Information processing method and device, equipment and storage medium

Abstract

The invention provides an information processing method and device, equipment and a storage medium, and relates to the technical field of computers, in particular to the technical fields of artificial intelligence, data processing, large models, intelligent customer service and the like. According to the specific implementation scheme, multiple pieces of artificial question and answer data are obtained, wherein the artificial question and answer data are generated by an intelligent customer service system and an artificial customer service system; analyzing and clustering the artificial question and answer data, and establishing a knowledge module based on an analysis and clustering result; updating a knowledge base of the intelligent customer service system based on the knowledge module; and reconfiguring a recommendation strategy of the intelligent customer service system based on the updated knowledge base. The service quality and efficiency of the intelligent customer service system can be improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to the technical fields of artificial intelligence, data processing, large models, intelligent customer service, etc. Background Art

[0002] In recent years, with the rapid development of the Internet, intelligent customer service systems have become an important tool for improving user service quality. However, in actual application, current intelligent customer service systems have problems such as redundant processing of similar questions, delayed knowledge base updates, and inaccurate content recommendations for users. These problems not only affect customer service efficiency, but also reduce user experience. Summary of the invention

[0003] The present disclosure provides an information processing method, apparatus, device and storage medium.

[0004] According to one aspect of the present disclosure, there is provided an information processing method, comprising:

[0005] Obtain multiple pieces of manual question and answer data, which are generated by an intelligent customer service system and a manual customer service system;

[0006] Analyze and cluster the manual question-answering data, and establish a knowledge module based on the analysis and clustering results;

[0007] Based on the knowledge module, the knowledge base of the intelligent customer service system is updated;

[0008] Based on the updated knowledge base, the recommendation strategy of the intelligent customer service system is reconfigured.

[0009] According to another aspect of the present disclosure, there is provided an information processing device, comprising:

[0010] An acquisition module is used to acquire multiple pieces of manual question and answer data, which are generated by an intelligent customer service system and a manual customer service system;

[0011] A processing module, used for analyzing and clustering the manual question and answer data, and establishing a knowledge module based on the analysis and clustering results;

[0012] A knowledge base updating module, used to update the knowledge base of the intelligent customer service system based on the knowledge module;

[0013] The reconfiguration module is used to reconfigure the recommendation strategy of the intelligent customer service system based on the updated knowledge base.

[0014] According to another aspect of the present disclosure, there is provided an electronic device, comprising:

[0015] at least one processor; and

[0016] a memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute any method in the embodiments of the present disclosure.

[0018] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute any method according to the embodiments of the present disclosure.

[0019] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which implements any method according to the embodiments of the present disclosure when executed by a processor.

[0020] The present invention analyzes and processes multiple manual question and answer data generated in the intelligent service process, establishes a knowledge module based on the processing results, and updates the knowledge base and recommendation mechanism of the intelligent customer service system based on the knowledge module. It can effectively transfer the experience and knowledge of manual customer service to the intelligent customer service system, and timely update the knowledge base and recommendation mechanism of the intelligent customer service system, thereby improving the service quality and efficiency of the intelligent customer service system.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0023] Figure 1 is a schematic diagram of an application scenario according to an embodiment of the present disclosure;

[0024] Figure 2 is a flowchart of an implementation of an information processing method according to an embodiment of the present disclosure;

[0025] Figure 3 is a flowchart of an information processing method according to an embodiment of the present disclosure;

[0026] Figure 4 is a schematic diagram of a recommendation pool according to an embodiment of the present disclosure;

[0027] Figure 5 is a schematic structural diagram of an information processing device 500 according to an embodiment of the present disclosure;

[0028] Figure 6A schematic block diagram of an example electronic device 600 that may be used to implement embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0029] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0030] The “and / or” in the embodiments of the present disclosure indicates that three relationships may exist. For example, A and / or B may indicate the following three situations: A exists alone, A and B exist at the same time, and B exists alone. The term “at least one” herein indicates any combination of at least two of any one or more of a plurality of types. For example, at least one of A, B, and C may indicate any one or more elements selected from the set consisting of A, B, and C. The terms “first” and “second” herein refer to and distinguish between multiple similar technical terms, and do not mean to limit the order or to limit the meaning to only two. For example, the first feature and the second feature refer to two types / two features. The first feature may be one or more, and the second feature may also be one or more.

[0031] In the digital age, customer service has become one of the key areas of corporate competition. The traditional manual customer service model faces challenges such as slow response, high cost, and difficulty in handling large-scale concurrent requests. With the rapid development of artificial intelligence technology, intelligent customer service systems have emerged. Among them, Retrieval-Augmented Generation (RAG) technology has gradually become a popular choice for building intelligent customer service systems due to its unique advantages.

