Business question and answer method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a business question and answer method and device, electronic equipment and a storage medium, and belongs to the technical field of artificial intelligence and finance. The method comprises the following steps: extracting voice emotion features and facial emotion features from historical service audio and video data; target work performance features are extracted from the historical service audio and video data; constructing a target customer group portrait based on historical customer data; screening out target business case data from a business database based on the target work performance characteristics and the target customer group portrait, so as to carry out model optimization on the original business question and answer model to obtain a target business question and answer model; answering the user question data through the target business question and answer model to obtain target answer content; performing voice synthesis based on the voice emotion features and the target reply content to obtain target reply voice; and synthesizing a target digital person for replying based on the target reply voice and the facial emotion features. According to the embodiment of the invention, the business question-answering efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the fields of artificial intelligence and financial technology, and in particular to a business question-answering method and device, an electronic device, and a storage medium. Background Art

[0002] Business Q&A is a service that answers business-related questions, aiming to provide customers with accurate, timely information and professional business advice. Business Q&A can be used in a variety of scenarios, such as in the financial sector, to answer customers' questions about financial services.

[0003] Currently, business Q&A is mainly achieved through manual responses from specific business agents. However, in actual application scenarios, business agents may be unable to provide timely answer services to customers due to busy business, thus affecting the efficiency of business Q&A.

[0004] Therefore, how to improve the efficiency of business question answering has become a technical problem that needs to be solved urgently. Summary of the Invention

[0005] The main purpose of the embodiments of the present application is to provide a business question-and-answer method and device, an electronic device, and a storage medium, aiming to improve the efficiency of business question-and-answer.

[0006] To achieve the above objectives, a first aspect of an embodiment of the present application provides a business question-answering method, the method comprising:

[0007] Obtaining historical business data of the target object; wherein the historical business data includes historical customer data and historical business audio and video data;

[0008] Extracting emotional features from the historical business audio and video data to obtain voice emotional features and facial emotional features;

[0009] Extracting features from the historical business audio and video data to obtain target work performance features of the target object;

[0010] Constructing a user profile for the historical customer data to obtain a target customer profile;

[0011] Filtering target business case data from a preset business database based on the target work performance characteristics and the target customer group portrait;

[0012] Optimizing the preset original business question-answering model based on the target business case data to obtain a target business question-answering model;

[0013] The target business question-answering model is used to answer the pre-acquired user question data to obtain the target answer content; wherein the user question data is raised by the target questioner;

[0014] Perform speech synthesis based on the speech emotion feature and the target answer content to obtain the target answer speech;

[0015] A digital human is synthesized based on the target answering voice and the facial emotion features to obtain a target digital human, and the target digital human is used to answer the target questioning object.

[0016] In some embodiments, the business database includes a plurality of business expert data, each of which includes expert work performance characteristics, business customer group profiles, and expert business cases; and filtering target business case data from a preset business database based on the target work performance characteristics and the target customer group profiles includes:

[0017] For each of the business expert data, similarity calculation is performed on the target work performance characteristics and the expert work performance characteristics to obtain work performance similarity data, and similarity calculation is performed on the target customer group portrait and the business customer group portrait to obtain customer group similarity data;

[0018] Performing an aggregation operation on the work performance similarity data and the customer group similarity data to obtain original business similarity data;

[0019] Filtering the plurality of original business similarity data based on a preset business similarity threshold to obtain target business similarity data;

[0020] determining target expert data from the business database based on the target business similarity data;

[0021] The expert business case of the target expert data is acquired to obtain the target business case data.

[0022] In some embodiments, optimizing a preset original business question-and-answer model based on the target business case data to obtain a target business question-and-answer model includes:

[0023] Perform case feature analysis on the target business case data to obtain target case parameter information;

[0024] The parameters of the original business question-and-answer model are adjusted based on the target case parameter information to obtain the target business question-and-answer model.

[0025] In some embodiments, the historical customer data includes multiple pieces of historical customer information, and constructing a user profile for the historical customer data to obtain a target customer group profile includes:

[0026] Classifying the historical customer information by customer type to obtain historical customer type data;

[0027] Performing clustering processing on the historical customer type data to obtain original customer group clustering data; wherein the original customer group clustering data includes a plurality of clustering points, each of the clustering points corresponding to one piece of historical customer information;

[0028] The customer type to which the largest number of cluster points belong is selected from the original customer group clustering data to obtain the target customer group portrait.

[0029] In some embodiments, the historical service audio and video data includes historical service voice data and historical service video data; and extracting emotional features from the historical service audio and video data to obtain voice emotional features and facial emotional features includes:

[0030] Performing audio extraction on the historical service video data to obtain original video and voice data;

[0031] Performing voice emotion extraction on the original video voice data and the historical service voice data to obtain the voice emotion features;

[0032] Extracting video frames from the historical service video data to obtain original video frame images;

[0033] Performing image preprocessing on the original video frame image to obtain an original face video image;

[0034] Perform expression extraction on the original face video image to obtain the facial emotion features.

[0035] In some embodiments, extracting features from the historical business audio and video data to obtain target work performance features of the target object includes:

[0036] Extracting speech features from the historical service speech data to obtain target speech performance features;

[0037] Extracting image features from the historical business video data to obtain target image performance features;

[0038] The target work performance feature of the target object is obtained by combining the target voice performance feature and the target image performance feature.

