized reference answer information of multiple user questions, providing basic reference data for the second language model to generate reply information.

[0102] Step S606: Fill the reference answer information and the problem description information into the set prompt template to generate a prompt text.

[0103] Step S607: Input the prompt text into the set second language model to obtain the target reply information.

[0104] As described above, by filling the reference answer information and the problem description information into the set prompt template to generate a prompt text, the generated reference answer information can be effectively utilized to provide a reliable reference direction for the second language model to generate reply information, improving the accuracy of the target reply information and enabling the effective solution of user problems.

[0105] Figure 7 It is a structural block diagram of an intelligent customer service implementation device provided by an embodiment of the present application. The device is configured to execute the intelligent customer service implementation method provided by the above embodiment, and has corresponding functional modules and beneficial effects for executing the method. As Figure 7 shown, the device specifically includes:

[0106] An acquisition module 101, configured to acquire the problem to be processed and the user identifier submitted by the client;

[0107] A keyword generation module 102, configured to input the problem to be processed into the set first language model to obtain a semantic keyword set;

[0108] A label determination module 103, configured to calculate the average semantic similarity between the semantic keyword set and multiple set candidate service labels respectively, and determine the candidate service label with the highest average semantic similarity as the target service label corresponding to the problem to be processed;

[0109] An information determination module 104, configured to query the problem-related information from the set business database according to the target service label, the semantic keyword set and the user identifier;

[0110] A reply determination module 105, configured to generate a prompt text based on the problem-related information, the problem to be processed and the set prompt template, and input the prompt text into the set second language model to obtain the target reply information.

[0111] As described above, by obtaining the problem to be processed submitted by the client and the user identifier, inputting the problem to be processed into the set first language model, a set of semantic keywords is obtained; calculating the average semantic similarity between the set of semantic keywords and each of the set of candidate business tags respectively, and determining the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed; querying problem-related information from the set business database according to the target business tag, the set of semantic keywords and the user identifier; generating a prompt text based on the problem-related information, the problem to be processed and the set prompt word template, and inputting the prompt text into the set second language model to obtain the target reply information. In the above solution, by inputting the problem to be processed into the set first language model to obtain a set of semantic keywords, the key information of the problem to be processed can be effectively captured, and the user's question requirements can be accurately identified; by matching the corresponding target business tag for the problem to be processed based on the average semantic similarity and querying the problem-related information, the relevant business field of the problem to be processed can be accurately determined, and the background information of the problem to be processed can be reasonably expanded, which is beneficial to helping the language model to more accurately identify the user's intention, give an accurate reply, improve the reply ability of the intelligent customer service, adapt to the diverse question scenarios of different users, and optimize the user experience.

[0112] In a possible embodiment, the tag determination module 103 is further configured to:

[0113] Query the first vocabulary description corresponding to each semantic keyword in the set of semantic keywords from the set vocabulary library, and the second vocabulary description corresponding to each of the set of candidate business tags respectively;

[0114] Convert the first vocabulary description and the second vocabulary description into corresponding first text vectors and second text vectors respectively;

[0115] Calculate the similarity between each of the multiple first text vectors corresponding to the set of semantic keywords and the second text vector corresponding to each candidate business tag, and take the average of the obtained similarity results to obtain the average semantic similarity corresponding to each candidate business tag.

[0116] In a possible embodiment, it further includes a tag generation module, configured to:

[0117] Obtain multiple historical conversations within a preset time range, and extract the first preset number of keywords with the highest frequency of occurrence from the multiple historical conversations;

[0118] Calculate the semantic similarity between the keywords and each of the multiple preset business tags respectively, and add the keywords to the keyword set associated with the preset business tag with the highest semantic similarity;

[0119] Select the second preset number of candidate business tags with the largest number of associated keywords from multiple preset business tags.

[0120] In a possible embodiment, the information determination module 104 is further configured to:

[0121] Query the target file identifier and the target storage path from the set association information table according to the target business tag;

[0122] Query at least one target log file from the set business database according to the target file identifier and the target storage path;

[0123] Extract the problem association information matching the user identifier and the semantic keyword set from at least one target log file.

[0124] In a possible embodiment, the information determination module 104 is further configured to:

[0125] Extract the associated log file corresponding to the user identifier from at least one target log file;

[0126] Retrieve at least one target record item from the associated log file based on the semantic keyword set;

[0127] Integrate at least one target record item to obtain the problem association information.