[0032] RAG technology is an advanced method that combines information retrieval (IR) and natural language generation (NLG). The core idea is to not only rely on the model's own language understanding ability when generating responses, but also enhance the accuracy and richness of responses by retrieving external knowledge bases (such as FAQs (FAQs) (FAQs are documents or databases that collect and answer questions frequently asked by users), product manuals, etc.). The specific process includes:

[0033] 1. User input analysis: First, the system receives and analyzes the user's query or question to extract key information.

[0034] 2. Information retrieval: Use efficient search engines or retrieval models to find the documents or fragments most relevant to the user's question in the knowledge base.

[0035] 3. Generate responses: Use the retrieved information as input and combine it with natural language generation technology to generate responses that are both grammatical and rich in practical content.

[0036] An intelligent customer service system based on RAG technology usually includes the following key components:

[0037] User interaction interface: Provides a friendly user interface and supports multiple interaction methods such as text and voice.

[0038] Natural Language Processing (NLP) module: responsible for tasks such as parsing user input and identifying intent.

[0039] Information retrieval engine: efficiently retrieve relevant information from the knowledge base.

[0040] Response generation module: combines the retrieval results and the generation model to generate the final response.

[0041] Knowledge base management system: maintain and update knowledge base content to ensure the timeliness and accuracy of information.

[0042] In the prior art, the update of the knowledge base mainly relies on the regular maintenance method of manual operation. Although this method can maintain the accuracy and integrity of the knowledge base content to a certain extent, its inherent limitations cannot be ignored. Specifically, since it relies entirely on manual updating, it is difficult to achieve the goal of automatic and instant updating of the knowledge content in the knowledge base. This means that once new information, solutions or technical developments appear, the knowledge base cannot quickly incorporate these latest contents into it, resulting in a certain lag in the knowledge base content relative to the development of the real world, and cannot meet the user's demand for the latest information.

[0043] In order to solve the above problems, the present disclosure provides an information processing method. Figure 1 Schematic diagram of an application scenario according to an embodiment of the present disclosure. Figure 1As shown, the application scenario diagram of the embodiment of the present disclosure may include but is not limited to an intelligent customer service system 110, a manual customer service system 120, an information processing system 130 and a knowledge base 140. Both the intelligent customer service system 110 and the manual customer service system 120 can provide services for user questions. When the intelligent customer service system 110 receives a user question, it first uses NLP technology to parse and identify the user's input question; then, based on the results of the parsing and intent identification, it uses an information retrieval engine to search for relevant information from the knowledge base 140; then, a final reply is generated based on the search results and fed back to the user. The information processing method proposed in the embodiment of the present disclosure can be applied to the intelligent customer service system 110, and can realize the update of the knowledge base 140.

[0044] The present disclosure provides an information processing method. Figure 2 is a flowchart of an implementation of an information processing method according to an embodiment of the present disclosure, comprising:

[0045] S210, obtaining a plurality of manual question and answer data, where the manual question and answer data is generated by an intelligent customer service system and a manual customer service system;

[0046] S220, analyzing and clustering the manual question and answer data, and establishing a knowledge module based on the analysis and clustering results;

[0047] S230: Based on the knowledge module, update the knowledge base of the intelligent customer service system;

[0048] S240: reconfigure the recommendation strategy of the intelligent customer service system based on the updated knowledge base.

[0049] By collecting multiple manual question and answer data, processing the manual question and answer data using large language models (LLM), NLP and clustering technology, and updating the knowledge base of the intelligent customer service system based on the processing results, the intelligent customer service system can improve the accuracy of its answers and the overall service quality. Based on the updated knowledge base, the recommendation mechanism of the intelligent customer service system is updated to improve the accuracy and timeliness of the recommended content.

[0050] Figure 3 It is a flowchart of an information processing method according to an embodiment of the present disclosure.

[0051] like Figure 3 As shown, the information processing flow includes the following steps:

[0052] S301. A user asks a question to an intelligent customer service system, and the intelligent customer service system provides support and answers based on a preset knowledge base.

[0053] S302: The intelligent customer service system determines a solution to the user's problem data.

[0054] If the intelligent customer service system can answer the question data raised by the user, the information processing flow ends; if the intelligent customer service system cannot answer the user's question data, or the user requests to be transferred to manual customer service to answer his question data, enter S303.

[0055] S303: The intelligent customer service forwards the user's question data to the manual customer service, and the manual customer service answers the user's question data.

[0056] In some implementations, the manual question and answer data includes first question data and first answer data;

[0057] First, the question data is received by the intelligent customer service system and forwarded to the manual customer service system;

[0058] The first answer data is generated by the manual customer service system based on the first question data.

[0059] Through the above method, when a user requests to use the services of the intelligent customer service system and the intelligent customer service system cannot meet the user's needs, the user request can be transferred to the manual customer service system, and the reply information of the manual customer service can be further collected to update the knowledge base of the intelligent customer service system using the manual question and answer data, thereby effectively transferring the experience and knowledge of the manual customer service to the intelligent customer service system, avoiding the continuous repetition of old knowledge in the knowledge base and the inability to expand it in time.