[0039] In some embodiments, performing speech synthesis based on the speech emotion feature and the target response content to obtain the target response speech includes:

[0040] Performing text preprocessing on the target reply content to obtain target reply text data;

[0041] Performing phoneme conversion on the target reply text data to obtain original phoneme data;

[0042] Performing emotion adjustment on the original phoneme data based on the speech emotion feature to obtain target phoneme data;

[0043] Speech synthesis is performed based on the target phoneme data to obtain the target response speech.

[0044] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides a business question-and-answer device, comprising:

[0045] A historical business data acquisition module is used to acquire the historical business data of the target object; wherein the historical business data includes historical customer data and historical business audio and video data;

[0046] An emotion feature extraction module is used to extract emotion features from the historical business audio and video data to obtain speech emotion features and facial emotion features;

[0047] A work performance feature extraction module, configured to extract features from the historical business audio and video data to obtain target work performance features of the target object;

[0048] A customer group portrait construction module is used to construct a user portrait based on the historical customer data to obtain a target customer group portrait;

[0049] A business case screening module is used to screen target business case data from a preset business database based on the target work performance characteristics and the target customer group portrait;

[0050] A model optimization module is used to optimize the preset original business question-answering model based on the target business case data to obtain a target business question-answering model;

[0051] A question answering module is used to answer pre-acquired user question data using the target business question answering model to obtain target answer content; wherein the user question data is raised by the target questioning object;

[0052] A speech synthesis module, configured to perform speech synthesis based on the speech emotion feature and the target response content to obtain a target response speech;

[0053] The digital human synthesis module is used to synthesize a digital human based on the target answer voice and the facial emotion features to obtain a target digital human, and use the target digital human to answer the target questioner.

[0054] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0055] To achieve the above-mentioned purpose, the fourth aspect of the embodiments of the present application proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the method described in the first aspect.

[0056] The business question-and-answer method and device, electronic device, and storage medium proposed in the present application obtain historical business data of a target object, wherein the historical business data includes historical customer data and historical business audio and video data; perform emotional feature extraction on the historical business audio and video data to obtain voice emotional features and facial emotional features; perform feature extraction on the historical business audio and video data to obtain target work performance features of the target object; construct a user profile on the historical customer data to obtain a target customer group portrait; thereby, based on the target work performance features and the target customer group portrait, target business case data is screened out from a preset business database to optimize the original business question-and-answer model to obtain a target business question-and-answer model, so that the target business question-and-answer model can be used to answer user question data raised by the target questioning object to obtain target answer content, thereby improving the accuracy of the business answer; further, speech synthesis is performed based on the voice emotional features and the target answer content to obtain a target answer speech, digital human synthesis is performed based on the target answer speech and facial emotional features to obtain a target digital human, and the target digital human is used to answer the target questioning object, thereby enhancing the interactivity and real-time nature of the business question-and-answer, and effectively improving the efficiency of the business question-and-answer and user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a flowchart of the business question-and-answer method provided in an embodiment of the present application;

[0058] Figure 2 yes Figure 1 Flowchart of step S102 in FIG.

[0059] Figure 3 yes Figure 1 Flowchart of step S103 in FIG.

[0060] Figure 4 yes Figure 1 Flowchart of step S104 in FIG.

[0061] Figure 5 yes Figure 1 Flowchart of step S105 in FIG.

[0062] Figure 6 yes Figure 1 Flowchart of step S106 in FIG.

[0063] Figure 7 yes Figure 1 Flowchart of step S108 in FIG.

[0064] Figure 8 This is a schematic diagram of the structure of the business question-and-answer device provided in an embodiment of the present application;

[0065] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0067] It should be noted that although the device schematics illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps shown or described may be performed in a sequence that differs from the module divisions in the device or the sequence in the flowcharts. The terms "first," "second," and so on, in the specification, claims, and drawings, are used to distinguish similar items and are not necessarily used to describe a specific sequence or precedence.

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0069] First, let’s analyze some of the terms used in this application:

[0070] Artificial intelligence (AI) is a new technical discipline that studies and develops theories, methods, technologies, and application systems for simulating, extending, and expanding human intelligence. A branch of computer science, AI seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thinking. It also encompasses the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0071] Natural language processing (NLP): NLP uses computers to process, understand, and apply human languages ​​(such as Chinese and English). A branch of artificial intelligence, NLP is an interdisciplinary field between computer science and linguistics, often referred to as computational linguistics. Natural language processing encompasses grammatical analysis, semantic analysis, and discourse comprehension. Natural language processing is commonly used in technical fields such as machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information intent recognition, information extraction and filtering, text classification and clustering, public opinion analysis, and opinion mining. It encompasses data mining, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research related to language processing, and linguistics research related to language computing.

[0072] Text-to-speech (TTS) is a technology that converts text information stored, edited, or transmitted in a computer into natural and fluent speech output. Speech synthesis uses advanced phonetics, linguistics, and acoustic technologies to analyze and process the input text data, then generates the corresponding speech waveform, and finally outputs it as audible sound through playback devices such as speakers. During the speech synthesis process, the speech synthesis system selects appropriate parameters such as voice, intonation, and rhythm based on the semantics, grammar, and context of the text to simulate the natural way of human speech as much as possible, making the generated speech not only clear and accurate, but also expressive and appealing. Speech synthesis technology has a wide range of applications in many fields, such as intelligent customer service, audio reading, assisted education, in-car navigation, etc., greatly enriching people's interactive experience and information acquisition methods.

