Intelligent dialogue importing method and system

Abstract

An intelligent dialogue import method and system, the system proposing a cloud server, in which a processing circuit executes the intelligent dialogue import method, in which an online dialogue program is started, a first chat robot is imported, the first chat robot receives content input by a user through a dialogue interface, and the content is input to the cloud server. The method comprises the following steps: obtaining semantic features of contents input by a user, user data and real-time environment information obtained through an external system, and importing a second chat robot in a specific field to a dialogue interface according to the information; and dialogue contents conforming to the semantic meaning, the preference and the click environment information of the user are generated through a natural language model operated by the second chat robot and a generative artificial intelligence technology, and then the dialogue contents are output through a dialogue interface.

Description

Technical Field

[0001] The disclosed book is about a chat robot, and more particularly, it relates to an intelligent conversation importing method and system which can dynamically import a chat robot in a specific field according to the semantics of the conversation content. Background Art

[0002] Among the artificial intelligence (AI) that is currently developing rapidly in various fields, one type is a natural language chatbot that can process natural language and automatically generate content, such as the Chat Generative Pre-trained Transformer (ChatGPT) developed by OpenAI. This type of natural language chatbot uses generative artificial intelligence technology to train a large amount of data first, and then generate new data related to the original data, and build an intelligent model after deep learning (such as generative adversarial networks (GAN)).

[0003] Taking ChatGPT as an example, ChatGPT is trained by learning a large amount of online information and can communicate with users in natural language. However, the content it commonly responds to users is standard answers learned through learning, and it cannot adapt in real time to provide answers relevant to the user's current status. Although it is a natural language chatbot, it lacks content that is relevant to the user and conforms to the real-time situation.

[0004] Furthermore, in addition to the above defects, the services provided by current natural language chatbots are only for general discussions and cannot meet the needs of all fields. There is still a lot of room for improvement; but it is understandable that for specific fields, chatbots in specific fields can still be trained with data in specific fields to provide conversation services in that field. However, the existing technology still lacks chatbots that integrate various fields and can effectively solve the needs of enterprises and users. Summary of the invention

[0005] In order to improve the shortcomings of existing chatbots, the public book proposes an intelligent dialogue introduction method and system, which can dynamically introduce chatbots in corresponding fields according to the semantics of the dialogue content. Furthermore, natural language processing (NLP) and generative artificial intelligence (generative AI) technology can be used to implement this chatbot, and in the process of dialogue with the user, it can further refer to the user's preferences and real-time environmental information to provide dialogue content that meets personal needs and current situations.

[0006] The intelligent dialogue import system includes a cloud server, and a processing circuit executes an intelligent dialogue import method. In the method, an online dialogue program is started, and a first chat robot is imported, and the content input by the user is received through a dialogue interface, and then the semantic features, user data and real-time environmental information of the content input by the user are obtained, so that the semantic features of the content input by the user, the user preferences obtained from the user data and the real-time environmental information can be imported into the second chat robot according to any information or any combination of information. Then, the dialogue content can be generated through the natural language model running by the domain chat robot, and then the dialogue content is imported into the online dialogue program, and the dialogue content is output on the dialogue interface.

[0007] The first chat robot may be the default main chat robot proposed by the system, and the second chat robot may be a domain chat robot for a specific domain introduced by the system based on any or any combination of user semantics, user preferences and real-time environmental information.

[0008] The cloud server provides an external system interface for connecting to one or more external systems to obtain the real-time environmental information, including one or any combination of real-time weather, real-time traffic conditions, real-time news and real-time location-related network information.

[0009] Furthermore, the natural language model running in the cloud server uses a conversion model to perform machine translation, document summarization and document generation and other procedures to generate conversation content; and further executes a vector algorithm on the content input by the user, the user's preferences and the real-time environmental information, annotates the obtained text, calculates the vector of each word, obtains the related content according to the vector distance between the words, and generates conversation content that conforms to the user's preferences and the real-time environmental information.

[0010] Furthermore, a vector algorithm is further executed on the historical conversation records recorded in the database or memory of the cloud server to generate conversation content that can match the user's current emotions.

[0011] According to an embodiment, a cloud server runs a social media, and a corresponding social media application is executed in a user device, so that a user can click a conversation link icon on a page of the social media application to enter an online conversation program.

[0012] The page may be a map interface, on which a plurality of link points associated with geographic locations are marked, including one or more video link points.

[0013] Furthermore, in the cloud server, statistics are collected on the number of times the domain chatbot is introduced, the running time, the number of conversations, and the number of clicks and time of the recommended content by the domain chatbot, to provide a domain customer report.

[0014] The cloud server provides a video database, allowing the domain chat robot to query the video database based on the user's semantic features, user preferences and real-time environmental information to introduce one or more video contents into the dialogue interface.

[0015] Furthermore, the cloud server provides a multi-domain robot database, which includes multiple domain models trained by learning various domain expertise through machine learning algorithms. The multiple domain models use natural language processing technology and generative artificial intelligence technology to implement multiple domain chat robots with natural language processing capabilities. The domain chat robot that imports the dialogue interface is provided by this multi-domain robot database.

[0016] Furthermore, in addition to generating dialogue content, the domain chat robot also introduces one or more audio and video content into the dialogue interface based on the user's semantic features, user preferences and real-time environmental information.