[0060] During the manual customer service Q&A stage, based on the specific type of question data raised by the user, the system will assign this question data to manual customer service staff who have a high resolution rate for similar question data and can obtain high customer satisfaction. Based on the existing LLM, the system can pre-generate a preliminary answer to a question data. This answer has a certain degree of accuracy and completeness after being analyzed and reasoned by the model. Subsequently, this preliminary answer will be directly provided to the manual customer service staff, so that the manual customer service staff does not need to start from scratch to carry out tedious information search and sorting work. They can also directly paste this pre-generated answer to quickly respond to user inquiries, thereby greatly shortening the user's waiting time and improving service efficiency.

[0061] Based on a series of conversations between users and human customer service, the user's emotions are analyzed from the content of the conversation to deeply understand the user's emotional state. This includes a detailed assessment of the positive or negative tendency of the user's emotions, the satisfaction level, and potential needs or dissatisfaction. At the same time, through the user's emotions, the correctness of the human customer service's answers and the resolution rate of user problems can be evaluated. Based on the above indicators, each human customer service can be assigned a question type that better matches their expertise to ensure that they can play a role in the areas they are best at. Not only can it significantly improve the work efficiency and problem-solving ability of customer service, but it can also further optimize the customer experience.

[0062] S304: Analyze multiple pieces of manual question and answer data.

[0063] The key questions extracted from the question data of multiple manual question-answering data are extracted through LLM. For each key question, the entities and attributes in the question are extracted through named entity recognition (NER) technology. For example, the user's question is "When will the version of a certain operating system be updated?" In this question, the key entity is the name of the operating system. There is a subordinate relationship between the operating system and the update version time, so the attribute is the update version time.

[0064] S305. According to the analysis results, multiple manual question and answer data are integrated to generate multiple question and answer pairs.

[0065] In some implementations, analyzing and clustering the manual question and answer data based on the knowledge module includes:

[0066] Identify similarities between manual question and answer data and generate multiple question and answer pairs based on the identification results;

[0067] By clustering the question-answer pairs, the category corresponding to each question-answer pair is determined; the category includes a single-level or multi-level category.

[0068] Through the above method, questions and answers with similar or similar content in multiple manual question and answer data can be merged or classified to form multiple question and answer pairs. This avoids data redundancy and improves data utilization efficiency. In addition, the clustered question and answer pairs can more clearly reflect the correlation between questions and answers, and improve the accuracy of obtaining answers to user questions.

[0069] In some implementations, similarity recognition is performed on the artificial question and answer data, and multiple question and answer pairs are generated based on the recognition results, including:

[0070] Extracting at least two first key question data from the first question data;

[0071] Determining semantic similarity between at least two first key question data;

[0072] Acquire second question data according to at least two second key question data whose semantic similarity is greater than or equal to the first threshold, wherein the first key question includes the second key question;

[0073] Determine a plurality of second answer data corresponding to the second key question data;

[0074] Determine the third answer data corresponding to the second question data based on the second answer data and the satisfaction information of each second answer data;

[0075] The second question data and the third answer data are determined as a question-answer pair.

[0076] In the text clustering task, the semantic similarity algorithm is a core technology, which is mainly used to measure the closeness of two texts in meaning and cluster texts with similar meanings or similar types. For such problems, bidirectional encoder representations based on Transformers (BERT) are usually used to more accurately cluster texts based on deep semantic features, thereby improving the accuracy and robustness of clustering. For example, in the automatic classification of forum posts, BERT can help correctly classify posts discussing similar topics, even if the vocabulary and sentence structures used in these posts may be completely different.

[0077] In an embodiment of the present disclosure, a plurality of first key question data are extracted from the first question data through a large language model. The semantic similarity between the plurality of first key question data is determined using semantic similarity calculation technology, and the semantic similarity determined by the plurality of first key question data is compared with a preset first threshold value to determine at least two second key question data. The first threshold value represents a judgment criterion for distinguishing first key question data with sufficiently high similarity. When the semantic similarity of the plurality of first key question data is greater than or equal to the first threshold value, it means that the objectives of the question queries they represent are the same, and therefore these first key question data can be integrated together.

[0078] Through the above method, key issues with high semantic similarity are integrated, further simplifying the complexity of problem processing.

[0079] In some implementations, determining the third answer data corresponding to the second question data based on the second answer data and the satisfaction information of each second answer data includes:

[0080] determining the second answer data having the highest value of satisfaction information as the third answer data;

[0081] The satisfaction information is determined by LLM based on the analysis results of multiple manual question and answer data.

[0082] For example, one user's question data is "When will version 1.0 of software A be released?", and the answer data of the first manual customer service representative is "early November"; another user's question data is "When will version 1.0 of software A be released?", and the answer data of the second manual customer service representative is "November 1st". Both users asked about the release time of version 1.0 of software A, so these two question data can be integrated into one question. At the same time, the answer data of the second manual customer service representative is more detailed than the answer data of the first customer service representative, and the value of the satisfaction information will be higher. Therefore, the answer data of the second manual customer service representative can be determined as the third answer.