[0073] Business Q&A is a service that answers business-related questions. It's typically provided by internal business experts, consultants, or agents. It aims to leverage expertise and experience to gain a deep understanding of a company's business environment, goals, and current status, providing clients with accurate, timely information and professional business advice. Business Q&A can be applied in a variety of scenarios, such as in the financial sector, answering clients' questions about financial services.

[0074] Currently, business Q&A is mainly achieved through manual responses from specific business agents. However, in actual application scenarios, business agents may be unable to provide timely answer services to customers due to busy business, thus affecting the efficiency of business Q&A.

[0075] Based on this, the embodiments of the present application provide a business question and answer method and device, an electronic device, and a storage medium, aiming to improve the efficiency of business question and answer.

[0076] The business question-and-answer method and device, electronic device, and storage medium provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, the business question-and-answer method in the embodiments of the present application is described.

[0077] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0078] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0079] The business question-and-answer method provided in the embodiment of the present application relates to the field of artificial intelligence technology. The business question-and-answer method provided in the embodiment of the present application can be applied to a terminal, can be applied to a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or as a server cluster or distributed system composed of multiple physical servers, or as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the business question-and-answer method, etc., but is not limited to the above forms.

[0080] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0081] It should be noted that in each specific embodiment of the present application, when it comes to the need to perform relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of such data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0082] Figure 1 This is an optional flowchart of the business question-and-answer method provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S109.

[0083] Step S101, obtaining historical business data of the target object; wherein the historical business data includes historical customer data, historical business audio and video data;

[0084] Step S102: extracting emotional features from historical service audio and video data to obtain voice emotional features and facial emotional features;

[0085] Step S103, extracting features from the historical business audio and video data to obtain target work performance features of the target object;

[0086] Step S104: construct a user profile based on historical customer data to obtain a target customer profile;

[0087] Step S105 , filtering target business case data from a preset business database based on target work performance characteristics and target customer profile;

[0088] Step S106: Optimizing the preset original business question-answering model based on the target business case data to obtain a target business question-answering model;

[0089] Step S107: answering the pre-acquired user question data through the target business question-answering model to obtain target answer content; wherein the user question data is raised by the target questioner;

[0090] Step S108, performing speech synthesis based on the speech emotion feature and the target answer content to obtain the target answer speech;

[0091] Step S109: synthesize a digital human based on the target answering voice and facial emotion features to obtain a target digital human, and use the target digital human to answer the target questioner.

[0092] Steps S101 to S109 shown in the embodiment of the present application are performed by obtaining historical business data of the target object; wherein the historical business data includes historical customer data and historical business audio and video data; performing emotional feature extraction on the historical business audio and video data to obtain voice emotional features and facial emotional features; performing feature extraction on the historical business audio and video data to obtain target work performance features of the target object; constructing a user profile on the historical customer data to obtain a target customer group portrait; thereby, based on the target work performance features and the target customer group portrait, target business case data is screened out from a preset business database to optimize the original business question and answer model to obtain a target business question and answer model, so that the target business question and answer model can be used to answer the user question data raised by the target questioning object to obtain the target answer content, thereby improving the accuracy of the business answer; further, based on the voice emotional features and the target answer content, speech synthesis is performed to obtain the target answer speech, based on the target answer speech and facial emotional features, digital human synthesis is performed to obtain the target digital human, and the target digital human is used to answer the target questioning object, thereby enhancing the interactivity and real-time nature of the business question and answer, and effectively improving the efficiency of the business question and answer and user satisfaction.

[0093] In step S101 of some embodiments, the target object may be a business agent, a business consultant, a manual customer service representative, etc., who can connect with customers and answer their questions.

[0094] Among them, historical business data includes the archived historical customer data and historical business audio and video data of the target object; specifically, historical customer data includes multiple historical customer information, and historical customer information is relevant information of users that the target object has connected with; historical business audio and video data includes historical business voice data and historical business video data, and historical business voice data can be telephone communication recording data between the target object and the customer, and historical business video data can be communication video data between the target object and the customer, etc., but is not limited to this.

[0095] See also Figure 2 In some embodiments, step S102 may include but is not limited to steps S201 to S205:

[0096] Step S201, extracting audio from historical service video data to obtain original video and voice data;

[0097] Step S202: extracting speech emotion from the original video speech data and the historical service speech data to obtain speech emotion features;

[0098] Step S203, extracting video frames from historical service video data to obtain original video frame images;

[0099] Step S204, performing image preprocessing on the original video frame image to obtain an original face video image;

[0100] Step S205: extract facial expressions from the original facial video image to obtain facial emotion features.

[0101] Steps S201 to S205 shown in the embodiment of the present application perform audio extraction on historical business video data to obtain original video voice data, perform voice emotion extraction on the original video voice data and historical business voice data to obtain voice emotion features of the target object; perform video frame extraction on historical business video data to obtain original video frame images, perform image preprocessing on the original video frame images to obtain original facial video images, perform expression extraction on the original facial video images to obtain facial emotion features of the target object, which can more comprehensively understand and analyze the emotional elements in business exchanges, facilitate the subsequent synthesis of a digital person similar to the target object, thereby enhancing the interactivity and real-time nature of business questions and answers.