[0017] To further understand the features and technical contents of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are only for reference and description and are not intended to limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A diagram showing an embodiment of a system architecture for running an intelligent dialogue import method;

[0019] Figure 2 A diagram showing an embodiment of a data structure for running an intelligent dialogue import method;

[0020] Figure 3 A diagram showing an embodiment of a multi-domain robot database;

[0021] Figure 4 One of the flow charts of the embodiment showing the intelligent dialogue import method;

[0022] Figure 5 The second flow chart of the embodiment showing the intelligent dialogue import method;

[0023] Figure 6 A flowchart showing an implementation example of establishing a domain chatbot;

[0024] Figure 7 A flowchart showing an implementation example of generating a domain customer report;

[0025] Figures 8 to 12 Displaying a sample diagram of a graphical user interface provided by the intelligent dialogue import system; and

[0026] Figures 13 to 17 A diagram showing an implementation example of a graphical user interface for an application related to an import domain chat robot. DETAILED DESCRIPTION

[0027] The following is an explanation of the embodiments of the present invention through specific embodiments. Those skilled in the art can understand the advantages and effects of the present invention from the contents disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and the details in this specification can also be modified and changed in various ways based on different viewpoints and applications without departing from the concept of the present invention. In addition, the drawings of the present invention are only simple schematic illustrations and are not depicted according to actual sizes. It is stated in advance. The following embodiments will further explain the relevant technical contents of the present invention in detail, but the disclosed contents are not intended to limit the scope of protection of the present invention.

[0028] It should be understood that, although the terms "first", "second", "third", etc. may be used herein to describe various components or signals, these components or signals should not be limited by these terms. These terms are mainly used to distinguish one component from another component, or one signal from another signal. In addition, the term "or" used herein may include any one or more combinations of the associated listed items depending on the actual situation.

[0029] The specification discloses a method and system for introducing intelligent dialogue. The method runs in a cloud server and uses natural language processing technology to provide intelligent dialogue services. The cloud server can provide social media services through the Internet, allowing users to share text, pictures and audio and video content after joining. The cloud server also provides chat robots in various fields, which can provide users with services provided by the cloud server to conduct dialogues, and further provide dialogue services in corresponding fields according to the features in the picture. Artificial intelligence technology is used, including using machine learning algorithms and natural language processing (NLP) technology to learn data from various fields to train chatbots to provide dialogue services, and by learning the user's activities in social media to obtain the user's preferences, so that the chatbot can provide dialogue content that is more in line with the user's personal needs and current environmental characteristics based on the user's dialogue meaning, user preferences, and real-time environmental information.

[0030] The above system implemented by the cloud server can refer to Figure 1 The system architecture embodiment diagram for running the intelligent dialogue import method is shown.

[0031] The figure shows a cloud server 100 implemented by a computer system, a database and a network, wherein various functional modules are implemented by software and hardware collaboration, such as a natural language processing module 101 for performing natural language processing as shown in the figure, wherein the natural language processing module 101 implements a chat robot with natural language processing capability; a machine learning module 103 runs a machine learning algorithm, and in addition to training a natural language model, can also learn the user's activities on the network through a deep learning method to obtain information about the user's preferences, so that the chat robot can provide conversation content that meets the user's preferences; the cloud server 100 provides an external system interface module 105, wherein circuits and related application software for connecting to an external system (such as an external system 1 111 and an external system 2 112) (such as via a network 10) and obtaining data through an application programming interface (API) are run; the cloud server 100 provides a user interface module 107, wherein the network connection function in the user interface module 107 allows a user device 150 to connect to the cloud server 100, and a server (web server) providing network services can be run, so that the corresponding application (such as a social media application) executed in the user device 150 can obtain the service provided by the cloud server 100.

[0032] According to the architecture shown in the figure, the cloud server 100 is provided with a built-in or externally connected database, and data services are provided through the cloud server 100, such as the audio and video database 110 shown in the figure, which provides the user device 150 with access via the network 10 to store the audio and video content uploaded and shared by users at each end, and can include text and images; the user database 120 stores user data, including user personal data, uploaded text, images and audio and video content, and obtains the user's activity data in the network service provided by the cloud server 100, such as the content browsed, tracking, liking, sharing and subscription network activities, which can form a user profile, and as the conversation content continues to be generated over time, the user database 120 can store and update the user data according to the time dimension, including recording the user's historical conversation record, which becomes a conversation record for the machine learning algorithm in the natural language model; the vector database 130 is a structured information that records various texts, images and audio and video content after vectorization calculation, which can be used to compare various data that conform to the user's personalization. The data provided by the system running the intelligent conversation import method also includes a multi-domain robot database 140, which can be referenced Figure 3 Shown embodiment, in the multi-domain robot database 140.

[0033] When the domain model is used in the intelligent dialogue import method, the system can determine the domain chat robot corresponding to the user's needs based on the semantic features in the dialogue information provided by the user. For example, when the user talks about his own needs for exercise, after processing by the natural language processing module 101, the exercise-related semantics are obtained, that is, the exercise-related domain chat robot is imported from the multi-domain robot database 140, so that the user can obtain exercise-related information from the chat robot in the sports field during the chat process, so that the system can provide better dialogue services in various fields.

[0034] Furthermore, the multi-domain robot database 140 may include domain chatbots for specific business needs. When an enterprise has a need for commercial promotion, it can submit the need to the intelligent dialogue import system implemented by the computer system, and the machine learning module 103 in the cloud server 100 learns the data provided by the enterprise, such as the introduction of advertising goods (services, products), and the public content related to the enterprise on the Internet, and obtains a domain chatbot for the business promotion needs of the enterprise. After being imported into the dialogue process with the user, the user can obtain the business data of interest from the dialogue process to achieve the purpose of commercial promotion.