[0083] Through this method, the data of multiple answers to the second question can be combined to determine the best answer, which helps to improve the accuracy of the answer.

[0084] S306: Cluster the multiple question-answer pairs to determine the category corresponding to each question-answer pair.

[0085] Clustering is to divide a data set into different classes or clusters according to a specific standard (such as distance), so that the similarity of data objects in the same cluster is as large as possible, and the difference of data objects in different clusters is as large as possible. That is, after clustering, data of the same class is gathered together as much as possible, and data of different classes are separated as much as possible. When clustering multiple question-answer pairs, the following four methods can be used:

[0086] (1) Hierarchical clustering

[0087] When dealing with multiple question-answer pairs, we can use the hierarchical clustering method to perform hierarchical clustering analysis on the question-answer pairs, thereby revealing the multi-level structure hidden behind them. Hierarchical clustering is an unsupervised machine learning method that recursively merges or splits data to construct a tree-like clustering structure called a "dendrogram". This method is often used to explore the hierarchical structure of data. For a given data set object, a hierarchical clustering algorithm is used to obtain a process of combining data subsets with a hierarchical structure.

[0088] In the disclosed embodiment, hierarchical clustering is used to perform hierarchical analysis on a series of question-answer pairs. In this process, all question-answer pairs are first regarded as independent clusters, and then the most similar clusters are gradually merged according to certain criteria (such as minimum distance, etc.) until the stopping condition is met (such as reaching a preset number of clusters or the distance between clusters is small enough). Through this process, the logical relationship and hierarchical structure between question-answer pairs can be clearly revealed. For example, there are multiple question-answer pairs for a product, and a series of question-answer pairs about product features may first be clustered into a large subject category, such as "product function introduction"; then, within this subject category, the question-answer pairs may be further subdivided into more specific subcategories, such as "how to use", "troubleshooting", etc.

[0089] (2) Density-Based Spatial Clustering of Applications with Noise (DBSCAN)

[0090] DBSCAN does not need to pre-specify the number of clusters when clustering, and the final number of clusters is uncertain. DBSCAN marks all points as core points, boundary points, or noise points, deletes noise points, assigns an edge between all core points within the distance threshold, and each group of connected core points forms a cluster, assigning each boundary point to a cluster of core points associated with it. DBSCAN can cluster dense data sets of any shape, is insensitive to outliers in the data set, and the setting of the initial value will hardly affect the clustering results. If the density of the sample set is uneven and the cluster spacing is very different, the clustering quality is poor.

[0091] In the disclosed embodiment, the DBSCAN algorithm identifies dense regions in the data by analyzing the spatial distance between question-answer pairs and clusters similar objects in these regions. The algorithm is particularly suitable for processing question-answer pair datasets that are irregularly distributed, complex in shape, and contain noise.

[0092] (3) K-Means Clustering

[0093] K-Means clustering algorithm is one of the more classic clustering algorithms in the partitioning method. Due to its high efficiency, it is widely used in clustering large-scale data. K-Means clustering algorithm uses k as a parameter to divide n objects into k clusters, so that the similarity within the cluster is high, while the similarity between clusters is low.

[0094] In the disclosed embodiment, a K-Means clustering algorithm is used to perform cluster analysis on the semantic vectors of question-answer pairs, and question-answer pairs that are similar or have common features at the semantic level are classified into the same cluster.

[0095] (4) Topic Model

[0096] In order to more effectively mine the information value of multiple question-answer pairs, we can use topic models, such as Latent Dirichlet Allocation (LDA). Through topic models, we can deeply explore the hidden semantic connections between question-answer pairs, thereby identifying the core topics and categories they revolve around. This process not only helps to understand the content of question-answer pairs more comprehensively, but also provides a key information basis for subsequent applications such as text classification, information retrieval, and recommendation systems. Specifically, topic models can analyze the distribution of vocabulary in question-answer pairs, and reveal the potential topic structure behind question-answer pairs by counting word frequencies and calculating co-occurrence relationships. In this way, a large number of question-answer pairs can be organized and classified in an orderly manner, which reduces the load on the intelligent customer service system and improves the efficiency of knowledge base updates.

[0097] In the disclosed embodiment, the topic model can be used to identify the topics and categories involved in these question-answer pairs. For example, a user may ask questions about the usage, performance evaluation, after-sales service, etc. of a certain product. These questions can be classified into different topic categories such as "product usage", "product evaluation" and "after-sales service". In addition, the topic model will learn and identify the words, phrases or concepts that appear repeatedly in the question-answer pairs, thereby inferring the topic structure of the question-answer pairs.

[0098] S307. Establish a knowledge module based on the analysis and clustering results.

[0099] In some embodiments, the knowledge module includes at least one of the following:

[0100] Question and answer right;

[0101] The category corresponding to each question-answer pair;

[0102] The update time of each question and answer pair.