[0102] In step S201 of some embodiments, audio extraction can be performed on historical business video data using preset audio software or video editing tools to obtain original video and voice data, wherein the data format of the original video and voice data needs to be selected according to the data requirements of a preset voice emotion analysis model, for example, FLAC format or WAV format can be selected, but is not limited to this.

[0103] In addition, the data format of the historical business voice data also needs to be adjusted according to the data requirements of the preset voice emotion analysis model, that is, the data format of the original video voice data is the same as the data format of the historical business voice data.

[0104] In some embodiments, step S202 may include but is not limited to the following steps:

[0105] Speaker separation is performed on the original video and voice data using a preset speaker separation algorithm to obtain target video and voice data of the target object;

[0106] Perform speaker separation on the historical service voice data using a preset speaker separation algorithm to obtain the target service voice data of the target object;

[0107] The target video voice data and target business voice data of the target object are subjected to voice emotion extraction through a preset voice emotion analysis model to obtain the voice emotion features of the target object.

[0108] Among them, the speaker separation algorithm can identify and separate different speakers in a speech based on sound features (such as pitch, speaking speed, voiceprint, etc.). Specifically, the speaker separation algorithm can be a U-Net model, a TasNet model, etc., and the specific selection needs to be based on the actual application scenario, but is not limited to this.

[0109] The preset speech emotion analysis model can extract the speaker's speech emotion characteristics from speech data. Specifically, it can be a hybrid model based on deep learning (such as convolutional neural network model (CNN) and long short-term memory network model (LSTM)), BERT model, ResNet model, etc. The specific selection needs to be based on the actual application scenario, but is not limited to this.

[0110] In step S203 of some embodiments, video frame extraction can be performed on historical business video data through a preset video frame extraction model, video processing tools, etc. to obtain the original video frame image, wherein the video frame extraction model can be a deep learning model, a machine learning model, etc., and the video processing tool can be an OpenCV library, an FFmpeg command line tool, etc. The specific selection needs to be made in combination with the actual application scenario, but is not limited to this.

[0111] In some embodiments, step S204 may include but is not limited to the following steps:

[0112] Perform face recognition on the original video frame image to obtain the original face image of the target object;

[0113] The size and resolution of the original face image are adjusted, and the initial video frame image is denoised, standardized, and enhanced to obtain the original face video image of the target object.

[0114] Among them, the specific preprocessing operation of the original facial image needs to be selected in combination with the actual application scenario, and the range of size and resolution adjustment needs to be set according to the data requirements of the preset facial emotion analysis model, but is not limited to this.

[0115] In step S205 of some embodiments, facial expression extraction is performed on the original facial video image using a preset facial emotion analysis model to obtain facial emotion features of the target object.

[0116] The preset facial emotion analysis model can extract facial emotion features of a person from image data. Specifically, it can be a convolutional neural network model (CNN), a recurrent neural network model (RNN), an emotion-enhanced multimodal head speaking video generation model (EMOPortraits), a FaceNet model, etc. The specific selection needs to be based on the actual application scenario, but is not limited to this.

[0117] See also Figure 3 In some embodiments, step S103 may include but is not limited to steps S301 to S303:

[0118] Step S301, extracting speech features from historical service speech data to obtain target speech performance features;

[0119] Step S302, extracting image features from historical business video data to obtain target image performance features;

[0120] Step S303 : obtaining target work performance features of the target object based on the combination of the target voice performance features and the target image performance features.

[0121] In steps S301 to S303 shown in the embodiment of the present application, target voice performance features are obtained by extracting voice features from historical business voice data, and target image performance features are obtained by extracting image features from historical business video data. Then, the target work performance features of the target object are obtained by combining the target voice performance features and the target image performance features. This can more comprehensively and accurately evaluate the actual work performance of the target object, enhance the accuracy of the evaluation, and provide a basis for subsequent matching of similar business experts based on the target work performance features of the target object.

[0122] It should be noted that the target job performance characteristics are used to evaluate the target object's work ability. For example:

[0123] Target object A, after analysis, the target work performance characteristics of target object A are: elite image, sunny, positive, good at communication, with excellent business capabilities, and able to quickly guide customers to express product needs.

[0124] Target object B. After analysis, the target work performance characteristics of target object B are: friendly image, reasonable, high understanding of financial products, able to answer users' questions on the spot, and high customer satisfaction.

[0125] In some embodiments, step S301 may include but is not limited to the following steps:

[0126] Convert historical business voice data into text using a preset speech recognition algorithm to obtain historical business text data;

[0127] The historical business text data is extracted through the preset semantic feature extraction model to obtain the target speech performance features.

[0128] In some embodiments, step S301 may also include but is not limited to the following steps:

[0129] Convert historical business voice data into text using a preset speech recognition algorithm to obtain historical business text data;

[0130] The original video voice data is converted into text using a preset speech recognition algorithm to obtain business video text data;

[0131] The preset semantic feature extraction model is used to extract features from historical business text data and business video text data to obtain the target speech performance features.

[0132] Specifically, the preset speech recognition algorithm can be a Wav2Vec algorithm, a DeepSpeech algorithm, a CNN model, an RNN model, etc., and the specific selection needs to be based on the actual application scenario, but is not limited to this.

