Interactive digital human driving method

Abstract

The invention discloses an interactive digital human driving method, which relates to the technical field of digital human interaction, and comprises the following steps of: capturing and recording interaction between a user and a digital human through an input mode of voice, character and touch screen instructions by a system, and preprocessing input data; identifying the intention and emotion of the user by using a natural language processing technology, and determining the demand of the user; generating a response according to the identified intention and context, and judging the demand and emotional state of the user; generating response content, feeding back the generated response to the user, and displaying the response on the interface; and updating the conversation state and the context. According to the method, through comprehensive sentiment analysis, intention-based features, classification model establishment and user behavior prediction scoring, the next operation of the user can be accurately predicted, so that the intelligence of the digital person in the use process is improved, the recommendation efficiency of the digital person is higher, and better use prospects and commercial values are brought.

Description

Technical Field

[0001] The present invention relates to the technical field of digital human interaction, and in particular to an interactive digital human driving method. Background Art

[0002] Digital human interaction refers to the interaction with virtual or artificial intelligence-driven characters in a digital way. These digital humans usually exist in the form of virtual images, 3D models, voice assistants or chat robots, which can simulate human conversations and behaviors. Human-computer interaction technology refers to the technology of realizing human-computer dialogue in an effective way through computer input and output devices. Human-computer interaction technology includes machines providing people with a large amount of relevant information and prompts through output or display devices, and people inputting relevant information to machines through input devices, answering questions and prompts, etc. Human-computer interaction technology is one of the important contents in computer user interface design. Human-computer interaction technology refers to the technology of realizing human-computer dialogue in an effective way through computer input and output devices. Human-computer interaction technology includes machines providing people with a large amount of relevant information and prompts through output or display devices, and people inputting relevant information to machines through input devices, answering questions and prompts, etc. Human-computer interaction technology is one of the important contents in computer user interface design. With the development of society, the demand for digital humans is increasing.

[0003] Existing digital humans are not comprehensive enough when interacting and cannot make predictions and recommendations based on multiple factors, which makes the digital humans' recommendations inefficient. Summary of the invention

[0004] In order to solve the above technical problems, an interactive digital human driving method is provided. This technical solution solves the above problems.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] An interactive digital human driving method, comprising:

[0007] S1: The system captures and records the user's interaction with the digital human through voice, text and touch screen commands, and pre-processes the input data;

[0008] S2: Use natural language processing technology to identify user intentions and emotions and determine user needs;

[0009] S3: Generate responses based on the identified intent and context, judging the user’s needs and emotional state;

[0010] S4: Generate response content, feed back the generated response to the user, and display it on the interface;

[0011] S5: Update the conversation state and context.

[0012] Preferably, S1 specifically includes:

[0013] Receive user's voice, text and touch screen input;

[0014] Use sensors and cameras as input methods, and use weighted average to combine gesture recognition with touch screen input;

[0015] The combined formula is: F 总 =αF1+(1-α)F2

[0016] Among them, F 总 is the comprehensive result of touch input and gesture recognition data, α is the fusion weight, 1-α is the weight of gesture feature in fusion, F1 is the touch input feature, and F2 is the gesture feature;

[0017] Clean and transform input data;

[0018] Automatically correct spelling and grammatical errors in user input using natural language techniques;

[0019] Using advanced data indexing and storage techniques, the processed data is stored in the database.

[0020] Preferably, S2 specifically includes:

[0021] Extract features from text that help understand intent, including keywords, phrases, and contextual information;

[0022] Based on the features of intent, a classification algorithm is used to classify intent;

[0023] Extract emotion-related features from text, including emotion vocabulary and emotion intensity;

[0024] Based on the emotional characteristics, a sentiment analysis model is used to judge the emotional tendency of the text, which includes positive, negative and neutral;

[0025] The sentiment analysis model formula is:

[0026] Where T is the text data to be analyzed, S(T) is the final sentiment classification result of text T, P is the threshold for determining that the text is classified as positive sentiment, and N is the threshold for determining that the text is classified as negative sentiment;

[0027] Based on sentiment analysis models, the intent and emotions are identified and the specific needs of users are inferred.