[0103] By adopting the above-mentioned knowledge module data storage method, the question and answer pairs in the knowledge module can be systematically and modularly organized and stored. In this way, when the knowledge module is used to update the knowledge base and recommendation strategy in the future, the efficiency and accuracy of the knowledge base update and recommendation can be improved.

[0104] S308. Based on the knowledge module, the knowledge base of the intelligent customer service system is updated.

[0105] In some implementations, the knowledge base of the intelligent customer service system is updated based on the knowledge module, including:

[0106] Inputting the second question data into the intelligent customer service system, and obtaining response information output by the intelligent customer service system;

[0107] Performing a first match between the response information and the third answer data corresponding to the second question data to obtain a first matching result;

[0108] Based on the first matching result, the knowledge base is updated.

[0109] In one example, updating the knowledge base based on the first matching result includes:

[0110] When the matching degree between the response information and the third answer data is less than or equal to the second threshold and the update time of the question-answer pair to which the third answer data belongs is later than the timestamp of the response information, the response information in the knowledge base is updated according to the third answer data.

[0111] The above method is to recall based on questions, that is, to ask the questions in the question-answer pair of the knowledge module to the intelligent customer service system to determine whether the intelligent customer service system can recall the answer information corresponding to the question. In this way, it is possible to determine not only whether the question-answer pair in the knowledge module has been saved in the knowledge base, but also whether the intelligent customer service system can effectively recall based on the question.

[0112] In the above-mentioned knowledge base update process, the second question data is input into the intelligent customer service system, and the intelligent customer service system feeds back the answer information. The answer information is matched with the third answer data corresponding to the second question data. Based on the first matching result, the knowledge base is updated. Specifically, this update mechanism is triggered when the following two conditions are met: first, the matching degree between the answer information and the third answer data is less than or equal to the preset second threshold, and the second threshold can be set according to the accuracy requirements and update frequency of the knowledge base. This condition ensures that the update will only be considered when there is a large difference or incomplete consistency in the content of the answer information and the third answer data.

[0113] Secondly, in addition to the matching requirements, the time condition must also be met, that is, the update time of the question and answer pair to which the third answer data belongs must be later than the timestamp of the response information. This condition ensures the timeliness of the update.

[0114] In some other implementations, updating the knowledge base of the intelligent customer service system based on the knowledge module further includes:

[0115] Performing a second matching of the third answer data with the knowledge base of the intelligent customer service system to obtain a second matching result;

[0116] Based on the second matching result, the knowledge base is updated.

[0117] In some implementations, updating the knowledge base based on the second matching result includes:

[0118] When the knowledge base of the intelligent customer service system contains first knowledge information matching the third answer data, the first knowledge is marked based on the second question data corresponding to the third answer data.

[0119] In some implementations, updating the knowledge base based on the second matching result includes:

[0120] When the knowledge base of the intelligent customer service system does not contain the first knowledge information matching the third answer data, the question-answer pair to which the third answer data belongs is saved in the knowledge base.

[0121] Different from the aforementioned question-based recall method, this method is based on answer recall, that is, the answer to the question and answer pair in the knowledge module is matched with the knowledge base of the intelligent customer service system to determine whether the answer has been saved in the knowledge base.

[0122] The above two recall methods can be used alone or in combination. Combining the two recall methods can more comprehensively recall the content in the knowledge base.

[0123] In the above-mentioned knowledge base update process, when the second question data cannot be used to feedback the answer information, the third answer data corresponding to the second question data is matched in the knowledge base of the intelligent customer service system. Based on the second matching result, the knowledge base is updated. If there is first knowledge information matching the third answer data in the knowledge base, this means that the question and answer pair has been recorded in the knowledge base. At this time, based on the second question data corresponding to the third answer data, the second question data and the first knowledge are paired and marked in the knowledge base so that the first knowledge can be used to answer the user's second question later. If there is no first knowledge information matching the third answer data in the knowledge base of the intelligent customer service system, the question and answer pair to which the third answer belongs is saved in the knowledge base.

[0124] By matching the first answer with the knowledge base and updating the knowledge base based on the matching results, it is possible to avoid the situation where it is impossible to determine whether the knowledge base needs to be updated when the response information feedback fails. At the same time, this process also enables the system to absorb new information, optimize existing question-answer pairs, and improve the accuracy and timeliness of the knowledge base.

[0125] S309: Update the recommendation pool.

[0126] In some implementations, reconfiguring the recommendation strategy of the intelligent customer service system based on the updated knowledge base includes:

[0127] Update the recommendation pool based on the updated knowledge base, where the recommendation pool includes the question-answer pairs corresponding to each user or user type, and the update time of the question-answer pairs in the knowledge base;

[0128] Reconfigure the recommendation weight of the question-answer pair based on the update time.

[0129] Based on the updated knowledge base, the system will update its internal recommendation pool accordingly. The recommendation pool is a highly sophisticated tool for indicating and matching specific question-answer pairs corresponding to different users or user types. More importantly, the recommendation pool also records in detail the specific update time of each question-answer pair in the knowledge base, which is crucial to ensure that users can receive updated and more accurate information.