[0133] The preset semantic feature extraction model can be a BERT model, a ResNet model, etc., and the specific selection needs to be based on the actual application scenario, but is not limited to this.

[0134] In some embodiments, before feature extraction is performed on text data through a preset semantic feature extraction model, the text data may be further cleaned to improve the accuracy of speech feature extraction; specifically, text cleaning includes but is not limited to removing punctuation marks, special symbols, stop words, etc., and the specific selection needs to be made in combination with the actual application scenario, but is not limited to this.

[0135] In step S302 of some embodiments, image features may be extracted from the original face video image using a preset image feature extraction model to obtain target image performance features;

[0136] Specifically, the preset image feature extraction model can be a pre-trained neural network model, a DeepFace model, a FaceChain-FACT model, a Holistic method model based on global information (such as principal component analysis PCA, linear discriminant analysis LDA, etc.) and a Local method model based on local information (such as HOG, LBP, SIFT, SURF, etc.), etc. The specific selection needs to be combined with the actual application scenario, but is not limited to this.

[0137] In step S303 of some embodiments, the target voice performance features and the target image performance features are combined to obtain target work performance features that can characterize the work ability of the target object, which is conducive to subsequent matching with more similar business experts based on the target work performance features.

[0138] See also Figure 4 In some embodiments, step S104 may include but is not limited to steps S401 to S403:

[0139] Step S401, classify historical customer information by customer type to obtain historical customer type data;

[0140] Step S402: clustering the historical customer type data to obtain original customer group clustering data; wherein the original customer group clustering data includes a plurality of clustering points, each clustering point corresponding to a piece of historical customer information;

[0141] Step S403: Filter out the customer type to which the largest number of cluster points belong from the original customer group clustering data to obtain a target customer group portrait.

[0142] Steps S401 to S403 shown in the embodiment of the present application, by classifying and clustering historical customer information by customer type, and then screening out the customer type to which the largest number of cluster points belong, can accurately identify and portray the main customer group of the target customer, that is, the target customer group portrait, effectively improving the accuracy and pertinence of customer segmentation, and providing data support for subsequent business expert matching.

[0143] In step S401 of some embodiments, first, a plurality of preset customer types are set based on the application scenario, and each historical customer information is classified into a customer type using a pre-trained customer type classification model, thereby obtaining a plurality of historical customer type data.

[0144] Specifically, the customer type classification model can be a neural network model, a deep learning model, an SVM model, a decision tree model, a random forest model, etc. The specific selection needs to be based on the actual application scenario, but is not limited to this.

[0145] In step S402 of some embodiments, multiple historical customer type data can be clustered, such as K-means clustering, X-means clustering, etc., to obtain at least one original customer group clustering data. Each original customer group clustering data can be specifically understood as a customer group. Each original customer group clustering data includes multiple clustering points, and each clustering point corresponds to a customer, that is, corresponds to a historical customer information.

[0146] It is understandable that the customers that a target object connects with may be of various types, that is, there is at least one customer group, and the customer group with the largest number of people (that is, the customer group with the largest number of clustering points) is the target customer group.

[0147] Based on this, the customer group with the largest number of clustering points is selected from at least one original customer group clustering data, and the corresponding customer type is the target customer group portrait.

[0148] In step S105 of some embodiments, the business database is constructed based on a specific application scenario, and the business database includes data of multiple business experts, each of which includes expert work performance characteristics, business customer group portraits, and expert business cases;

[0149] Specifically, the business database can be understood as an intelligent entity that integrates different types of professional business experts, and the business expert data can be understood as the business-related data of excellent business agents, including expert work performance characteristics, business customer group portraits and expert business cases. Among them, expert work performance characteristics and business customer group portraits are pre-predicted and calibrated, while expert business cases are business data that are combined with specific scenarios and have universal commonalities. They can be used to optimize the preset original business question-and-answer model, thereby generating a target business question-and-answer model with the work ability of an excellent business agent, and then using the target business question-and-answer model to answer questions raised by customers, thereby improving the accuracy of business questions and answers.

[0150] See also Figure 5 In some embodiments, step S105 may also include but is not limited to steps S501 to S505:

[0151] Step S501: For each business expert data, similarity calculation is performed between the target work performance characteristics and the expert work performance characteristics to obtain work performance similarity data, and similarity calculation is performed between the target customer group portrait and the business customer group portrait to obtain customer group similarity data;

[0152] Step S502: performing aggregation operation on the work performance similarity data and the customer group similarity data to obtain original business similarity data;

[0153] Step S503, screening multiple original business similarity data based on a preset business similarity threshold to obtain target business similarity data;

[0154] Step S504, determining target expert data from the business database based on the target business similarity data;

[0155] Step S505: Acquire expert business cases of target expert data to obtain target business case data.

[0156] In steps S501 to S505 shown in the embodiment of the present application, the work performance similarity data and customer group similarity data between the target object and the business expert data are calculated for each business expert data, and the work performance similarity data and customer group similarity data are aggregated to obtain the original business similarity data. Finally, based on the target business similarity data, the target expert data is determined from the business database, which can effectively identify the most similar business experts, obtain the expert business cases of the target expert data, and obtain the target business case data, thereby providing strong support for business decision-making and problem solving.

