Cloud conference virtual response method and system based on artificial intelligence

Through AI semantic analysis and logic optimization, the response efficiency and accuracy of cloud conferencing systems are improved, the problem of insufficient response mechanism in the existing technology is solved, the smoother and smarter user experience is provided, and the system upgrade and function expansion is supported. .

CN117951262BActive Publication Date: 2024-11-19BEIJING UNITED MEDIA TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311676831.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-07
Publication Date
2024-11-19
Estimated Expiration
2043-12-07

AI Technical Summary

Technical Problem

The existing cloud conferencing system has insufficient response efficiency and accuracy in response mechanism, which cannot meet the needs of user personalized and efficient interaction, and lacks an effective optimization mechanism, which affects the user experience and the expansion of system functions.

Method used

Using the cloud conference virtual response method based on artificial intelligence, we generate response logic sequences through AI semantic analysis, determine the logical processing unit to be optimized and the response logic implantation point map, optimize the adjustable control node, generate customized response control nodes, and enhance the logical processing unit. , improve response efficiency and accuracy.

Benefits of technology

It realizes the efficiency and accuracy of the response of the cloud conference system, meets the user's personalized interaction needs, improves the flexibility and robustness of the system, provides users with a smooth and intelligent meeting experience, and lays the foundation for system upgrades and function expansion. The foundation is.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117951262B_ABST
    Figure CN117951262B_ABST
Patent Text Reader

Abstract

The embodiment of the present application provides a cloud conference virtual response method and system based on artificial intelligence. By obtaining the optimized cloud conference interactive system framework and the preset automatic response script library, AI semantic analysis is performed on the automatic response script library to generate a response logic sequence. Subsequently, the logic processing unit to be optimized is identified from the cloud conference interactive system framework, and the response logic implantation point map is determined based on the response logic sequence and the corresponding logic processing parameter sequence. Through the response logic implantation point map, the adjustable control nodes associated with each response instruction are determined, and optimization operations are performed on these adjustable control nodes to create customized response control nodes. Finally, these customized response control nodes are used to generate an enhanced version of the logic processing unit, and the cloud conference interactive system framework is updated accordingly to enhance its response mechanism.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of cloud conference technology, and in particular to an artificial intelligence-based cloud conference virtual response method and system. Background Art

[0002] With the rapid development of information technology and network communication technology, cloud conference systems, as a new remote communication method, have been widely used in business, education and government departments. Cloud conference systems can transcend geographical restrictions, realize real-time audio and video communication and data sharing, and greatly improve work efficiency and communication convenience. However, with the continuous growth and diversification of user needs, existing cloud conference systems face a series of challenges in response mechanisms.

[0003] First, traditional cloud conferencing systems often use static logic processing units when processing user requests, and cannot be dynamically adjusted according to different scenarios and user habits. This limits the response efficiency and accuracy of the cloud conferencing system and cannot meet the user's needs for personalization and efficient interaction.

[0004] Secondly, due to the lack of an effective optimization mechanism, the response logic processing unit of the cloud conference system is difficult to make fine adjustments according to actual usage, which not only affects the user experience, but also limits the expansion and upgrade of the cloud conference system functions. Summary of the invention

[0005] In order to at least overcome the above-mentioned deficiencies in the prior art, the purpose of the embodiments of the present application is to provide a cloud conference virtual answering method and system based on artificial intelligence.

[0006] Obtaining an optimized cloud conference interactive system framework and a preset automatic answering script library, performing AI semantic analysis on the automatic answering script library, and generating an answering logic sequence;

[0007] Determine the logic processing unit to be optimized from the cloud conference interactive system framework, and determine the response logic implantation point map according to the response logic sequence and the logic processing parameter sequence corresponding to the logic processing unit to be optimized;

[0008] Determine the adjustable control nodes associated with each response instruction in the response logic sequence from the logic processing parameter sequence according to the response logic implantation point map, and perform optimization operations on the corresponding adjustable control nodes according to each response instruction to generate each customized response control node, wherein the positions of the adjustable control nodes determined by the response logic implantation point map are dispersed;

[0009] Based on the various customized response control nodes, an enhanced logic processing unit corresponding to the logic processing unit to be optimized is generated, and based on the enhanced logic processing unit, a cloud conference interactive system framework with an enhanced cloud conference response mechanism is generated. The cloud conference interactive system framework with an enhanced cloud conference response mechanism is used to interactively process any cloud conference request data to be responded to.