[0028] Preferably, based on the features of the intent, using a classification algorithm to classify the intent includes:

[0029] Build a classification model based on the features of intent;

[0030] The classification model expression is:

[0031] In the formula, w is the weight vector, x is the feature vector, and w t x is the dot product of the weight vector and the feature vector, b is the bias term, and z is the result of the linear combination. is the predicted probability value, e is the natural logarithm, the probability that the sample belongs to category 1, e is the natural logarithm, σ(z) is the logistic function, and P(Y=1x) represents the probability that the sample belongs to category 1 given feature x;

[0032] Classification is performed based on the output of the model.

[0033] Preferably, S3 specifically includes:

[0034] Get the user's previous conversation records;

[0035] Perform analysis based on background data of the user's personal information;

[0036] Based on the calculated sentiment analysis results and classification results, as well as the data analysis results, the identified intentions and emotions, the specific needs of the users are inferred;

[0037] Based on the specific needs of users, determine the urgency and priority of user needs;

[0038] Select response templates based on users’ specific needs and priorities;

[0039] Adapt responses based on contextual information to ensure personalization and relevance;

[0040] Use natural language generation techniques to generate natural and contextual responses;

[0041] Adjust the tone of the response based on the user’s emotional state, including using a soothing and understanding tone for users with negative emotions and an encouraging and positive tone for users with positive emotions;

[0042] Providing emotional support, including apologies, thanks, and comfort;

[0043] Set up user feedback testing to evaluate the effectiveness of your responses.

[0044] Preferably, analyzing the background data based on the user's personal information includes:

[0045] Based on the user's historical usage data, predict and score the user's behavior;

[0046] The prediction score formula is:

[0047]

[0048] in, is the predicted score, α and β are the weight coefficients of C and B respectively, C is the predicted score of the collaborative filtering method, and B is the predicted score of the content-based method;

[0049] Based on the predicted score, determine the user's next action.

[0050] Preferably, S4 specifically includes:

[0051] Use the model to generate response content;

[0052] Format the generated response content into a suitable HTML data structure;

[0053] Embed dynamic data into responses;

[0054] Transmit the formatted response to the front-end application through the API;

[0055] The front end receives the response and renders the content on the user interface.

[0056] Preferably, the front end receives the response and renders the content on the user interface including:

[0057] Select front-end components and elements based on the response content, including text boxes, dialog boxes, lists, and forms; pass the processed data to the front-end components for rendering, including:

[0058] Insert text content into appropriate HTML elements;

[0059] Set the src attribute of the image to display the picture;

[0060] Update dynamic data and apply animation and dynamic effects;

[0061] Bind user interaction events to corresponding front-end components to respond to user operations.

[0062] Preferably, S5 specifically includes:

[0063] Capture new messages and instructions from users;

[0064] Analyze user input to identify key information and intent;

[0065] Store user input and related data into the current dialog state;

[0066] Record user actions, contextual information, and system responses;

[0067] Adapt context based on input, including adding or modifying session history and user preferences;

[0068] Save the updated dialog state and context information to a database or memory;

[0069] Generate appropriate response content using the updated context.

[0070] Preferably, generating appropriate response content using the updated context includes:

[0071] Get the updated context;

[0072] Extract the latest contextual information from the conversation state, including the user’s previous messages, emotional state, intent, and classification results;

[0073] Format the generated content, including paragraphing, bullet points, and use of visual elements;

[0074] Perform syntax checking and language optimization;

[0075] Transmit the formatted response to the front-end application via the API or directly;

[0076] Display the generated response content on the user interface;

[0077] Record generated responses and user feedback;

[0078] Collect feedback from users.

[0079] Compared with the prior art, the present invention has the following beneficial effects:

[0080] The present invention proposes comprehensive sentiment analysis, establishes a classification model based on the characteristics of intention, and predicts and scores the user's behavior, which can accurately predict the user's next operation, thereby improving the intelligence of the digital human during use and making the digital human's recommendation efficiency higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 It is a flow chart of the steps of the present invention. DETAILED DESCRIPTION

[0082] In order to better understand the present application, a more detailed description will be made of various aspects of the present application with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present application, and are not intended to limit the scope of the present application in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0083] In the accompanying drawings, the size, dimensions and shapes of the elements have been slightly adjusted for ease of illustration. The drawings are for illustration only and are not strictly drawn to scale. As used herein, the terms "substantially", "approximately" and similar terms are used as terms of approximation, not as terms of degree, and are intended to illustrate the inherent deviations in measurements or calculations that would be recognized by a person of ordinary skill in the art. In addition, in this application, the order in which the processing of each step is described does not necessarily represent the order in which these processes occur in actual operation, unless otherwise specified or can be derived from the context.

