Intelligent voice and man-machine interaction system

Abstract

The invention provides an intelligent voice and man-machine interaction system, which belongs to the field of artificial intelligence and comprises a registration center module, a gateway module, a voice receiving module, a voice processing module, a voice responding module and a background management module. A traditional one-to-one manual customer-to-customer relationship is changed into intelligent voice to help customers to answer questions, so that the manpower cultivation cost is reduced, a more intelligent voice system is used for replacing manual answering at any time in 24 hours, the problem that customers queue up to wait in rush hours or at night is solved, the user experience is improved, and the user experience is improved. The satisfaction degree of the user is increased, and the expenditure of an enterprise can be greatly reduced.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a system for intelligent voice and human interaction. Background Art

[0002] With the development and growth of Internet technology, our lives and work have also entered the Internet era. Voice interaction, as a new interactive technology, has been widely used in various industries: home smart speakers, voice-controlled car terminals, personal voice assistants, conference system voice control, etc. At present, there are still a large number of companies that still rely heavily on people to solve most of the problems of their customers. The cost of human resource training, the cost of manpower and time, and the expenses are very huge. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a system for intelligent voice and human interaction.

[0004] The technical solution of the present invention is:

[0005] A system for providing intelligent voice and human interaction, comprising:

[0006] Registration center: used to connect the components in the system in series and register them through the registration center. It is used to register various services in the system, such as gateway module, voice receiving module, customer voice processing module, and customer voice response module.

[0007] Gateway module: Responsible for request forwarding, filtering and forwarding received requests to the telephone, video, and online modules for processing. Request forwarding, synthesis, and protocol conversion. All requests from clients must first pass through the gateway, and then route these requests to the corresponding microservices. The gateway will often process a request by calling multiple microservices and aggregate the results of multiple services. It can convert between web protocols and non-web-friendly protocols used internally, such as HTTP and WebSocket.

[0008] Voice reception: Voice reception module, used to predict the voice feedback time when the customer receives the intelligent voice customer service to reply to the customer voice data. The voice feedback time includes the customer voice data transmission time Tre, the customer voice data processing time Tpro, the customer service voice data generation time Tgen and the customer service voice data transmission time Tse. Specifically, the call process between the customer and the intelligent voice customer service is a process of back-and-forth interaction through voice data. During a call between the customer and the intelligent voice customer service, the customer will send the customer voice data to the intelligent voice customer service system through the client program many times. After receiving the customer voice data each time, the intelligent voice customer service system will identify the customer voice data received at that time and generate the corresponding reply content, and then convert it into customer service voice data and send it to the client program. That is, the customer service voice data is the reply voice generated by the intelligent voice customer service for each received customer voice data. According to the complexity of the content involved in the customer voice data, the time for generating the reply content in real time through the artificial intelligence model will also be different. When the customer's question is more complicated, the intelligent voice customer service system needs a long time to generate its reply content, which causes the customer to wait for a long time to receive the reply of the intelligent voice customer service after asking the question. For some simple and common questions, the intelligent voice customer service system can generate a reply and send it to the customer in a very short time. Whether the reply is too fast or too slow, it will cause discomfort to the customer. According to research, the most comfortable response waiting time for customers during a call is between 600 milliseconds and 1200 milliseconds. The cloud management platform automatically monitors the expiration of services periodically. After the service expires, the license is released and the resources are temporarily stored for a certain retention period. When the retention period expires, the released resources can be automatically destroyed.

[0009] Backstage management:

[0010] Used to manage permissions, add and modify product categories, add and modify products, and review products. Permissions are assigned according to roles, and some functions are displayed in roles with permissions. Set as an administrator and other roles.

[0011] The beneficial effects of the present invention are

[0012] Given that most companies still use manual methods to answer customer questions by phone or video, the labor cost is huge and the training time cycle is long. There is a time limit for answering customer needs. Using a smarter voice system to replace manual has many advantages. It can answer customer questions 24 hours a day, solve the problem of customers waiting in line during peak hours or at night, thereby increasing user experience and user satisfaction, and greatly reducing expenses for companies. It can also expand the acquisition of information and the issuance of commands without the need for people to use input devices such as keyboards and mice, which is more convenient and fast. Whether in scenarios such as home, work, and study, intelligent voice can provide quick help. Intelligent voice can learn user voice habits through technologies such as deep learning and natural language processing, and can provide more appropriate services based on the user's interests, hobbies, habitual expressions and other personalized characteristics. In this way, in daily use, artificial intelligence voice can provide users with a more intimate and comfortable experience. In summary, intelligent voice brings people a convenient and fast way of interaction, intelligent service content and personalized interactive experience, while also improving work efficiency and developing more application scenarios, bringing humans a more intelligent, efficient and comfortable life experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a schematic diagram of the system structure of the present invention;

[0014] Figure 2 It is a schematic diagram of the workflow of the present invention. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0016] The present invention provides a system of intelligent voice and human interaction, which will change the traditional one-to-one relationship between human and customer to intelligent voice to help customers answer questions, thereby reducing the cost of human training, using a more intelligent voice system to replace human, answering questions at any time 24 hours a day, solving the problem of customers queuing during peak hours or at night, thereby increasing user experience and user satisfaction, and greatly reducing expenses for enterprises. It includes: registration center module, gateway module, voice reception, voice processing, voice response, and background management.