[0157] In step S501 of some embodiments, similarity calculation may adopt cosine similarity calculation method, Pearson similarity calculation method, Euclidean distance calculation method, Manhattan distance calculation method, etc., and the specific selection needs to be made in combination with the actual application scenario, and is not limited thereto.

[0158] In addition, the similarity data has a value between 0 and 1, wherein the closer the value is to 1, the higher the similarity.

[0159] In step S502 of some embodiments, the work performance similarity data and the customer group similarity data can be added to obtain the original business similarity data; or the work performance similarity data and the customer group similarity data can be weighted and summed based on different weights to obtain the original business similarity data. The specific selection needs to be made in combination with the actual application scenario, and is not limited to this.

[0160] It should be noted that the preset business similarity threshold is set based on a specific application scenario, for example, the business similarity threshold is 0.9, the business similarity threshold is 0.95, and the present invention is not limited thereto.

[0161] In some embodiments, target expert data is determined from a business database based on target business similarity data, and the number of target expert data is one or more, so as to obtain expert business cases corresponding to the target expert data, obtain target business case data, and use it as an excellent case to optimize the original business question-and-answer model.

[0162] If the multiple original business similarity data do not exceed the preset business similarity threshold, the data with the largest similarity value among the original business similarity data is selected as the target business similarity data.

[0163] In step S106 of some embodiments, the preset original business question-answering model may be a large language model, a neural network model pre-trained based on an application scenario, or the like.

[0164] See also Figure 6 In some embodiments, step S106 includes but is not limited to steps S601 to S602:

[0165] Step S601: Analyze the case characteristics of the target business case data to obtain target case parameter information;

[0166] Step S602 : Adjust the parameters of the original business question-and-answer model based on the target case parameter information to obtain a target business question-and-answer model.

[0167] In steps S601 to S602 shown in the embodiment of the present application, target case parameter information is obtained by performing case feature analysis on the target business case data, and the parameters of the original business question and answer model are adjusted according to the target case parameter information, so that the business question and answer model can more accurately understand and answer relevant business questions, thereby improving the accuracy and efficiency of the model in actual application scenarios.

[0168] Since the target business case data may be audio and video data or text data, the target business case data first needs to be uniformly converted into text data.

[0169] In some embodiments, step S601 may include but is not limited to the following steps:

[0170] Performing text conversion on target business case data to obtain target case text data;

[0171] Perform feature analysis on the target case text data to obtain target case features; target case features include user preference features, communication strategies, communication skills, etc.

[0172] Parameterize the target case features to obtain target case parameter information.

[0173] Specifically, feature analysis of target case text data can be performed using a pre-trained case learning model.

[0174] In step S602 of some embodiments, the parameters of the original business question and answer model may be adjusted based on a preset optimization strategy and by calling target case parameter information to obtain a target business question and answer model.

[0175] Specifically, the preset optimization strategy can be the proximal policy optimization method (PPO), the trust region policy optimization method (TRPO), the deep deterministic policy gradient method (DDPG), etc., or it can be the gradient descent method, the back propagation method and other methods. The specific selection needs to be combined with the actual application scenario, but is not limited to this.

[0176] In step S106 of some embodiments, if the original business question-and-answer model is a large language model, the target business case data can be input into the large language model so that the large language model can automatically learn the communication strategies, closing skills, etc. of the business experts in the target business case data, thereby possessing the business capabilities of the business experts in the target business case data.

[0177] In step S107 of some embodiments, the target questioning object may be a customer with whom the target object has interacted, or a new customer of the target object. When the target object is unable to respond to the target questioning object's business question in a timely manner, the optimized target business question-answering model may be used to answer the question.

[0178] Therefore, after obtaining the user question data raised by the target questioner, the target business question-answering model is used to answer the question and obtain accurate target answer content.

[0179] See also Figure 7 In some embodiments, step S108 may include but is not limited to steps S701 to S704:

[0180] Step S701, performing text preprocessing on the target reply content to obtain target reply text data;

[0181] Step S702, performing phoneme conversion on the target reply text data to obtain original phoneme data;

[0182] Step S703, emotion adjustment is performed on the original phoneme data based on the speech emotion feature to obtain target phoneme data;

[0183] Step S704: Perform speech synthesis based on the target phoneme data to obtain the target response speech.

[0184] In steps S701 to S704, the target response content is subjected to text preprocessing and phoneme conversion to obtain raw phoneme data for adaptation to the speech synthesis system. On this basis, the raw phoneme data is emotionally adjusted using speech emotion characteristics to obtain target phoneme data, which can make the synthesized speech more consistent with the target subject's speech emotion expression and enhance its expressiveness. Ultimately, speech synthesis based on the target phoneme data produces a target response speech that not only accurately conveys the information but also is rich in the target subject's emotional color, improving the naturalness of the speech synthesis and increasing the intimacy of the target questioner.

[0185] In step S701 of some embodiments, regular expressions or NLP libraries can be used to implement text preprocessing of the target reply content, specifically including but not limited to removing punctuation marks, special symbols, stop words, etc.

[0186] In addition, the target reply content after text cleaning can also be segmented through a word segmentation tool (such as Jieba tool) to divide the target reply content into a sequence of multiple words or units to obtain target reply text data.