[0084] It should also be understood that expressions such as "include", "including", "have", "contain" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present application, "may" is used to mean "one or more embodiments of the present application". And, the term "exemplary" is intended to refer to an example or illustration.

[0085] Unless otherwise defined, all words (including engineering terms and scientific and technological terms) used in this article have the same meaning as commonly understood by ordinary technical citizens in the field to which this application belongs. It should also be understood that, unless explicitly stated in this application, words defined in commonly used dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0086] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0087] Reference Figure 1 As shown, an interactive digital human driving method includes:

[0088] S1: The system captures and records the user's interaction with the digital human through voice, text and touch screen commands, and pre-processes the input data;

[0089] S2: Use natural language processing technology to identify user intentions and emotions and determine user needs;

[0090] S3: Generate responses based on the identified intent and context, judging the user’s needs and emotional state;

[0091] S4: Generate response content, feed back the generated response to the user, and display it on the interface;

[0092] S5: Update the conversation state and context.

[0093] Furthermore, S1 specifically includes:

[0094] Receive user's voice, text and touch screen input;

[0095] Use sensors and cameras as input methods, and use weighted average to combine gesture recognition with touch screen input;

[0096] The combined formula is: F 总 =αF1+(1-α)F2

[0097] Among them, F 总 is the comprehensive result of touch input and gesture recognition data, α is the fusion weight, 1-α is the weight of gesture feature in fusion, F1 is the touch input feature, and F2 is the gesture feature;

[0098] Clean and transform input data;

[0099] Automatically correct spelling and grammatical errors in user input using natural language techniques;

[0100] Using advanced data indexing and storage techniques, the processed data is stored in the database.

[0101] Furthermore, S2 specifically includes:

[0102] Extract features from text that help understand intent, including keywords, phrases, and contextual information;

[0103] Based on the features of intent, a classification algorithm is used to classify intent;

[0104] Extract emotion-related features from text, including emotion vocabulary and emotion intensity;

[0105] Based on the emotional characteristics, a sentiment analysis model is used to judge the emotional tendency of the text, which includes positive, negative and neutral;

[0106] The sentiment analysis model formula is:

[0107] Where T is the text data to be analyzed, S(T) is the final sentiment classification result of text T, P is the threshold for determining that the text is classified as positive sentiment, and N is the threshold for determining that the text is classified as negative sentiment;

[0108] Based on sentiment analysis models, the intent and emotions are identified and the specific needs of users are inferred.

[0109] Furthermore, based on the features of intent, the classification algorithm is used to classify intent, including:

[0110] Build a classification model based on the features of intent;

[0111] The classification model expression is:

[0112] In the formula, w is the weight vector, x is the feature vector, and w t x is the dot product of the weight vector and the feature vector, b is the bias term, and z is the result of the linear combination. is the predicted probability value, e is the natural logarithm, the probability that the sample belongs to category 1, e is the natural logarithm, σ(z) is the logistic function, and P(Y=1x) represents the probability that the sample belongs to category 1 given feature x;

[0113] Classification is performed based on the output of the model.

[0114] Furthermore, S3 specifically includes:

[0115] Get the user's previous conversation records;

[0116] Perform analysis based on background data of the user's personal information;

[0117] Based on the calculated sentiment analysis results and classification results, as well as the data analysis results, the identified intentions and emotions, the specific needs of the users are inferred;

[0118] Based on the specific needs of users, determine the urgency and priority of user needs;

[0119] Select response templates based on users’ specific needs and priorities;

[0120] Adapt responses based on contextual information to ensure personalization and relevance;

[0121] Use natural language generation techniques to generate natural and contextual responses;

[0122] Adjust the tone of the response based on the user’s emotional state, including using a soothing and understanding tone for users with negative emotions and an encouraging and positive tone for users with positive emotions;

[0123] Providing emotional support, including apologies, thanks, and comfort;

[0124] Set up user feedback testing to evaluate the effectiveness of your responses.