[0010] According to one aspect of an embodiment of the present application, a cloud conferencing system is provided, which includes a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores machine-executable instructions, and the machine-executable instructions are loaded and executed by the processor to implement an artificial intelligence-based cloud conferencing virtual response method in any one of the aforementioned possible implementations.

[0011] According to one aspect of the embodiments of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in various optional implementations of the above three aspects.

[0012] In the technical solutions provided in some embodiments of the present application, first, the optimized cloud conference interactive system framework and the preset automatic answering script library are obtained, and the automatic answering script library is subjected to AI semantic analysis to generate an answering logic sequence. Subsequently, the logic processing unit to be optimized is identified from the cloud conference interactive system framework, and the answering logic implantation point map is determined based on the answering logic sequence and the corresponding logic processing parameter sequence. Through the answering logic implantation point map, the adjustable control nodes associated with each answering instruction are determined, and the optimization operation is performed on these adjustable control nodes to create customized answering control nodes. Finally, these customized answering control nodes are used to generate an enhanced version of the logic processing unit, and the cloud conference interactive system framework is updated accordingly to enhance its answering mechanism. Thus, targeted optimization strategies can be implemented to effectively improve the answering efficiency and accuracy of the cloud conference interactive system framework, and meet the needs of user personalization and efficient interaction. The present application allows the cloud conference interactive system framework to accurately identify and adjust key control nodes according to the preset answering logic sequence, ensuring that each answering instruction can obtain the best response effect. Since the positions of the adjustable control nodes determined by the response logic implantation point map are decentralized, the optimization operation can be performed independently at multiple key points of the cloud conference interactive system framework, thus avoiding the chain effects and unstable factors that may be caused by centralized adjustments, significantly improving the flexibility and robustness of the cloud conference interactive system framework, and providing users with a smoother and smarter conference experience. In addition, the enhanced logic processing unit can be flexibly integrated into the existing cloud conference interactive system framework, laying the foundation for future system upgrades and function expansions. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be extracted in combination with these drawings without creative work.

[0014] Figure 1 A flowchart of a cloud conference virtual response method based on artificial intelligence provided in an embodiment of the present application;

[0015] Figure 2 A schematic block diagram of the structure of a cloud conference system for implementing the above-mentioned artificial intelligence-based cloud conference virtual response method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] The following description is intended to enable one of ordinary skill in the art to implement and combine the present application, and the description is provided in the context of a specific application scenario and its requirements. It will be apparent to one of ordinary skill in the art that various changes may be made to the disclosed embodiments, and that the general principles defined in the present application may be applied to other embodiments and application scenarios without departing from the principles and scope of the present application. Therefore, the present application is not limited to the described embodiments, but should be given the broadest scope consistent with the claims.

[0017] Figure 1 This is a flow chart of an artificial intelligence-based cloud conference virtual response method provided by an embodiment of the present application. The artificial intelligence-based cloud conference virtual response method is introduced in detail below.

[0018] Step S110, obtaining the optimized cloud conference interactive system framework and the preset automatic answering script library, performing AI semantic analysis on the automatic answering script library, and generating an answering logic sequence.

[0019] For example, if software development company A is upgrading its cloud conference system, it can pull the latest cloud conference interactive system framework code from the version control system and obtain a set of pre-written auto-response scripts. Next, it can use the natural language processing (NLP) model to perform AI semantic analysis on these auto-response scripts to understand the intent and context of each auto-response script.