[0187] In step S702 of some embodiments, a phoneme conversion tool, such as a Seq2Seq model, a convolutional neural network (CNN) model, etc., can be used to convert each word in the target response text data into a corresponding phoneme representation, and the duration of the phoneme representation is calculated to obtain the original phoneme data.

[0188] In step S703 of some embodiments, parameters such as phoneme duration, pitch, and intensity in the original phoneme data are adjusted based on the speech emotion characteristics to obtain target phoneme data.

[0189] In step S704 of some embodiments, the target phoneme data is synthesized into a target response speech through a pre-trained speech synthesis model, thereby giving the speech data the speech emotion expression of the target object and improving the naturalness and intimacy of the synthesized speech.

[0190] In step S109 of some embodiments, digital human synthesis technology can be called, such as a three-dimensional modeling and rendering engine (such as the Blender engine, the Unity engine), to synthesize a digital human based on the target reply voice and facial emotional features, thereby obtaining a digital human whose posture, facial expression, and voice emotional expression are similar to those of the target object. When the target digital human speaks the target reply voice, its body posture and facial expression will change, that is, the "digital version" of the target object.

[0191] For example, if the target person displays positive emotions when communicating with customers, the target reply voice will sound enthusiastic and cheerful, while the target digital person's facial expressions and body language will appear positive and open.

[0192] It is understandable that in financial scenarios, especially in the sales process of financial products, the target object (i.e., the business agent) has a strong personal color and personal style. Speech synthesis is performed based on the target object's voice emotional characteristics and the target answer content to obtain the target answer voice. Digital human synthesis is performed based on the target object's facial emotional characteristics and the target answer voice. The final target digital human has the target object's personal color and personal style, making it easier for customers to accept the digital human image and increasing the intimacy of the target questioner.

[0193] Moreover, when the target person is busy or taking a break, the target digital person can still handle business questions and answers with the customer on behalf of the target person, which improves customer satisfaction and helps to further improve the conversion rate of financial products.

[0194] See also Figure 8 The embodiment of the present application further provides a business question-answering device that can implement the above-mentioned business question-answering method, and the device includes:

[0195] The historical business data acquisition module 801 is used to acquire the historical business data of the target object; wherein the historical business data includes historical customer data and historical business audio and video data;

[0196] Emotional feature extraction module 802, used to extract emotional features from historical business audio and video data to obtain voice emotional features and facial emotional features;

[0197] The work performance feature extraction module 803 is used to extract features from historical business audio and video data to obtain target work performance features of the target object;

[0198] The customer group portrait construction module 804 is used to construct a user portrait based on historical customer data to obtain a target customer group portrait;

[0199] The business case screening module 805 is used to screen target business case data from a preset business database based on target work performance characteristics and target customer profiles;

[0200] The model optimization module 806 is used to optimize the preset original business question-answering model based on the target business case data to obtain a target business question-answering model;

[0201] The question answering module 807 is used to answer the pre-acquired user question data using the target business question answering model to obtain the target answer content; wherein the user question data is raised by the target questioner;

[0202] The speech synthesis module 808 is used to perform speech synthesis based on the speech emotion characteristics and the target response content to obtain the target response speech;

[0203] The digital human synthesis module 809 is used to synthesize a digital human based on the target answering voice and facial emotion features to obtain a target digital human, and use the target digital human to answer the target questioner.

[0204] The specific implementation of the business question-and-answer device is substantially the same as the specific implementation of the above-mentioned business question-and-answer method, and will not be described in detail here.

[0205] The present application also provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned business question-and-answer method when executing the computer program. The electronic device can be any smart terminal including a tablet computer, an in-vehicle computer, or the like.

[0206] See also Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is shown. The electronic device includes:

[0207] The processor 901 can be implemented as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0208] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 902 and is called by the processor 901 to execute the business question-and-answer method of the embodiments of this application;

[0209] Input / output interface 903, used to implement information input and output;

[0210] Communication interface 904, used to implement communication interaction between this device and other devices, which can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WiFi, Bluetooth, etc.);

[0211] Bus 905 , which transmits information between various components of the device (e.g., processor 901 , memory 902 , input / output interface 903 , and communication interface 904 );

[0212] The processor 901 , the memory 902 , the input / output interface 903 and the communication interface 904 are connected to each other in communication within the device via a bus 905 .

[0213] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned business question-and-answer method is implemented.

[0214] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0215] The business question-answering method and device, electronic device, and storage medium provided by the embodiments of the present application obtain historical business data of a target object, wherein the historical business data includes historical customer data and historical business audio and video data; perform emotional feature extraction on the historical business audio and video data to obtain voice emotional features and facial emotional features; perform feature extraction on the historical business audio and video data to obtain target work performance features of the target object; construct a user profile on the historical customer data to obtain a target customer group profile; thereby, based on the target work performance features and the target customer group profile, target business case data is filtered out from a preset business database to optimize the original business question-answering model to obtain a target business question-answering model, thereby enabling the target business question-answering model to answer user question data raised by the target question-asking object and obtain target answer content, thereby improving the accuracy of the business answer; further, perform speech synthesis based on the voice emotional features and the target answer content to obtain a target answer voice; perform digital human synthesis based on the target answer voice and facial emotional features to obtain a target digital human, and use the target digital human to answer the target question-asking object. Thus, the interactivity and real-time nature of the business question-answering are enhanced, and the efficiency of the business question-answering and user satisfaction are effectively improved.