[0125] Further, the analysis based on the background data of the user's personal information includes:

[0126] Based on the user's historical usage data, predict and score the user's behavior;

[0127] The prediction score formula is:

[0128]

[0129] in, is the predicted score, α and β are the weight coefficients of C and B respectively, C is the predicted score of the collaborative filtering method, and B is the predicted score of the content-based method;

[0130] Based on the predicted score, determine the user's next action.

[0131] Furthermore, S4 specifically includes:

[0132] Use the model to generate response content;

[0133] Format the generated response content into a suitable HTML data structure;

[0134] Embed dynamic data into responses;

[0135] Transmit the formatted response to the front-end application through the API;

[0136] The front end receives the response and renders the content on the user interface.

[0137] Furthermore, the front end receives the response and renders the content on the user interface including:

[0138] Select front-end components and elements based on the response content, including text boxes, dialog boxes, lists, and forms; pass the processed data to the front-end components for rendering, including:

[0139] Insert text content into appropriate HTML elements;

[0140] Set the src attribute of the image to display the picture;

[0141] Update dynamic data and apply animation and dynamic effects;

[0142] Bind user interaction events to corresponding front-end components to respond to user operations.

[0143] Furthermore, S5 specifically includes:

[0144] Capture new messages and instructions from users;

[0145] Analyze user input to identify key information and intent;

[0146] Store user input and related data into the current dialog state;

[0147] Record user actions, contextual information, and system responses;

[0148] Adapt context based on input, including adding or modifying session history and user preferences;

[0149] Save the updated dialog state and context information to a database or memory;

[0150] Generate appropriate response content using the updated context.

[0151] Further, using the updated context to generate appropriate response content includes:

[0152] Get the updated context;

[0153] Extract the latest contextual information from the conversation state, including the user’s previous messages, emotional state, intent, and classification results;

[0154] Format the generated content, including paragraphing, bullet points, and use of visual elements;

[0155] Perform syntax checking and language optimization;

[0156] Transmit the formatted response to the front-end application via the API or directly;

[0157] Display the generated response content on the user interface;

[0158] Record generated responses and user feedback;

[0159] Collect feedback from users.

[0160] The use process of the present invention is:

[0161] Step 1: Receive user's voice, text and touch screen input;

[0162] Step 2: Use sensors and cameras as input methods, and use weighted average to combine gesture recognition with touch screen input;

[0163] Step 3: Clean and transform the input data;

[0164] Step 4: Use natural language technology to automatically correct spelling and grammatical errors in user input;

[0165] Step 5: Extract features from the text that help understand the intent, including keywords, phrases, and contextual information;

[0166] Step 6: Based on the features of intent, use a classification algorithm to classify intent;

[0167] Step 7: Extract emotion-related features from the text, including emotion vocabulary and emotion intensity;

[0168] Step 8: Based on the sentiment characteristics, use the sentiment analysis model to determine the sentiment tendency of the text;

[0169] Step 9: Get the user's previous conversation records;

[0170] Step 10: Analyze the background data based on the user's personal information;

[0171] Step 11: Based on the calculated sentiment analysis results and classification results, as well as the data analysis results, the identified intentions and emotions, infer the specific needs of the user;

[0172] Step 12: Use the model to generate response content;

[0173] Step 13: Save the updated dialog state and context information to the database or memory.

[0174] In summary, the advantages of the present invention are as follows: the present invention proposes comprehensive sentiment analysis, establishes a classification model based on the characteristics of intention, and predicts and scores the user's behavior, which can accurately predict the user's next operation, thereby improving the intelligence of the digital human during use and making the digital human's recommendation efficiency higher.

[0175] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. An interactive digital human driving method, characterized in that: The drive methods include: S1: The system captures and records the user's interaction with the digital human through voice, text and touch screen commands, and pre-processes the input data; S2: Use natural language processing technology to identify user intentions and emotions and determine user needs; S3: Generate responses based on the identified intent and context, judging the user’s needs and emotional state; S4: Generate response content, feed back the generated response to the user, and display it on the interface; S5: Update the conversation state and context.

2. The interactive digital human driving method according to claim 1, characterized in that: S1 specifically includes: Receive user's voice, text and touch screen input; Use sensors and cameras as input methods, and use weighted average to combine gesture recognition with touch screen input; The combined formula is: F 总 =αF1+(1-α)F2 Among them, F 总 is the comprehensive result of touch input and gesture recognition data, α is the fusion weight, 1-α is the weight of gesture feature in fusion, F1 is the touch input feature, and F2 is the gesture feature; Clean and transform input data; Automatically correct spelling and grammatical errors in user input using natural language techniques; Using advanced data indexing and storage techniques, the processed data is stored in the database.