[0130] Figure 4 is a schematic diagram of a recommendation pool according to an embodiment of the present disclosure. Figure 4 As shown in the figure, there are multiple question-answer pairs in the knowledge base, such as Q-A1, Q-A2…Q-An, and each question-answer pair has a corresponding update time. Assume that recommendation pool 1 recommends knowledge for category A users, and the recommended content is Q-A1, Q-A2, and Q-A3; recommendation pool 2 recommends knowledge for category B users, and the recommended content is Q-A4, Q-A5, and Q-An. For recommendation pool 1, the corresponding recommendation weights of the question-answer pairs are reconfigured according to the update time of Q-A1, Q-A2, and Q-A3. For example, if the update time of Q-A1 is the latest, the recommendation weight of this question-answer pair will be increased so that category A users can understand the knowledge updates in this field.

[0131] In the disclosed embodiment, recommendations based on the needs of user type groups are an effective strategy for deeply exploring and utilizing the common interests and needs of user type groups, aiming to provide users with content that is more in line with their expectations and preferences. This strategy accurately captures and understands the needs and preference characteristics of different user type groups by collecting and analyzing user behavior data, interest tags, historical access records and other information. At the same time, in order to maintain the timeliness and accuracy of the knowledge base, it is also necessary to increase the recommendation weight of the most recently updated knowledge content so as to force it to all users in the corresponding user type. This is because with the continuous development and updating of information, new knowledge, viewpoints and technologies continue to emerge, and users often have a strong interest in fresh and cutting-edge knowledge content.

[0132] Through collaborative filtering and content-based recommendation algorithms, users' historical behavior data and personality characteristics can be deeply mined and analyzed, thus providing users with accurate and personalized question-answer pair recommendation services. Collaborative filtering algorithms predict question-answer pairs that users may be interested in and make recommendations by mining similarities between users.

[0133] Content-based recommendation algorithms focus on analyzing the content features of question-answer pairs and users' preferences for these features. They build user interest models by extracting information such as text descriptions, keywords, and tags of question-answer pairs, as well as user feedback on these information. Then, by comparing the degree of match between the user interest model and the content features of the question-answer pairs, they recommend the question-answer pairs that best match the user's interests.

[0134] Through these methods, the system can prioritize the latest knowledge, and users can promptly receive recommended content that is more closely related to their needs and more timely.

[0135] In addition, through in-depth statistical analysis of manual question-and-answer data, we can comprehensively track and evaluate the efficiency of handling various user questions and user satisfaction. By carefully analyzing the interaction between manual customer service and users, we can reveal the inherent connection between the speed and accuracy of handling problems and user feedback. Based on the statistical analysis of manual question-and-answer data, we can comprehensively track and optimize the efficiency of handling various problems and user satisfaction, and then continuously optimize the knowledge base and recommendation mechanism to improve the overall service capabilities of the system.

[0136] The present disclosure also provides an information processing device, Figure 5 is a schematic diagram of the structure of an information processing device 500 according to an embodiment of the present disclosure, comprising:

[0137] An acquisition module 510 is used to acquire a plurality of manual question and answer data, where the manual question and answer data is generated by an intelligent customer service system and a manual customer service system;

[0138] Processing module 520, used to analyze and cluster the manual question and answer data, and establish a knowledge module based on the analysis and clustering results;

[0139] A knowledge base updating module 530, used to update the knowledge base of the intelligent customer service system based on the knowledge module;

[0140] The reconfiguration module 540 is used to reconfigure the recommendation strategy of the intelligent customer service system based on the updated knowledge base.

[0141] In some implementations, the manual question and answer data includes first question data and first answer data;

[0142] The first question data is received by the intelligent customer service system and forwarded to the manual customer service system;

[0143] The first answer data is generated by the manual customer service system based on the first question data.

[0144] In some embodiments, the processing module 520 is configured to:

[0145] Identify similarities between manual question and answer data and generate multiple question and answer pairs based on the identification results;

[0146] By clustering the question-answer pairs, the category corresponding to each question-answer pair is determined; the category includes a single-level or multi-level category.

[0147] In some embodiments, the processing module 520 is configured to:

[0148] Extracting at least two first key question data from the first question data;

[0149] Determine the semantic similarity between the first key question data;

[0150] Acquire second question data according to at least two second key question data whose semantic similarity is greater than or equal to a first threshold, wherein the first key question includes the second key question;

[0151] Determine a plurality of second answer data corresponding to the second key question data;

[0152] Determine the third answer data corresponding to the second question data based on the second answer data and the satisfaction information of each second answer data;

[0153] The second question data and the third answer data are determined as a question-answer pair.

[0154] In some embodiments, the processing module 520 is configured to:

[0155] determining the second answer data having the highest value of satisfaction information as the third answer data;

[0156] The satisfaction information is determined by a large language model based on the analysis results of multiple manual question and answer data.