[0020] For example, an automatic response script for voice quality issues may be: "Please check your microphone settings or move the microphone closer and speak again." The NLP model can be used to perform AI semantic analysis on the text in the automatic response script library. In this process, each automatic response script is processed and its main function, intent, and the type of information that the user may be seeking are determined.

[0021] Taking the script about sound quality issues as an example, the NLP model will analyze the text and recognize that its intention is to solve the sound transmission problem and involves relevant operational instructions for "microphone settings".

[0022] After the analysis is completed, a structured response logic sequence can be created based on the results of the semantic analysis. This response logic sequence will reorganize the original script content in a format that the program can understand and execute.

[0023] For a script targeting sound quality issues, the response logic sequence might look like this:

[0024] ```json

[0025] {

[0026] "intent": "solve_audio_issue",

[0027] "topic": "microphone_settings",

[0028] "actions": [

[0029] {

[0030] "action_type": "check_microphone",

[0031] "message": "It looks like there might be a problem with your microphone. Please check your microphone settings."

[0032] },

[0033] {

[0034] "action_type": "repeat_request",

[0035] "message": "If you still have problems after adjusting, please repeat your question or provide more details."

[0036] } ]

[0038] }

[0039] ```

[0040] In this response logic sequence, "intent" indicates that the purpose of the response is to solve the problem about sound; "topic" specifies that the specific field of the problem is microphone settings; "actions" defines the actions that should be taken, such as first instructing the user to check the microphone, and then asking the user to repeat the question for further confirmation.

[0041] In order to support the semantic analysis of the automatic answering script in the cloud conference system, the training of the NLP model is implemented based on artificial intelligence technology. The task of this NLP model is to extract the intent, topic, and corresponding action instructions from the answering script written by the user. The following is an example of how to train the NLP model:

[0042] First, a large amount of sample automatic response script data is collected, including user queries and system responses. The data preprocessing stage includes steps such as cleaning (e.g., removing noise, correcting errors), word segmentation, part-of-speech tagging, and named entity recognition, with the goal of converting the original text into a format suitable for processing by machine learning models. Next, features that can represent language patterns and contextual meanings, such as word embedding vectors, sentence structure features, etc., are selected or constructed. On this basis, AI architectures such as sequence-to-sequence (Seq2Seq) models, Transformer, or BERT can be used to deal with the problem of natural language understanding. The model can be trained using sample automatic response script data to enable it to perform accurate semantic parsing of new response scripts. After the model training is completed, the performance of the model is evaluated through cross-validation and independent test sets. They may use indicators such as accuracy, recall, and F1 score to measure the effectiveness of the model in identifying intent, topics, and action instructions. Based on the test results, the model structure may be adjusted, feature engineering may be re-performed, or different pre-trained models may be tried to further improve accuracy and response quality. Finally, the trained NLP model is deployed to the cloud conference system. During deployment, the model’s performance can be closely monitored and regularly fine-tuned and updated to adapt to changes in user behavior and new meeting scenarios.

[0043] Through these steps, a series of text replies written by users can be converted into a machine-executable logical sequence, so that the automatic answering system can not only provide standard replies according to preset scripts, but also dynamically generate more appropriate and personalized responses based on the user's specific questions.

[0044] Step S120, determining the logic processing unit to be optimized from the cloud conference interactive system framework, and determining the response logic implantation point map according to the response logic sequence and the logic processing parameter sequence corresponding to the logic processing unit to be optimized.

[0045] For example, the existing cloud conference interaction system can be further examined, especially to identify the logic processing units to be optimized in the automatic answering function. By mapping the relationship between the logic processing units to be optimized and their logic processing parameters, a map of answering logic implantation points can be created, which marks where to insert or modify the logic to achieve the best automatic answering performance.

[0046] Exemplarily, this step can be explained by a more specific example:

[0047] Assume that the automatic answering function in the cloud conference system is currently being optimized. First, it is necessary to determine which logical processing units in the cloud conference interactive system framework are to be optimized. To simplify the discussion, assume that it is decided to focus on optimizing the user question response module, that is, the logical processing unit to be optimized is the user question response module, because it directly interacts with the automatic answering script.