[0035] According to the system architecture diagram shown in the figure, the cloud server 100 obtains data from the external system through the network 10 or the connection under a specific protocol, such as the external system 1 111 and the external system 2 112 schematically shown in the figure, which are, for example, servers established by the government or enterprises to provide open data, so that the cloud server 100 can obtain demand-compliant and real-time information through the external system interface module 105 using the application programming interface provided by each external system, such as real-time weather, real-time traffic conditions, real-time news and real-time location-related network information.

[0036] The user device 150 executes an application that can obtain services provided by the cloud server 100. For example, if the cloud server 100 provides social media services, the user device 150 executes the corresponding social media application and obtains the social media services through the user interface module 107. In particular, the cloud server 100 provides a natural language chat robot through the natural language processing module 101, so that the user can have a conversation with the chat robot through the dialogue interface 115. On the other hand, the cloud server 100 can learn the activity data of the user using various services of the cloud server 100 through the machine learning module 103, including obtaining the activity data of the user using the social media application and the dialogue interface 115, so that the machine learning module 103 can learn the user's interest characteristics and establish user data.

[0037] It is worth mentioning that the various texts, images and audio-visual contents obtained by the cloud server 100 are unstructured information, which can be converted into vectorized data through encoding to facilitate the acquisition of the meaning, which is conducive to data search. Furthermore, the vectorized data can be used to compare the user's search keywords, and a distance function is used to calculate the distance between the search keywords and the vectorized data in the database. The closer the distance, the closer the data, so that users can search for data through the vector database 130. Similarly, the semantics can also be judged according to the user's conversation keywords to automatically import the corresponding domain chat robot.

[0038] According to an embodiment, the vector database 130 of the cloud server 100 can support multi-modal search services such as text and images, wherein structured information is provided, such as various texts, patterns and audio-visual contents are converted into text, and vectorized data is obtained by vector algorithm calculation, which can be applied to search services and can be used in natural language processing programs. The natural language processing program will use a natural language model to map the vectorized data to a vector space. Taking the words input by the user as an example, a word vector is obtained after vector calculation.

[0039] According to the embodiment, the intelligent dialogue introduction method proposed in the open book is implemented as a multi-domain chat robot running on the cloud server 100. The chat robot can communicate with the user in natural language, including text and voice. In addition to responding to the information input by the user, it can also obtain the user data through the cloud server before the conversation, from which the user's personality and habits can be derived. In addition, the real-time status can be obtained from the external system (111, 112), for example, the local weather and news can be obtained according to the user's location, so that the reply content can not only be targeted at the user's preferences, but also reflect the actual status.

[0040] For example, when a user mentions his or her needs during a chat with the chatbot proposed in the open book, such as mentioning that it is time to eat, the chatbot will provide recommendations for meals and restaurants based on the user's favorite meals learned in the past, as well as the user's current location or the location information mentioned by the user in the conversation, rather than just obtaining answers from the database that has already been learned.

[0041] Furthermore, trained chatbots in various fields can be set up in the cloud server running the intelligent dialogue import method. When the user expresses the need for further information in the dialogue, the chatbot can introduce chatbots in related fields (such as business / product / field robots of a restaurant, food court, or night market), so that the chatbots in related fields can continue to communicate with the user in natural language and provide more professional and accurate dialogue content. For example, when the user expresses the need to find a suitable restaurant in the dialogue, but also has calorie considerations and fitness needs, the system can import chatbots in the field related to healthy diet and exercise, and provide users with relevant information on how to exercise through sports that the user is interested in after obtaining excess calories in the dialogue.

[0042] The conversation service provided by the above system can be a function running in social media. The user's activities in social media are also the data for the system to learn the user's preferences and form structured data in the system. Figure 2 The data structure embodiment diagram of the intelligent dialogue import method is shown, wherein the data is divided into social media platform data (social media platform) 21, user data (such as user profile) 23 and user activity data (user profile data) 25.

[0043] The social media platform data 21 is non-public data in the system. The system that runs the intelligent dialogue import method obtains the viewer data (viewers) 211 of users accessing various contents provided by the cloud server, and the creator data (creators) 212 of users who provide various contents in the system. It also includes business data (business) 213 provided by the system to enterprises to establish corporate data for advertising, and, because the system can provide localized services, various location data (locations) 214 related to geographic locations will be obtained.

[0044] User data 23 is publicly available data in the system, covering data edited by users themselves, including viewer data 231 obtained by the system from various user activity data, which may include the user's interest data as a viewer obtained through machine learning, such as recent interest data, historical interest data, and location-related interest data.

[0045] The creator data 232 in the user data 23 is the relevant data of the user as a creator, covering the data related to the creator's preference types and the creator's location learned by the system through machine learning. For example, it may include the data of the user as a creator, and the types and locations of the creator's preferences learned through learning, including geographic location, or a specific location of a place.

[0046] The business data 233 in the user data 23 includes the business type of the enterprise and its product features obtained by the system through machine learning when the user is an enterprise.

[0047] The user activity data 25 is non-public data in the system, which includes statistical data of the user's activities in various services provided by the cloud server, and includes data obtained through machine learning, mainly including viewer data 251, creator data 252 and business data 253.

[0048] Viewer data 251 refers to the browsing rate, browsing time, and activity data such as following, liking, commenting, and subscribing of users using the services provided in the cloud server; creator data 252 refers to the statistical data when the user acts as a creator, such as the followers of the channel or account, the number of views of the invented content and the account browsing rate; business data 253 refers to the followers, content views, and overall impression data obtained when the user is a business.

[0049] The social media platform data 21, user data 23 and user activity data 25 collected and learned by the cloud server above become the basis for the dialogue service proposed in the open book using natural language processing and generative artificial intelligence technology. The cloud server calculates the above data through processing circuits to provide a chat robot that meets the personalized and real-time needs of users.