[0017] 1) Registration center: used to connect the components in the system in series and register through the registration center. It is used to register various services in the system, such as gateway module, customer voice receiving module, customer voice processing module, and customer voice response module. It detects the health of the system and makes it easy to build dynamic service discovery, configuration management, and service management platforms for cloud-native applications. It helps to quickly implement dynamic service discovery, service configuration, service metadata, and traffic management. It builds a service infrastructure for modern application architectures centered on "services" (such as microservice paradigm and cloud-native paradigm).

[0018] 2) Gateway module: responsible for request forwarding, synthesis and protocol conversion. All requests from clients must first pass through the gateway, which then routes these requests to the corresponding services. The gateway will often process a request by calling multiple microservices and aggregate the results of multiple services. It can convert between web protocols and non-web-friendly protocols used internally, such as HTTP and WebSocket. Responsible for request forwarding, filtering and forwarding received requests to the telephone, video, and online modules for processing.

[0019] 3) Voice reception: The voice reception module is used to receive the voice feedback time when the customer receives the intelligent voice customer service's reply to the customer voice data; the voice feedback time includes the customer voice data transmission time Tre, the customer voice data processing time Tpro, the customer service voice data generation time Tgen and the customer service voice data transmission time Tse. Specifically, the call process between the customer and the intelligent voice customer service is a process of back-and-forth interaction through voice data. During a call between the customer and the intelligent voice customer service, the customer will send the customer voice data to the intelligent voice customer service system through the client program many times. After receiving the customer voice data each time, the intelligent voice customer service system will identify the customer voice data received at that time and generate the corresponding reply content, and then convert it into customer service voice data and send it to the client program. That is, the customer service voice data is the reply voice generated by the intelligent voice customer service for each received customer voice data. According to the complexity of the content involved in the customer voice data, the time for generating the reply content in real time through the artificial intelligence model will also be different. When the customer's question is more complicated, the intelligent voice customer service system needs a long time to generate its reply content, which causes the customer to wait for a long time to receive the reply from the intelligent voice customer service after asking the question. For some simple and common questions, the intelligent voice customer service system can generate a reply and send it to the customer in a very short time. Whether the reply is too fast or too slow, it will cause discomfort to the customer. According to research, the most comfortable response waiting time for customers during a call is between 600 milliseconds and 1200 milliseconds.

[0020] 4) Speech processing: ① Speech preprocessing, including amplitude normalization, frequency response correction, framing, windowing, and start and end point detection. ② Speech acoustic parameter analysis, including parameters such as speech formant frequency and amplitude, as well as analysis of speech linear prediction parameters and cepstrum parameters. ③ Parameter normalization, mainly normalization on the time axis, commonly used methods include dynamic time warping (DTW) or dynamic programming (DP). ④ Pattern matching, which can use distance criteria or probability rules, or syntactic classification. ⑤ Recognition judgment, giving the recognition result through the final discriminant function. Speech recognition can be classified according to different recognition contents: phoneme recognition, syllable recognition, word or phrase recognition; it can also be classified according to vocabulary: small vocabulary (less than 50 words), medium vocabulary (50 to 500 words), large vocabulary (more than 500 words) and super large vocabulary (tens to tens of thousands of words). Classification according to pronunciation characteristics: it can be divided into recognition of isolated sounds, connected sounds and continuous sounds. According to the requirements for the speaker, there are person-recognition, that is, recognition of only a specific speaker, and person-independent recognition, that is, recognition of anyone regardless of the speaker. Obviously, the most difficult speech recognition is the one that meets the requirements of large vocabulary, continuous sound and person-independent recognition at the same time. Speech understanding uses artificial intelligence technologies such as knowledge expression and organization to automatically recognize sentences and understand the meaning of words. The main difference from speech recognition is the degree of full utilization of grammatical and semantic knowledge. Using the understanding ability, the system can improve performance: ① It can eliminate noise and clutter; ② It can understand the meaning of the context and use it to correct errors and clarify uncertain semantics; ③ It can handle ungrammatical or incomplete sentences.