[0157] In some embodiments, the knowledge module includes at least one of the following:

[0158] Question and answer right;

[0159] The category corresponding to each question-answer pair;

[0160] The update time of each question and answer pair.

[0161] In some implementations, the knowledge base updating module 530 is used to:

[0162] Inputting the second question data into the intelligent customer service system, and obtaining response information output by the intelligent customer service system;

[0163] Performing a first match between the response information and third answer data corresponding to the second question data to obtain a first matching result;

[0164] Based on the first matching result, the knowledge base is updated.

[0165] In some implementations, the knowledge base updating module 530 is used to:

[0166] When the matching degree between the response information and the third answer data is less than or equal to the second threshold and the update time of the question-answer pair to which the third answer data belongs is later than the timestamp of the response information, the response information in the knowledge base is updated according to the third answer data.

[0167] In some implementations, the knowledge base updating module 530 is further configured to:

[0168] Performing a second matching of the third answer data with the knowledge base of the intelligent customer service system to obtain a second matching result;

[0169] Based on the second matching result, the knowledge base is updated.

[0170] In some implementations, the knowledge base updating module 530 is used to:

[0171] When the knowledge base of the intelligent customer service system contains first knowledge information matching the third answer data, the first knowledge is marked based on the second question data corresponding to the third answer data.

[0172] In some implementations, the knowledge base updating module 530 is used to:

[0173] When the knowledge base of the intelligent customer service system does not contain the first knowledge information matching the third answer data, the question-answer pair to which the third answer data belongs is saved in the knowledge base.

[0174] In some implementations, the reconfiguration module 540 is configured to:

[0175] Update the recommendation pool based on the updated knowledge base, where the recommendation pool includes the question-answer pairs corresponding to each user or user type, and the update time of the question-answer pairs in the knowledge base;

[0176] According to the update time, the recommendation weight of the question-answer pair is reconfigured.

[0177] For the description of specific functions and examples of each module and submodule of the device in the embodiment of the present disclosure, reference can be made to the relevant description of the corresponding steps in the above method embodiment, which will not be repeated here.

[0178] In the technical solution disclosed in the present invention, the acquisition, storage and application of personal information of users involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0179] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0180] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0181] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0182] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0183] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as detection methods. For example, in some embodiments, the detection method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the detection method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the detection method in any other appropriate manner (e.g., by means of firmware).

[0184] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0185] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0186] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0187] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0188] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0189] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0190] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0191] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. An information processing method, comprising: Acquire multiple pieces of manual question and answer data, where the manual question and answer data are generated by an intelligent customer service system and a manual customer service system; Analyzing and clustering the manual question and answer data, and establishing a knowledge module based on the analysis and clustering results; Based on the knowledge module, updating the knowledge base of the intelligent customer service system; Based on the updated knowledge base, the recommendation strategy of the intelligent customer service system is reconfigured.

2. The method according to claim 1, wherein: The manual question and answer data includes first question data and first answer data; The first question data is received by the intelligent customer service system and forwarded to the manual customer service system; The first answer data is generated by the manual customer service system based on the first question data.

3. The method according to claim 1 or 2, wherein: The analyzing and clustering the manual question and answer data based on the knowledge module includes: Performing similarity recognition between the artificial question and answer data, and generating a plurality of question and answer pairs based on the recognition results; By clustering the question-answer pairs, the categories corresponding to the question-answer pairs are determined; the categories include single-level or multi-level categories.

4. The method according to claim 3, wherein: The performing similarity recognition on the artificial question and answer data and generating a plurality of question and answer pairs based on the recognition results includes: Extracting at least two first key question data from the first question data; Determining semantic similarity between at least two of the first key question data; Acquire second question data according to at least two second key question data whose semantic similarity is greater than or equal to a first threshold, wherein the first key question includes the second key question; Determine a plurality of second answer data corresponding to the second key question data; Determine third answer data corresponding to the second question data based on the second answer data and satisfaction information of each of the second answer data; The second question data and the third answer data are determined as the question-answer pair.

5. The method according to claim 4, wherein: The determining, based on the second answer data and the satisfaction information of each of the second answer data, the third answer data corresponding to the second question data comprises: determining the second answer data having the highest value of the satisfaction information as the third answer data; The satisfaction information is determined by a large language model based on analysis results of the plurality of manual question and answer data.

6. The method according to claim 5, wherein: The knowledge module contains at least one of the following: The question and answer pair; The categories corresponding to each of the question-answer pairs; The update time of each of the question-answer pairs.

7. The method according to claim 6, wherein: The updating of the knowledge base of the intelligent customer service system based on the knowledge module includes: Inputting the second question data into the intelligent customer service system to obtain response information output by the intelligent customer service system; Performing a first match between the response information and the third answer data corresponding to the second question data to obtain a first matching result; Based on the first matching result, the knowledge base is updated.