[0048] In this user question response module, there may be multiple logic processing parameters, such as:

[0049] ResponseTime: The time required to respond to user questions.

[0050] RecognitionAccuracy: Determines the threshold for speech recognition accuracy.

[0051] FollowUpPrompt: A parameter that prompts the user to ask additional questions or provide more information if the question is not understood.

[0052] With the answer logic sequence generated in the previous step, the intent and expected behavior of each auto-answer script are now determined. For example, there is an auto-answer script that is used to answer audio quality related questions such as "Does my voice sound clear?"

[0053] Here is an example of how to determine the graph of response logic implant points:

[0054] For auto-answer scripts that deal with audio quality issues, you may choose to adjust the RecognitionAccuracy parameter, as high accuracy of speech recognition is critical to understanding user queries regarding audio issues.

[0055] In addition, the answer to this question may not require too many follow-up prompts, so the FollowUpPrompt parameter may be set to a lower sensitivity or simply turned off this feature.

[0056] Next, we will create a response logic implantation point map, which contains the following information:

[0057] 1. Each parameter in the user question response module (such as ResponseTime, RecognitionAccuracy, FollowUpPrompt).

[0058] 2. How each parameter affects the execution of a specific script in the response logic sequence.

[0059] 3. The range of parameter values ​​that need to be adjusted.

[0060] For example, a response logic placement map might show:

[0061] {

[0062] "Audio Quality Query": {

[0063] "ResponseTime": "default", / / No need to change the response time

[0064] "RecognitionAccuracy": "high", / / Improve speech recognition accuracy

[0065] "FollowUpPrompt": "low" / / Reduce follow-up prompts

[0066] },

[0067] / / Parameter adjustment for other scripts...

[0068] }

[0069] Through this response logic implantation point map, we can determine which control nodes need to be adjusted and how to adjust these control nodes to achieve the best automatic response experience. Each adjustment is based on the needs of the response logic sequence, ensuring that the automatic response mechanism is both fast and accurate.

[0070] Step S130, determining the adjustable control nodes associated with each response instruction in the response logic sequence from the logic processing parameter sequence according to the response logic implantation point map, and performing optimization operations on the corresponding adjustable control nodes according to each response instruction to generate each customized response control node, and the positions of the adjustable control nodes determined by the response logic implantation point map are dispersed.

[0071] For example, the response logic implantation point map can be used to analyze which control parameters need to be adjusted for different response logics. For example, it may be found that some response logics require shorter response times, so the timeout parameters of those control nodes are adjusted. Other response logics may require higher accuracy, so the data thresholds of the relevant control nodes need to be adjusted. In this way, each response instruction is optimized to suit their specific needs and context. The optimization of control nodes is distributed in different modules throughout the system.

[0072] Exemplarily, continuing based on the above scenario, the following describes how to use the answering logic implantation point map to optimize the automatic answering function in the cloud conference system.

[0073] Assume that an automatic response logic sequence has been determined through semantic analysis, especially for user questions about voice clarity, such as "Does my voice sound clear?", and a response logic implantation point map has been created based on this response logic sequence, including control parameters such as response time (ResponseTime), speech recognition accuracy (RecognitionAccuracy), and follow-up prompt (FollowUpPrompt).

[0074] Now, it is necessary to determine and optimize the adjustable control nodes associated with the response instructions based on the response logic implantation point map.

[0075] First, determine the adjustable control nodes associated with the automatic response to the question "Does my voice sound clear?" Based on the response logic implantation point map, the following adjustable control nodes can be found:

[0076] Audio input processor (responsible for capturing and initially processing user voice)

[0077] Speech recognition engine (converts speech to text)

[0078] Autoresponse generator (generates reply messages)

[0079] Next, each control node is optimized specifically.

[0080] For audio input processors, their gain or noise cancellation settings may be adjusted to ensure that the user's voice is clearly captured even in the presence of loud background noise.