[0128] In a possible embodiment, it further includes an information table generation module, configured to:

[0129] Obtain multiple candidate business tags set;

[0130] Query the storage path of the log file associated with each candidate business tag from the set business database;

[0131] Construct an association information table based on each candidate business tag, the file identifier of the associated log file, and the storage path.

[0132] In a possible embodiment, the reply determination module 105 is further configured to:

[0133] Integrate the problem association information and the problem to be processed to obtain the problem description information;

[0134] Query the reference answer information matching the problem description information from the set answer database;

[0135] Fill the reference answer information and the problem description information into the set prompt word template to generate the prompt word text.

[0136] Figure 8The following is a schematic structural diagram of an intelligent customer service implementation device provided by an embodiment of the present application. As Figure 8 shown, the device includes a processor 201, a memory 202, an input device 203, and an output device 204. The number of processors 201 in the device can be one or more. Figure 8 Here, one processor 201 is taken as an example. The processor 201, the memory 202, the input device 203, and the output device 204 in the device can be connected through a bus or other means. Figure 8 Here, connection through a bus is taken as an example. The memory 202, as a computer-readable storage medium, can be configured to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the intelligent customer service implementation method in the embodiments of the present application. The processor 201 executes various functional applications and data processing of the device by running the software programs, instructions, and modules stored in the memory 202, that is, implements the above-mentioned intelligent customer service implementation method. The input device 203 can be configured to receive input digital or character information, and generate key signal inputs related to the user settings and function control of the device. The output device 204 can include a display device such as a display screen.

[0137] The above-provided intelligent customer service implementation device can be used to execute the intelligent customer service implementation method provided in any of the above embodiments, and has corresponding functions and beneficial effects.

[0138] An embodiment of the present application also provides a non-volatile storage medium containing computer-executable instructions. When the computer-executable instructions are executed by a computer processor, they are configured to execute an intelligent customer service implementation method described in one of the above embodiments, including: obtaining a problem to be processed and a user identifier submitted by a client, inputting the problem to be processed into a set first language model to obtain a set of semantic keywords; calculating the average semantic similarity between the set of semantic keywords and a set of multiple candidate service tags respectively, and determining the candidate service tag with the highest average semantic similarity as the target service tag corresponding to the problem to be processed; querying problem-related information from a set business database according to the target service tag, the set of semantic keywords, and the user identifier; generating a prompt text based on the problem-related information, the problem to be processed, and a set prompt word template, and inputting the prompt text into a set second language model to obtain a target reply message.

[0139] Storage Medium - Any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media such as CD-ROMs, floppy disks, or magnetic tape devices; computer system memory or random access memory such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory such as flash memory, magnetic media, optical storage; registers or other similar types of memory elements, etc. The storage medium may also include other types of memory or combinations thereof. Additionally, the storage medium may be located in a first computer system in which the program is executed, or may be located in a different second computer system that is connected to the first computer system via a network such as the Internet. The second computer system may provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media residing in different locations (e.g., in different computer systems connected via a network). The storage medium may store program instructions executable by one or more processors (e.g., embodied as a computer program).

[0140] Of course, for a storage medium containing computer-executable instructions provided by an embodiment of the present application, the computer-executable instructions are not limited to the intelligent customer service implementation method as described above, and may also execute related operations in the intelligent customer service implementation methods provided by any embodiment of the present application.

[0141] It should be noted that the numbering of each step in this solution is only used to describe the overall design framework of this solution, and does not represent an inevitable sequence between steps. On the basis that the overall implementation process conforms to the overall design framework of this solution, it all falls within the protection scope of this solution. The sequential order in the form of words during description is not an exclusive limitation on the specific implementation process of this solution. Those skilled in the art should understand that the embodiments of the present application may be provided as a method, a system, or a computer program product. In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM), and / or non-volatile memory in the form of, for example, read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0142] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, commodity or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising the element.