3. The interactive digital human driving method according to claim 2, characterized in that: S2 specifically includes: Extract features from text that help understand intent, including keywords, phrases, and contextual information; Based on the features of intent, a classification algorithm is used to classify intent; Extract emotion-related features from text, including emotion vocabulary and emotion intensity; Based on the emotional characteristics, a sentiment analysis model is used to judge the emotional tendency of the text, which includes positive, negative and neutral; The sentiment analysis model formula is: Where T is the text data to be analyzed, S(T) is the final sentiment classification result of text T, P is the threshold for determining that the text is classified as positive sentiment, and N is the threshold for determining that the text is classified as negative sentiment; Based on sentiment analysis models, the intent and emotions are identified and the specific needs of users are inferred.

4. The interactive digital human driving method according to claim 3, characterized in that: Based on the features of intent, the classification algorithm used for intent classification includes: Build a classification model based on the features of intent; The classification model expression is: In the formula, w is the weight vector, x is the feature vector, and w t x is the dot product of the weight vector and the feature vector, b is the bias term, and z is the result of the linear combination. is the predicted probability value, e is the natural logarithm, the probability that the sample belongs to category 1, e is the natural logarithm, σ(z) is the logistic function, and P(Y=1x) represents the probability that the sample belongs to category 1 given feature x; Classification is performed based on the output of the model.

5. The interactive digital human driving method according to claim 4, characterized in that: S3 specifically includes: Get the user's previous conversation records; Perform analysis based on background data of the user's personal information; Based on the calculated sentiment analysis results and classification results, as well as the data analysis results, the identified intentions and emotions, the specific needs of the users are inferred; Based on the specific needs of users, determine the urgency and priority of user needs; Select response templates based on users’ specific needs and priorities; Adapt responses based on contextual information to ensure personalization and relevance; Use natural language generation techniques to generate natural and contextual responses; Adjust the tone of the response based on the user’s emotional state, including using a soothing and understanding tone for users with negative emotions and an encouraging and positive tone for users with positive emotions; Providing emotional support, including apologies, thanks, and comfort; Set up user feedback testing to evaluate the effectiveness of your responses.

6. The interactive digital human driving method according to claim 5, characterized in that: Analysis based on the background data of the user's personal information includes: Based on the user's historical usage data, predict and score the user's behavior; The prediction score formula is: in, is the predicted score, α and β are the weight coefficients of C and B respectively, C is the predicted score of the collaborative filtering method, and B is the predicted score of the content-based method; Based on the predicted score, determine the user's next action.

7. The interactive digital human driving method according to claim 6, characterized in that: S4 specifically includes: Use the model to generate response content; Format the generated response content into a suitable HTML data structure; Embed dynamic data into responses; Transmit the formatted response to the front-end application through the API; The front end receives the response and renders the content on the user interface.

8. The interactive digital human driving method according to claim 7, characterized in that: The front end receives the response and renders the content on the user interface including: Select front-end components and elements based on the response content, including text boxes, dialog boxes, lists, and forms; pass the processed data to the front-end components for rendering, including: Insert text content into appropriate HTML elements; Set the src attribute of the image to display the image; Update dynamic data and apply animation and dynamic effects; Bind user interaction events to corresponding front-end components to respond to user operations.

9. The interactive digital human driving method according to claim 8, characterized in that: S5 specifically includes: Capture new messages and instructions from users; Analyze user input to identify key information and intent; Store user input and related data into the current dialog state; Record user actions, contextual information, and system responses; Adapt context based on input, including adding or modifying session history and user preferences; Save the updated dialog state and context information to a database or memory; Generate appropriate response content using the updated context.

10. The interactive digital human driving method according to claim 9, characterized in that: Generating an appropriate response using the updated context includes: Get the updated context; Extract the latest contextual information from the conversation state, including the user’s previous messages, emotional state, intent, and classification results; Format the generated content, including paragraphing, bullet points, and use of visual elements; Perform syntax checking and language optimization; Transmit the formatted response to the front-end application via the API or directly; Display the generated response content on the user interface; Record generated responses and user feedback; Collect feedback from users.

Images