[0050] exist Figure 2 Under the data structure for running the intelligent dialogue import method shown, for users who are enterprises, specific organizations or individuals that provide commercial services or products, the intelligent dialogue import system can use the commercial data 213 of the social media platform data 21 to train domain chatbots in related fields, thereby providing business promotion for enterprises, organizations or individuals. In another embodiment, the intelligent dialogue import system can also use data from various professional fields to train domain robots in different fields, and can provide professional dialogue content in specific fields according to user needs.

[0051] according to Figure 3The diagram of an embodiment of a multi-domain robot database is shown, which describes a multi-domain robot database 140 that is externally connected to a cloud server 100 or built into a cloud server 100, wherein a plurality of chat robots are provided, including a main chat robot 300 that provides general conversation services to users, which may also be a preset chat robot at the beginning of a conversation, wherein natural language processing technology is run to conduct natural language conversations and parse out the user's semantic features, and it can refer to the real-time environmental information of the conversation to determine whether a specific domain chat robot needs to be introduced.

[0052] The diagram schematically shows that the multi-domain robot database 140 provides multiple domain chat robots trained by learning data from different domains, such as domain chat robot 1 301, domain chat robot 2 302 and domain chat robot 3 303. When used, chat robots in different domains can also be merged to form another domain chat robot. For example, sports domain robots can provide conversational capabilities for various sports professions, and the system can also provide domain chat robots for training specific sports for specific sports; the intelligent conversation import system can also train domain chat robots that meet the needs of the user according to the user's requirements. Such chat robots can assist the user in performing conversation services for promoting specific products. The multi-domain robot database 140 provides a robot import interface 30 that interfaces with the cloud server 100. One of the domain chat robots can be selected according to the instructions generated by the cloud server 100, and imported into the cloud server 100 through the robot import interface 30.

[0053] According to an embodiment, in a cloud server, the natural language model running therein can first perform a vector algorithm on the content input by the user through the dialogue interface, the user's preferences, and the real-time environmental information to annotate the obtained text, calculate the vector of each word, and obtain the relevant content after querying the database based on the vector distance between the words. Based on this, the chatbot using the natural language model can generate dialogue content that conforms to the user's preferences and real-time environmental information. During the online dialogue process, the natural language model running behind the chatbot can use the transformer model to perform machine translation, document summarization, and document generation on the textual data. Please refer to Figure 4 The description of the process of the embodiment shown in the figure is to obtain the user's meaning in the conversation, so that the chat robot (such as Figure 3 The displayed main chat robot 300, domain chat robot one 301, domain chat robot two 302 and domain chat robot three 303, etc.) can generate dialogue content.

[0054] The conversation program obtained by implementing the intelligent conversation importing method described in the above embodiment can actually be run by starting a graphical user interface (GUI) through an application (such as a social media application) executed in the user's device to display the conversation content between the chat robot and the user. The intelligent conversation importing method executed by the system using the cloud server architecture can be referred to Figures 4 to 7 The embodiment flow chart shown in FIG. Figures 8 to 12 The example diagram of the graphical user interface provided by the intelligent dialogue import system shown, and Figures 13 to 17 The diagram shown is a diagram of an implementation example of a graphical user interface for an application related to an imported domain chat robot. The example shown in the diagram is not intended to limit the actual operation method.

[0055] exist Figure 4 In the process shown, at the beginning, the user uses the application running in the user device to connect to the cloud server, use the services provided therein, and browse the content therein. The cloud server provides (such as through Figure 1 The graphical user interface module 107 of the display can be used to browse text, graphics and audio and video content. Figure 8 The map interface 80 with an electronic map as the background is started after the user device executes the application program, wherein the map interface 80 is schematically displayed, and a link icon (icon) thereof is marked on the map interface 80 according to the geographic information associated with each piece of video content. For example, the multiple link points associated with the geographic location in the figure may include one or more video link points, and the accompanying figure uses the video link points 801, 802 and 803 as examples. According to an embodiment, the multiple video link points marked on the map interface 80 may include points of interest (POIs) recommended by the system within the geographic range.

[0056] One or more audio-visual contents of a point of interest (POI).

[0057] According to an embodiment, taking social media as an example, a corresponding social media application is executed in a user device of a user terminal, wherein a page is started, such as Figure 8 The map interface 80 shown, while browsing the content on the map interface 80, has several icons at the bottom indicating link icons such as play 811, dialogue 812, assistant 813, search 814 and user homepage 815. At this time, the user can click (touch or specific gesture) the dialogue link icon 812, or click a link point marked somewhere on the map interface 80 that provides a dialogue, to start the online dialogue program (step S401).

[0058] on the other hand, Fig. 9Another schematic diagram of a link method for starting an online dialogue program is provided. For example, when a user selects any audio and video link point (801, 802 or 803) on the map interface 80, the screen starts as shown in FIG. Fig. 9 The video playback page 90 is displayed for playing a video shared or created by a user. The figure also shows the geographic location 901 of the video. The sidebar displays several link icons, such as favorite 903, conversation 904, collection 905 and share 906. The user can also start the online conversation program by clicking the conversation link icon of conversation 904 (step S401).

[0059] Then, a dialogue interface is started, allowing the user to input text, graphics or specific audio and video content (e.g., input a link to share audio and video content) through the dialogue interface, so that the cloud server receives the content input by the user through the user interface module (step S403). According to an embodiment, the online dialogue program is implemented as a chat robot using a natural language model (refer to Figure 3 The multi-domain robot database 140 is displayed. At the beginning, a preset main chat robot 300 can be imported, which can communicate with the user through a dialogue interface, receive the dialogue content input by the user, and perform natural language processing for the content input by the user each time.