[0216] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0217] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0218] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0219] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0220] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0221] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0222] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the above-mentioned units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0223] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0224] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0225] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes multiple instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: various media that can store programs, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0226] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A business question and answer method, characterized in that: The method comprises: Acquire historical business data of the target object; wherein the historical business data includes historical customer data and historical business audio and video data; Extracting emotional features from the historical business audio and video data to obtain voice emotional features and facial emotional features; Extracting features from the historical business audio and video data to obtain target work performance features of the target object; Constructing a user profile for the historical customer data to obtain a target customer group profile; Filtering target business case data from a preset business database based on the target work performance characteristics and the target customer group portrait; Optimizing the preset original business question-and-answer model based on the target business case data to obtain a target business question-and-answer model; The pre-acquired user question data is answered by the target business question-answering model to obtain target answer content; wherein the user question data is raised by the target questioning object; Perform speech synthesis based on the speech emotion feature and the target answer content to obtain the target answer speech; A digital human is synthesized based on the target answering voice and the facial emotion features to obtain a target digital human, and the target digital human is used to answer the target questioning object.

2. The method according to claim 1, characterized in that The business database includes a plurality of business expert data, each of which includes expert work performance characteristics, business customer group portraits, and expert business cases; the target business case data is screened out from the preset business database based on the target work performance characteristics and the target customer group portraits, including: For each of the business expert data, similarity calculation is performed on the target work performance characteristics and the expert work performance characteristics to obtain work performance similarity data, and similarity calculation is performed on the target customer group portrait and the business customer group portrait to obtain customer group similarity data; Performing aggregation operation on the work performance similarity data and the customer group similarity data to obtain original business similarity data; Filtering the plurality of original business similarity data based on a preset business similarity threshold to obtain target business similarity data; Determining target expert data from the business database based on the target business similarity data; The expert business case of the target expert data is acquired to obtain the target business case data.

3. The method according to claim 1, characterized in that The optimizing the preset original business question-and-answer model based on the target business case data to obtain the target business question-and-answer model includes: Perform case feature analysis on the target business case data to obtain target case parameter information; The parameters of the original business question and answer model are adjusted based on the target case parameter information to obtain the target business question and answer model.

4. The method according to claim 1, characterized in that: The historical customer data includes a plurality of historical customer information, and the user profile construction of the historical customer data to obtain the target customer group profile includes: Classifying the historical customer information by customer type to obtain historical customer type data; Performing clustering processing on the historical customer type data to obtain original customer group clustering data; wherein the original customer group clustering data includes a plurality of clustering points, each of which corresponds to one piece of historical customer information; The customer type to which the clustering points have the largest number is selected from the original customer group clustering data to obtain the target customer group portrait.

5. The method according to any one of claims 1 to 4, characterized in that: The historical business audio and video data includes historical business voice data and historical business video data; the emotional feature extraction of the historical business audio and video data to obtain voice emotional features and facial emotional features includes: Performing audio extraction on the historical service video data to obtain original video and voice data; Performing voice emotion extraction on the original video voice data and the historical business voice data to obtain the voice emotion features; Extracting video frames from the historical business video data to obtain original video frame images; Performing image preprocessing on the original video frame image to obtain an original face video image; The facial expression is extracted from the original face video image to obtain the facial emotion feature.

6. The method according to claim 5, characterized in that The extracting features of the historical business audio and video data to obtain the target work performance features of the target object includes: Extracting speech features from the historical service speech data to obtain target speech performance features; Extracting image features from the historical business video data to obtain target image performance features; The target work performance feature of the target object is obtained by combining the target voice performance feature and the target image performance feature.

7. The method according to any one of claims 1 to 4, characterized in that: The performing speech synthesis based on the speech emotion feature and the target answer content to obtain the target answer speech includes: Performing text preprocessing on the target reply content to obtain target reply text data; Performing phoneme conversion on the target reply text data to obtain original phoneme data; Performing emotion adjustment on the original phoneme data based on the speech emotion feature to obtain target phoneme data; Speech synthesis is performed based on the target phoneme data to obtain the target reply speech.

8. A business question-and-answer device, characterized in that: The device comprises: A historical business data acquisition module is used to acquire the historical business data of the target object; wherein the historical business data includes historical customer data and historical business audio and video data; An emotion feature extraction module is used to extract emotion features from the historical business audio and video data to obtain speech emotion features and facial emotion features; A work performance feature extraction module, used to extract features from the historical business audio and video data to obtain target work performance features of the target object; A customer group portrait construction module is used to construct a user portrait based on the historical customer data to obtain a target customer group portrait; A business case screening module, used to screen target business case data from a preset business database based on the target work performance characteristics and the target customer group portrait; A model optimization module, used to optimize the preset original business question-answering model based on the target business case data to obtain a target business question-answering model; A question answering module is used to answer the pre-acquired user question data through the target business question answering model to obtain target answer content; wherein the user question data is raised by the target questioning object; A speech synthesis module, used for performing speech synthesis based on the speech emotion feature and the target reply content to obtain the target reply speech; The digital human synthesis module is used to synthesize a digital human based on the target answering voice and the facial emotion features to obtain a target digital human, and use the target digital human to answer the target questioning object.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images