[0021] 5) Voice response: Due to the influence of various factors such as the customer's expression, related scenarios, and the comprehensiveness of the questions in the standard reply database, the standard reply content corresponding to the customer's voice data is determined from the standard reply database, and its reply content is often not as good as the reply content generated in real time by the intelligent voice customer service model to meet the customer's needs. When the customer is not satisfied with the reply to the standard reply content, there is a high probability that the same question will be raised again in the same call. In the technical scheme of the above implementation, while determining the standard reply content corresponding to the customer's voice data from the standard reply database, another thread calls the intelligent customer service voice data generation module to generate the customer service voice data through the pre-trained intelligent voice customer service model and cache it in the memory. When the customer asks the same question again, the reply content can be directly taken out from the cache and sent to the customer. The reply text generation submodule is used to input the customer text data into the pre-trained intelligent voice customer service model to generate customer service text data. The customer service text data is the reply text generated in real time by the intelligent voice customer service model based on the natural semantic understanding of the customer text data.

[0022] 6) Backstage management: You can set and modify various information and introductions of the company, view, modify and delete member information, and lock and unlock functions to send letters to members online. New database connection, online table editing, add data tables, edit databases, add and edit file connection websites, used for management permissions, product category addition and modification management, product addition and modification, and product review, and assign permissions according to roles. Some functions are displayed in roles with permissions. Set as administrator and other roles.

[0023] The above description is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, and is not used to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A system for intelligent voice and human interaction, characterized in that: include: Registration center: used to connect components in the system in series and register them through the registration center to register various services in the system; Gateway module: responsible for request forwarding, synthesis and protocol conversion; The voice receiving module is used to predict the time it takes for the customer to receive the voice feedback from the intelligent voice customer service to reply to the customer's voice data; Backstage management: used to manage permissions, add and modify product categories, add and modify products, and review products. Permissions are assigned according to roles, and some functions are displayed in roles with permissions. Set as administrator and other roles.

2. The system according to claim 1, characterized in that All requests from clients must first pass through the gateway, which then routes these requests to the corresponding microservices; It processes a request by calling several microservices and aggregating the results of several services, converting between web protocols and non-web-friendly protocols used internally.

3. The system according to claim 2, characterized in that Responsible for request forwarding, filtering and forwarding the received requests to the telephone, video, and online modules for processing.

4. The system according to claim 1, characterized in that Voice feedback time includes customer voice data transmission time, customer voice data processing time, customer service voice data generation time and customer service voice data transmission time.

5. The system according to claim 4, characterized in that During a call between a customer and the intelligent voice customer service, the customer sends voice data to the intelligent voice customer service system through the client program. After each time the intelligent voice customer service system receives the customer voice data, it will recognize the customer voice data received at that time and generate the corresponding reply content, and then convert it into customer service voice data and send it to the client program.

6. The system according to claim 5, characterized in that Customer service voice data is the reply voice generated by the intelligent voice customer service for each customer voice data received. Depending on the complexity of the content involved in the customer voice data, the time for generating the reply content in real time through the artificial intelligence model will also vary.

7. The system according to claim 6, characterized in that The customer's response waiting time during a call is between 600 milliseconds and 1200 milliseconds. The cloud management platform automatically monitors the service expiration period. After the service expires, the license is released and the resources are temporarily stored for a set retention period. When the retention period expires, the released resources can be automatically destroyed.

8. The system according to claim 1, characterized in that Also includes speech processing modules: 1) Speech preprocessing, including speech amplitude normalization, frequency response correction, framing, windowing, and start and end point detection; 2) Speech acoustic parameter analysis, including analysis of speech formant frequency, amplitude parameters, as well as speech linear prediction parameters and cepstrum parameters; 3) Parameter normalization, that is, normalization on the time axis, using dynamic time warping (DTW) or dynamic programming (DP); 4) Pattern matching, using distance criteria or probability rules, or syntactic classification; 5) Recognition judgment: the recognition result is given through the final discriminant function.

9. The system according to claim 1, characterized in that It also includes a voice response module: while determining the standard reply content corresponding to the customer voice data from the standard reply database, another thread calls the intelligent customer service voice data generation module to generate customer service voice data through a pre-trained intelligent voice customer service model and cache it in the memory. When the customer asks the same question again, the reply content can be directly taken out of the cache and sent to the customer.

10. The system according to claim 9, characterized in that The reply text generation submodule is used to input the customer text data into a pre-trained intelligent voice customer service model to generate customer service text data. The customer service text data is the reply text generated in real time by the intelligent voice customer service model based on the natural semantic understanding of the customer text data.

Images