[0060] The example of the dialogue interface can be found in Fig.10 Displayed dialogue interface 1000, Fig.11 The displayed dialog interface 1110 and Fig.12 The displayed dialogue interface 1200 and the like, each example dialogue interface will provide an input field for the user to input content, and a dialogue display area for displaying the dialogue content output by the chatbot and the content input by the user.

[0061] At this time, the cloud server obtains the content input by the user through the user interface module, wherein the content received through the dialogue interface can be text, voice or audio and video content. If the received content is voice or audio and video content, it is converted into text through a text conversion program, and then a natural language processing module is used to perform semantic analysis to obtain semantic features (step S405). During the execution of the above program, the cloud server obtains user data from the user database, and the external system (which can be Figure 1 The external system interface module 105 displayed obtains real-time environment information (step S407).

[0062] While the conversation is going on, the software program executing the intelligent conversation import method in the cloud server can judge in real time whether to import a domain chat robot in a specific domain according to the semantic features in the conversation, any information or any combination of information of user preferences and real-time environmental information. After judging to import the domain chat robot (step S409), the database is also queried to obtain the semantic features of the content that matches the user input, the user preferences obtained from the user data, and the content of the real-time environmental information (step S411). In the online conversation program, the natural language model running in the domain chat robot generates the conversation content (step S413). Afterwards, the conversation content is imported into the online conversation program, the conversation content is output in the conversation interface (step S415), and the above-mentioned step S403 and other embodiment processes are continuously executed.

[0063] Furthermore, when the natural language model of the cloud server is running, it uses a database or system memory to record information of multiple dimensions, which may include historical conversation records under the same online conversation program, so that before the chat robot generates a conversation, such as step S411, in addition to considering the semantic features of the user in the conversation, the user's preferences and real-time environmental information, it can also consider the user's historical conversation records in this online conversation program (step S417), so that the conversation content generated by the natural language model (step S311) is a conversation content that is in line with the current situation.

[0064] For example, the historical conversation records of users in the same online conversation program often contain the current context and can reflect the user's current emotions and needs. Figure 1 In the embodiment shown, the natural language processing model 101 in the cloud server 100 can use the natural language model to simultaneously consider the user's meaning, user preferences, real-time environmental information, and historical conversation records, and when generating conversation content, the same conversation context can be continued. For example, the same conversation topic can be continued, and when generating a conversation in natural language, the language can have the same tone as the previous conversation content (reflecting the user's emotions: joy, anger, sadness, happiness, etc.), so that the chat robot can learn the user's emotional expression through historical conversation records.

[0065] Related illustrations can be found in Fig.10 , Fig.10 The dialogue interface 1000 shows several dialogue contents 1001, 1002, and 1003 between the user and the chat robot, and the chat robot can also query the database based on the user semantic features obtained from the dialogue content 1002 to provide recommended audio and video content 1004. An input field 1005 is provided below the dialogue interface 1000 for the user to further input the dialogue content.

[0066] Another mode is, Fig.11The dialogue interface 1110 shown in this example shows that when the online dialogue program is started, the system directly provides natural language dialogue content 1111, 1112, 1114 based on the user's preferences and real-time information, and directly provides recommended audio and video content 1113. The user can then use the input field 1115 in the dialogue interface 1110 to respond to the above dialogue content.

[0067] According to another embodiment, the natural language model running in the chatbot generates the conversation content based on the semantic features of the content input by the user, the user's preferences and the real-time environment information, and may also include providing multiple recommendation options, multiple recommended audio and video content, and / or multiple recommended friend links. Fig.12 Display example.

[0068] In the online chat program, Fig.12 The dialogue interface 1200 shown includes a chat robot generating dialogue content 1201 based on the semantic features of the user, and when the program processing natural language in the system determines that the semantic features of a certain dialogue involve the content of a certain field, the system can import the corresponding field chat robot from the multi-field robot database in real time, and start a dialogue with the user to provide relevant information in the field until the next judgment is to return to the main chat robot or import another field chat robot. In the subsequent dialogue, the semantics in this example allow the chat robot to determine that the user is choosing a specific matter, so several recommended options 1202 are provided. And in particular, the chat robot provides the user with recommended options based on the real-time environmental information obtained by the system from the external system. For example, the chat robot can provide recommended options 1202 based on the real-time climate, road conditions, time, and the user's location. If the time is just at meal time, and referring to the user's eating habits, meal options can be provided based on restaurants that are open near the user's location.

[0069] Accordingly, if the user expresses a desire to watch video content, the recommendation options 1202 may be a plurality of recommended video content; if the user expresses a desire to find friends with similar interests, the recommendation options 1202 may be a plurality of recommended friend links.

[0070] Furthermore, the user then uses the input field 1206 to respond to the recommended options 1202 and inputs the dialogue content 1203, so that the chatbot responds to the dialogue content 1204 according to the semantics of the dialogue content 1203, and proposes a plurality of recommended contents 1205 according to the semantics of the above dialogue content. Continuing with the above example, when the user responds that he wants one of the meals, the chatbot can provide restaurant options corresponding to the meal the user wants to eat based on the real-time weather, road conditions and user location obtained by the system from the external system. If the weather is bad and there is a traffic jam on the road, the chatbot should recommend restaurant options that are convenient for the user to go to.