[0143] Note that the above is only a preferred embodiment of the present invention and the technical principles applied. Those skilled in the art will understand that the present invention is not limited to the specific embodiments here, and various obvious changes, re-adjustments and substitutions can be made by those skilled in the art without departing from the protection scope of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments only. Without departing from the concept of the present invention, more other equivalent embodiments can be included, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for implementing intelligent customer service, characterized in that: include: Obtaining a pending question and a user identifier submitted by a client, inputting the pending question into a set first language model, and obtaining a semantic keyword set; Acquire multiple historical conversations within a preset time range, and extract a first preset number of keywords with the highest occurrence frequency from the multiple historical conversations; Calculating semantic similarities between the keyword and a plurality of preset service tags respectively, and adding the keyword to a keyword set associated with the preset service tag having the highest semantic similarity; Filtering out a second preset number of candidate business tags having the largest number of associated keywords from the plurality of preset business tags; Calculate the average semantic similarity between the semantic keyword set and multiple candidate business tags respectively, and determine the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed; According to the target service tag, a target file identifier and a target storage path are obtained by querying from a set association information table; Obtain at least one target log file from a set business database according to the target file identifier and the target storage path; Extracting question-related information matching the user identifier and the semantic keyword set from the at least one target log file; A prompt word text is generated based on the question association information, the question to be processed and a set prompt word template, and the prompt word text is input into a set second language model to obtain target answer information.

2. The intelligent customer service implementation method according to claim 1, characterized in that: The calculating the average semantic similarity between the semantic keyword set and a plurality of set candidate service tags respectively includes: Querying a first vocabulary description corresponding to each semantic keyword in the semantic keyword set and second vocabulary descriptions corresponding to each of the plurality of candidate service tags from a set vocabulary library; Convert the first vocabulary description and the second vocabulary description into a corresponding first text vector and a second text vector respectively; Similarities are calculated between the multiple first text vectors corresponding to the semantic keyword set and the second text vector corresponding to each candidate business tag, and the obtained similarity results are averaged to obtain an average semantic similarity corresponding to each candidate business tag.

3. The intelligent customer service implementation method according to claim 1, characterized in that: The extracting the question-related information matching the user identifier and the semantic keyword set from the at least one target log file includes: Extracting an associated log file corresponding to the user identifier from the at least one target log file; Retrieving the associated log file based on the semantic keyword set to obtain at least one target record item; The at least one target record item is integrated to obtain problem-related information.

4. The intelligent customer service implementation method according to claim 3, characterized in that: Before obtaining the target file identifier and the target storage path from the set association information table according to the target service tag, the method further includes: Get multiple candidate business tags that are set; Querying a storage path of a log file associated with each candidate service tag from a set service database; A correlation information table is constructed based on each of the candidate service tags, the file identifier of the associated log file, and the storage path.

5. The method for implementing intelligent customer service according to any one of claims 1 to 4, characterized in that: The generating of prompt word text based on the question association information, the question to be processed and the set prompt word template includes: Integrate the problem-related information and the problem to be processed to obtain problem description information; Querying a set answer database for reference answer information that matches the question description information; Fill the reference answer information and the question description information into the set prompt word template to generate a prompt word text.

6. An intelligent customer service implementation device, characterized in that: include: An acquisition module configured to acquire pending issues and user identifiers submitted by a client; A keyword generation module, configured to input the problem to be processed into a set first language model to obtain a semantic keyword set; a tag generation module configured to obtain multiple historical conversations within a preset time range, extract a first preset number of keywords with the highest frequency of occurrence from the multiple historical conversations, calculate semantic similarities between the keywords and multiple preset business tags respectively, and add the keywords to a keyword set associated with the preset business tag with the highest semantic similarity, and select a second preset number of candidate business tags with the largest number of associated keywords from the multiple preset business tags; A tag determination module is configured to calculate the average semantic similarity between the semantic keyword set and a plurality of set candidate business tags, and determine the candidate business tag with the highest average semantic similarity as the target business tag corresponding to the problem to be processed; an information determination module, configured to query a set association information table to obtain a target file identifier and a target storage path according to the target business tag, query a set business database to obtain at least one target log file according to the target file identifier and the target storage path, and extract problem association information matching the user identifier and the semantic keyword set from the at least one target log file; The answer determination module is configured to generate a prompt word text based on the question association information, the question to be processed and a set prompt word template, and input the prompt word text into a set second language model to obtain target answer information.

7. An intelligent customer service implementation device, the device comprising: one or more processors; A storage device is configured to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the intelligent customer service implementation method described in any one of claims 1 to 5.

8. A non-volatile storage medium storing computer executable instructions, wherein the computer executable instructions are configured to execute the intelligent customer service implementation method described in any one of claims 1 to 5 when executed by a computer processor.

Images