[0081] For the speech recognition engine, its accuracy threshold will be raised so that it can accurately understand the content of the question even when the user asks about the sound quality.

[0082] For the automatic answer generator, judgment logic will be programmed so that when it recognizes that the user is asking about the sound clarity, it can provide a quick link to test the audio function, or give a confident reply confirming that the sound status is good.

[0083] After each optimization, the original adjustable control nodes are updated to specially customized versions of custom answer control nodes that specifically respond to user questions related to sound clarity. For example, the audio input processor may have a new profile called "clarity optimization", which is automatically enabled when it detects that the user is asking questions about sound. The speech recognition engine may have a special mode added to specifically recognize terms and sentence patterns related to audio quality. The automatic answer generator stores a variety of preset replies and can select the most appropriate one to send to the user based on the context.

[0084] In this process, although the relevant control nodes may be dispersed in the cloud conference interactive system framework (for example, audio processing is implemented on the front end, while speech recognition and response generation are on the back-end server), they together form a collaborative network to provide customized automatic response services for users' specific questions. Through such optimization, the cloud conference system can handle users' queries about sound clarity more accurately and effectively, improving the user experience.

[0085] Step S140, based on the respective customized response control nodes, an enhanced logic processing unit corresponding to the logic processing unit to be optimized is generated, and based on the enhanced logic processing unit, a cloud conference interactive system framework with an enhanced cloud conference response mechanism is generated, wherein the cloud conference interactive system framework with an enhanced cloud conference response mechanism is used to interactively process any cloud conference request data to be responded to.

[0086] For example, after the optimization of the above steps, the new answer control node can be integrated into the original logic processing unit to create an enhanced logic processing unit. For example, they may integrate an improved timeout control mechanism into the user question identification module. Finally, these enhanced logic processing units are recompiled and deployed on the cloud platform to form an updated version of the cloud conference interaction system framework. This new cloud conference interaction system framework can provide more intelligent and flexible automatic answering services to meet the various interaction needs of users in cloud conferences.

[0087] Exemplarily, based on the previous steps, the logical processing units to be optimized in the user question response module have been defined, and each customized response control node has been determined. Now, these adjustments need to be implemented in the cloud conference interactive system framework to generate enhanced logical processing units and an updated cloud conference interactive system framework. For example, any cloud conference request data to be responded to can be interactively processed based on the cloud conference response mechanism enhanced cloud conference interactive system framework.

[0088] Now that we have created the map of response logic placement points and identified the parameters to optimize, such as RecognitionAccuracy and FollowUpPrompt, we need to implement these changes at the code level.

[0089] For example, the code of the user question response module is modified according to the answer logic implantation point map. For example, if the answer logic implantation point map indicates that audio quality inquiries require high speech recognition accuracy, the relevant code may be updated to use a more advanced speech recognition algorithm or add more sample data to improve the recognition rate. For the FollowUpPrompt parameter, the prompt strategy can be dynamically adjusted according to the user's question type, rather than applying the same strategy across the board.

[0090] As each adjustable control node is optimized, they are integrated back into the user query response module. This may involve changing configuration files, adjusting parameter values ​​in the database, or modifying server-side logic.

[0091] After all customized answer control nodes are integrated, the enhanced user question response module will be recompiled and tested. The focus of the test is to verify whether the new logic processing unit works as expected and whether the automatic answer becomes more accurate, faster, and more in line with user needs. Through unit testing, integration testing, and user acceptance testing (UAT) and other processes, it is ensured that the modification does not introduce new errors and improves the original functions. Finally, after the test passes, the enhanced module will be deployed to the production environment, thus completing the update of the entire cloud conference system framework.

[0092] For example, after completing the above steps, when the user reenters the meeting and asks "Does my voice sound clear?", the enhanced user question response module will quickly start a more advanced voice recognition algorithm to parse the question and decide whether to give a follow-up prompt based on the nature of the question. If the voice is clear, it may automatically reply "Your voice is clear."; if a sound quality problem is detected, it may provide some help information or suggest that the user adjust the microphone settings. In this way, the cloud conference system can interact with users more intelligently, thereby significantly improving user satisfaction.