[0071] According to the above embodiments, it can be seen that, unlike the current generative artificial intelligence-implemented chat robot technology that can only respond to user questions with information obtained through past data training, the chat robot implemented by the intelligent dialogue introduction method proposed in the public book can generate dialogue content that meets user needs based on semantic features analyzed from user dialogues, user preferences obtained through system learning of user activity data, and real-time environmental information. Furthermore, it can also import domain robots based on needs or currently met conditions (any or any combination of semantic features, user preferences and real-time environmental information) to provide dialogue services trained for specific fields.

[0072] Next reference Figure 5 The flowchart of another embodiment of the intelligent dialogue import method shown in the figure, in which the first chat robot or the second chat robot is an automatic dialogue program executed in an online dialogue program, and may refer to the main chat robot and any one of the multiple field chat robots provided in the multi-field chat robot database provided by the system.

[0073] exist Figure 5 In the display process, the user starts the online dialogue program through the application and imports the first chat robot. According to the embodiment, the system default main chat robot can be imported at the beginning of the online dialogue program (step S501). The first chat robot can actively generate dialogue content, or wait for the user to input content and then receive the dialogue content input by the user (step S503) to further obtain the user's semantic features. According to the embodiment, the cloud server can use the natural language processing module to perform transformation calculation (transformer) and vector operation (vector operation) to obtain semantic features (step S505).

[0074] It is worth mentioning that according to the embodiment of the intelligent dialogue introduction method proposed in the open book, artificial intelligence technology is used to learn natural language, and after natural language understanding, text classification and grammar analysis are performed. When processing the dialogue content input by the user, a transformation model (transformer model, which was developed by Google in 2017) can be used. TM The deep learning method proposed by the Brain Team processes the natural language content input by users in a time sequence. If the input content is not text, it needs to be converted into text first. In this way, in the online dialogue program, this conversion model can be used to perform machine translation, document summarization, document generation, etc.

[0075] After obtaining the semantic features of the user's conversation content, the system combines the user's preferences and the user's current location obtained by the system or analyzes the location of interest to the user from the conversation content, and obtains real-time environmental information from an external system in real time based on this location (step S507). The real-time environmental information may include one or any combination of real-time weather, real-time traffic conditions, real-time news, and real-time location-related network information (such as POI on a map, POI evaluation, etc.) obtained from one or more external systems in real time.

[0076] At this time, the system runs the software program that imports the chat robot (interface Figure 3 The robot import interface 30 in the multi-domain robot database shown will determine whether to import the second chat robot (step S509) based on any one or any combination of user semantic features, user preferences and real-time environmental information. The second chat robot is a software program for importing the chat robot, which determines a domain chat robot for a specific domain based on any one or any combination of user semantics, user preferences and real-time environmental information.

[0077] In the above steps, if it is determined that the second chat robot does not need to be introduced (No), it means that the user is still communicating with the original chat robot (such as the preset main chat robot or any field chat robot), and the system will use the vector database to calculate the closest answer based on the user's semantic features, user preferences and real-time environmental information, or add historical dialogue records (step S511). It should be mentioned here that the data in the vector database is structured information obtained by using a vector algorithm, so that the system can obtain words with similar semantics from the obtained content based on the vector distance. Here is an example, the vector distance between the two words "computer" appearing in the dialogue content and the two words "calculation" in the database is closer, while the vector distance between "computer" and "running" is farther.

[0078] In this embodiment, the content input by the user, the content that the user is interested in, and the real-time environment information, and the historical conversation records can be added as needed, and the obtained text is annotated after executing the vector algorithm, and the vector of each word is calculated. The content with relevance is obtained according to the vector distance between the words, and the conversation content that meets the user's preferences and real-time environment information is generated accordingly. Further, according to the embodiment, when the vector algorithm is executed on the historical conversation records recorded in the cloud server, the conversation content that can meet the user's current emotions can be generated, such as continuing the same topic in the historical conversation records, and using words that are equivalent to the emotions obtained by analysis.

[0079] Furthermore, the system also searches the audio-visual database according to the above information to obtain the corresponding audio-visual content (step S513), and the main chat robot or the domain chat robot generates the dialogue content by natural language processing and generative artificial intelligence technology (step S515), and outputs the dialogue content on the dialogue interface (step S517). In addition, in one embodiment, during the chat process, the system will continue the above steps, so that the chat robot can communicate with the user through natural language (text or sound), and provide the user with interesting and real-time content (audio, text).

[0080] On the other hand, in the judgment of step S509, if it is judged that the current condition should import a domain chat robot of a specific domain (yes), the domain chat robot can be imported through the robot import interface in the multi-domain robot database (step S519), and the domain chat robot generates a dialogue through the natural language model running therein, and conducts a dialogue with the user in a specific domain (step S519). Similarly, the system obtains the semantic features in the dialogue through the natural language processing module (step S523), and can continue to obtain real-time environmental information, and execute the steps of generating dialogue content, such as steps S511 to S517. The process will repeatedly execute steps such as step S503 to generate dialogue content according to semantic features, user preferences and real-time environmental information obtained in real time, and determine whether to import other domain chat robots.

[0081] Related applications can be referred to first Fig.13 The user interface embodiment diagram of the related application of importing domain chatbots is shown in the figure, in which a domain chatbot provided by a manufacturer is taken as an example. When the system determines from the content of the conversation between the user and the chatbot that the topic involves the service or product provided by a certain manufacturer, it actively imports the domain chatbot of the specific manufacturer. At this time, the dialogue interface 1300 shown in the figure can display the manufacturer logo 1301. The domain chatbot generates the content of the conversation with the user using a natural language model, such as the dialogue content 1303 and 1305 in the figure, and can query the audio and video database (such as Figure 1 The recommended audio and video content can be introduced after the audio and video database 110 is added, and can be embedded in the dialogue interface 1300, such as the recommended product video 1304 in the icon, and the user can respond to the dialogue content 1306.