8. The method according to claim 7, wherein: The updating of the knowledge base based on the first matching result includes: When the matching degree between the response information and the third answer data is less than or equal to a second threshold and the update time of the question-answer pair to which the third answer data belongs is later than the timestamp of the response information, the response information in the knowledge base is updated according to the third answer data.

9. The method according to claim 7 or 8, wherein: The updating of the knowledge base of the intelligent customer service system based on the knowledge module further includes: Performing a second matching operation on the third answer data and the knowledge base of the intelligent customer service system to obtain a second matching result; Based on the second matching result, the knowledge base is updated.

10. The method according to claim 9, wherein: The updating of the knowledge base based on the second matching result includes: In a case where the knowledge base of the intelligent customer service system contains first knowledge information matching the third answer data, the first knowledge is marked based on the second question data corresponding to the third answer data.

11. The method according to claim 9, wherein: The updating of the knowledge base based on the second matching result includes: When the knowledge base of the intelligent customer service system does not contain the first knowledge information matching the third answer data, the question-answer pair to which the third answer data belongs is saved in the knowledge base.

12. The method according to claim 1, wherein: The reconfiguring the recommendation strategy of the intelligent customer service system based on the updated knowledge base includes: Update a recommendation pool based on the updated knowledge base, wherein the recommendation pool includes question-answer pairs corresponding to each user or user type, and update time of the question-answer pairs in the knowledge base; According to the update time, the recommendation weight of the question-answer pair is reconfigured.

13. An information processing device, comprising: An acquisition module, used to acquire multiple pieces of manual question and answer data, wherein the manual question and answer data are generated by an intelligent customer service system and a manual customer service system; A processing module, used for analyzing and clustering the manual question and answer data, and establishing a knowledge module based on the analysis and clustering results; A knowledge base updating module, used to update the knowledge base of the intelligent customer service system based on the knowledge module; The reconfiguration module is used to reconfigure the recommendation strategy of the intelligent customer service system based on the updated knowledge base.

14. The device according to claim 13, wherein: The manual question and answer data includes first question data and first answer data; The first question data is received by the intelligent customer service system and forwarded to the manual customer service system; The first answer data is generated by the manual customer service system based on the first question data.

15. The device according to claim 13 or 14, wherein: The processing module is used for: Performing similarity recognition between the artificial question and answer data, and generating a plurality of question and answer pairs based on the recognition results; By clustering the question-answer pairs, the categories corresponding to the question-answer pairs are determined; the categories include single-level or multi-level categories.

16. The device according to claim 15, wherein: The processing module is used for: Extracting at least two first key question data from the first question data; Determining semantic similarity between at least two of the first key question data; Acquire second question data according to at least two second key question data whose semantic similarity is greater than or equal to a first threshold, wherein the first key question includes the second key question; Determine a plurality of second answer data corresponding to the second key question data; Determine third answer data corresponding to the second question data based on the second answer data and satisfaction information of each of the second answer data; The second question data and the third answer data are determined as the question-answer pair.

17. The device according to claim 16, wherein: The processing module is used for: determining the second answer data having the highest value of the satisfaction information as the third answer data; The satisfaction information is determined by a large language model based on analysis results of the plurality of manual question and answer data.

18. The device according to claim 17, wherein: The knowledge module contains at least one of the following: The question and answer pair; The categories corresponding to each of the question-answer pairs; The update time of each of the question-answer pairs.

19. The device according to claim 18, wherein: The knowledge base updating module is used to: Inputting the second question data into the intelligent customer service system to obtain response information output by the intelligent customer service system; Performing a first match between the response information and the third answer data corresponding to the second question data to obtain a first matching result; Based on the first matching result, the knowledge base is updated.

20. The device according to claim 19, wherein The knowledge base updating module is used to: When the matching degree between the response information and the third answer data is less than or equal to a second threshold and the update time of the question-answer pair to which the third answer data belongs is later than the timestamp of the response information, the response information in the knowledge base is updated according to the third answer data.

21. The device according to claim 19 or 20, wherein: The knowledge base updating module is further used for: Performing a second matching operation on the third answer data and the knowledge base of the intelligent customer service system to obtain a second matching result; Based on the second matching result, the knowledge base is updated.

22. The device according to claim 21, wherein The knowledge base updating module is used to: In a case where the knowledge base of the intelligent customer service system contains first knowledge information matching the third answer data, the first knowledge is marked based on the second question data corresponding to the third answer data.

23. The device according to claim 21, wherein The knowledge base updating module is used to: When the knowledge base of the intelligent customer service system does not contain the first knowledge information matching the third answer data, the question-answer pair to which the third answer data belongs is saved in the knowledge base.

24. The device according to claim 13, wherein: The reconfiguration module is used to: Update a recommendation pool based on the updated knowledge base, wherein the recommendation pool includes question-answer pairs corresponding to each user or user type, and update time of the question-answer pairs in the knowledge base; According to the update time, the recommendation weight of the question-answer pair is reconfigured.

25. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 12.

26. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-12.

27. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 12.

Images