[0082] Fig.14 Another graphical user interface embodiment diagram for importing recommended manufacturers is displayed. This diagram shows that during the conversation, the domain chat robot determines the conditions that meet the services or products provided by a specific manufacturer based on the user's meaning, user preferences and user location, and then displays one or more localized recommended contents of the geographical location 1401 related to the user's location on the conversation interface 1400, such as the recommended manufacturer 1 1403 and the recommended manufacturer 2 1405 associated with the geographical location 1401. In this way, the specific domain chat robot can provide conversation contents that meet the user's needs in a specific domain.

[0083] Another embodiment can refer to Fig.15The illustrated diagram is a graphical user interface implementation example of an enterprise page. This diagram shows that the intelligent dialogue import system allows users to create their own homepages through user-side applications, such as the enterprise page 1500 shown in the figure. It allows enterprise users to create content and provides management functions, such as customizable domain chat robots. For example, a domain robot link 1501 is provided on the enterprise page 1500.

[0084] According to the embodiment, the intelligent dialogue import system provides enterprise users with the ability to build domain chatbots that meet the needs of the enterprise, that is, to use the machine learning module (such as Figure 1 The machine learning module 103 is displayed to learn the content provided by the enterprise, so that the trained domain chat robot can provide users with conversation content about the enterprise's services or products in natural language. The method of establishing a domain chat robot can refer to Figure 6 The process embodiment diagram shown.

[0085] In the intelligent dialogue import system, enterprise users are domain customers and can first join a specific social media (step S601), and can establish Fig.15 The displayed enterprise page 1500 allows domain customers to upload audio and video content and descriptions to create their own homepage (step S603). In the method of establishing a domain chatbot belonging to a domain customer, a machine learning algorithm provided by the system is used to learn the data of various fields provided by the domain customer. After learning the characteristics of each field (step S605), a domain chatbot belonging to the domain customer is established using natural language processing technology and generative artificial intelligence technology (step S607).

[0086] Furthermore, for commercial users, the system can provide paid services, allowing domain customers to set conditions for using domain chat robots, such as setting the number of conversations and budgets per day, week, or month (step S609). Fig.17 The diagram of a graphical user interface embodiment of a budget setting page is shown, in which a budget setting page 1700 is displayed on an enterprise user-side application, wherein various charging plans are provided for reference and setting by field users.

[0087] When the domain customer completes the establishment of his own domain chat robot and related settings, Figure 1 The multi-domain robot database 140 provided by the cloud server 100 is shown as establishing a domain chat robot established by each domain customer. When the system executes the intelligent dialogue import method, the software program executing the intelligent dialogue import method analyzes the user's meaning and can import the corresponding domain chat robot into the online dialogue program according to the user's needs (step S611).

[0088] Here, a domain chatbot in the fitness category is listed as an example. When a user talks to the main chatbot or a specific domain chatbot, the software program that executes the intelligent dialogue import method can continuously analyze the user's semantic features. When it is determined that the user's semantics are about fitness-related topics, such as the user expressing the body shape he wants to exercise in the conversation, the software program that executes the intelligent dialogue import method can import the corresponding domain chatbot. The domain chatbot can understand the user's needs through the conversation, so that it can provide fitness courses and equipment that the company can provide, and plan a fitness plan that suits the user's age, time, location and current situation.

[0089] Furthermore, the intelligent dialogue import system can provide the usage status of the domain chatbot when serving customers in various fields. Figure 7 The process diagram for generating a domain customer report is shown in the figure, and you can also refer to Fig.16 A diagram showing an embodiment of a graphical user interface of a displayed statistics report page.

[0090] When the intelligent dialogue import system runs each domain chatbot, statistical data can be established for each domain chatbot, and domain customers can also evaluate the performance through the statistical data established in the system. In the intelligent dialogue import system, the software program in the cloud server counts the number of times the domain chatbot is imported, the running time, the number of dialogues (step S701), and the number of clicks and time of the recommended content of the domain chatbot (step S703), or the number of new followers (step S705) and the number of customer data views (step S707) can be obtained, and a domain customer report can be provided so that the domain customer can evaluate the performance of the domain chatbot accordingly (step S709).

[0091] For reference Fig.16 The embodiment of the graphical user interface of the statistical report page is shown, wherein the statistical report page 1600 provided to the domain customer is shown. The embodiment shows that the domain customer can have a dialogue with the chat robot provided by the system, such as the schematically shown dialogue content 1601, 1603, and ask the system to provide a report, such as the statistical report 1602 responded by the system. It is mentioned here that in addition to providing the domain customers with common statistical reports, the domain customers can also obtain their own domain chat robot statistical data through dialogue in the online dialogue program, and during the dialogue process, the system continues to provide recommended activities 1604 through the chat robot according to the user's semantics, preferences and real-time environmental information.

[0092] In summary, according to the embodiments of the above-mentioned intelligent dialogue import method and system, the cloud server proposed by the system executes the intelligent dialogue import method to provide natural language dialogue services, and also refers to the user preferences obtained through deep learning and obtains real-time environmental information during the dialogue process with the user, so as to realize a chat robot and a dialogue program that can adapt to user preferences and real-time information.

[0093] The contents disclosed above are only preferred feasible embodiments of the present invention, and are not intended to limit the claims of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention's specification and drawings are included in the claims of the present invention.

Claims

1. An intelligent dialogue import method, executed in a cloud server, Features The method includes: Starting an online conversation program, introducing a first chat robot, and receiving content input by a user through a conversation interface; Obtaining semantic features of the content input by the user; Obtain user data and real-time environmental information; Importing a second chat robot according to any one or any combination of semantic features of the content input by the user, user preferences derived from the user data, and the real-time environment information; Generate a conversation content through a natural language model run by the second chatbot; and The conversation content is imported into the online conversation program, and the conversation content is outputted on the conversation interface.

2. The intelligent dialogue introduction method according to claim 1, Features In addition to generating the dialogue content, the first chat robot or the second chat robot also introduces one or more audio and video content into the dialogue interface according to the user's semantic features, the user's preferences and the real-time environmental information.

3. The intelligent dialogue introduction method according to claim 2, It is characterized in that Furthermore, one or more location-appropriate recommended contents related to the user's location are displayed in the dialogue interface.

4. The intelligent dialogue introduction method according to claim 1, Features The cloud server runs a social media and executes a corresponding social media application in a user device, so that the user clicks a conversation link pattern on a page in the social media application to enter the online conversation program.

5. The intelligent dialogue introduction method according to claim 4, Features The page is a map interface, on which are marked a plurality of link points associated with geographic locations, including one or more audio and video link points.

6. The intelligent dialogue introduction method according to claim 1, It is characterized in that The natural language model running in the cloud server generates the conversation content by using a conversion model to perform machine translation, document summarization and document generation.

7. The intelligent dialogue introduction method according to claim 6, It is characterized in that In the cloud server, a vector algorithm is further executed on the content input by the user, the user's preferences and the real-time environmental information to annotate the obtained text, calculate the vector of each word, obtain related content based on the vector distance between words, and generate the conversation content that conforms to the user's preferences and the real-time environmental information.

8. The intelligent dialogue introduction method according to claim 1, Features The real-time environmental information includes one or any combination of real-time weather, real-time traffic conditions, real-time news and real-time location-related network information obtained from one or more external systems in real time.

9. The intelligent dialogue introduction method according to any one of claims 1 to 8, Features The first chat robot is a default main chat robot in the online dialogue program, and the second chat robot is a domain chat robot determined by the software program that executes the imported chat robot in the cloud server based on the user's semantics, the user's preferences and any one or any combination of the real-time environment information.

10. The intelligent dialogue introduction method according to claim 9, Features The cloud server provides a multi-domain robot database, and the domain chat robot is provided by the multi-domain robot database.

11. The intelligent dialogue introduction method according to claim 10, Features The multi-domain robot database includes multiple domain models trained by learning various domain expertise through machine learning algorithms. The multiple domain models use natural language processing technology and generative artificial intelligence technology to realize multiple domain chat robots with natural language processing capabilities.

12. An intelligent dialogue import system, Features The system comprises: A cloud server uses a processing circuit to execute an intelligent dialogue introduction method, including: Starting an online conversation program, introducing a first chat robot, and receiving content input by a user through a conversation interface; Obtaining semantic features of the content input by the user; Obtain user data and real-time environmental information; Importing a second chat robot according to any one or any combination of semantic features of the content input by the user, user preferences derived from the user data, and the real-time environment information; Generate a conversation content through a natural language model run by the second chatbot; and The conversation content is imported into the online conversation program, and the conversation content is outputted on the conversation interface.

13. The intelligent dialogue introduction system according to claim 12, Features The cloud server provides an external system interface for connecting to one or more external systems to obtain the real-time environmental information, including one or any combination of real-time weather, real-time traffic conditions, real-time news and real-time location-related network information.

14. The intelligent dialogue introduction system according to claim 12, It is characterized in that The natural language model running in the cloud server generates the conversation content by using a conversion model to perform machine translation, document summarization and document generation; A vector algorithm is further executed on the content input by the user, the user's preferences and the real-time environmental information to annotate the obtained text, calculate the vector of each word, obtain related content based on the vector distance between words, and generate the conversation content that conforms to the user's preferences and the real-time environmental information.

15. The intelligent dialogue introduction system according to claim 12, Features The cloud server runs a social media and executes a corresponding social media application in a user device, so that the user clicks a conversation link pattern on a page in the social media application to enter the online conversation program.

16. The intelligent dialogue introduction system according to claim 15, Features The page is a map interface, on which are marked a plurality of link points associated with geographic locations, including one or more audio and video link points.

17. The intelligent dialogue introduction system according to any one of claims 12 to 16, Features The first chat robot is a default main chat robot in the online dialogue program, and the second chat robot is a domain chat robot determined by the software program that executes the imported chat robot in the cloud server based on the user's semantics, the user's preferences and any one or any combination of the real-time environment information.

18. The intelligent dialogue introduction system according to claim 17, Features The cloud server provides a multi-domain robot database, and the domain chat robot is provided by the multi-domain robot database; the multi-domain robot database includes multiple domain models trained by learning various domain expertise through machine learning algorithms, and the multiple domain models use natural language processing technology and generative artificial intelligence technology to realize multiple domain chat robots with natural language processing capabilities.

19. The intelligent dialogue introduction system according to claim 17, It is characterized in that In the cloud server, statistics are collected on the number of times the chatbot in the field is imported, the running time, the number of conversations, and the number of clicks and time of the recommended content of the chatbot in the field are collected to provide a field customer report.

20. The intelligent dialogue introduction system according to claim 17, Features The cloud server provides an audio-visual database, and the domain chat robot queries the audio-visual database according to the user's semantic features, the user's preferences and the real-time environmental information to introduce one or more audio-visual contents into the dialogue interface.

Images