[0093] Based on the above steps, first, obtain the optimized cloud conference interactive system framework and the preset automatic answering script library, and perform AI semantic analysis on the automatic answering script library to generate an answering logic sequence. Then, identify the logic processing unit to be optimized from the cloud conference interactive system framework, and determine the answering logic implantation point map based on the answering logic sequence and the corresponding logic processing parameter sequence. Through the answering logic implantation point map, determine the adjustable control nodes associated with each answering instruction, and perform optimization operations on these adjustable control nodes to create customized answering control nodes. Finally, use these customized answering control nodes to generate an enhanced version of the logic processing unit, and update the cloud conference interactive system framework accordingly to enhance its answering mechanism. In this way, targeted optimization strategies can be implemented to effectively improve the answering efficiency and accuracy of the cloud conference interactive system framework. This application allows the cloud conference interactive system framework to accurately identify and adjust key control nodes according to the preset answering logic sequence to ensure that each answering instruction can get the best response effect. Since the positions of the adjustable control nodes determined by the response logic implantation point map are decentralized, the optimization operation can be performed independently at multiple key points of the cloud conference interactive system framework, thus avoiding the chain effects and unstable factors that may be caused by centralized adjustments, significantly improving the flexibility and robustness of the cloud conference interactive system framework, and providing users with a smoother and smarter conference experience. In addition, the enhanced logic processing unit can be flexibly integrated into the existing cloud conference interactive system framework, laying the foundation for future system upgrades and function expansions.

[0094] In a possible implementation, in step S120, determining the logic processing unit to be optimized from the cloud conference interaction system framework includes:

[0095] Step S121, obtaining the unit quantity of the target logic unit in the cloud conference interactive system framework, wherein the target logic unit refers to the functional layer in the cloud conference interactive system framework that is responsible for processing data streams and command sequences.

[0096] Step S122, calculating a hierarchical index value based on the total number of logical layers and preset target parameters, and determining a logic processing unit to be optimized for integrating a virtual response function in the cloud interaction system framework according to the hierarchical index value.

[0097] For example, suppose a cloud conference system has multiple logic units, including a user question response module, etc. Each logic unit has a series of logic processing parameters that determine how the logic unit operates. In order to optimize the automatic answering function, it is necessary to determine which logic units and which parameters in these logic units to adjust. Next, each step will be explained with examples using specific scenarios:

[0098] Assume that the automatic answering function of the cloud conference system is being optimized, especially the response to the user's question "Does my voice sound clear?" It is necessary to improve several logical units including the user question response module.

[0099] The cloud conference interactive system framework includes multiple logical units, such as audio processing unit, user question response module, network connection manager, etc. Through analysis, it is determined that the audio processing unit and the user question response module are the target logical units that need to be optimized.

[0100] The response logic sequence may include the following instructions: "detect audio input quality", "generate speech clarity feedback", etc. The implantation point step size for these instructions can be calculated, for example, the position of each instruction in the logic processing parameter sequence can be defined according to the order of the logic unit.

[0101] Next, an initial coordinate sequence is created, which may be the parameter settings of the current audio processing unit and the user question response module. Using the implant point step, the coordinates of the implant point can be determined from the starting point sequence. This involves selecting appropriate parameters and modifying them to best meet the needs of the response logic sequence.

[0102] Next, the selected parameters can be updated to the target values ​​to generate new implantation point coordinate feature data. For example, the noise filtering threshold of the audio processing unit may be increased to enhance the sound clarity.

[0103] Then, the corresponding control nodes can be found in the user question response module based on the feature data of these implantation point coordinates, such as the conditional statement used to determine whether to start the automatic clarity test. Finally, an updated answer logic implantation point map will be obtained, which clearly shows which parameters are adjusted and their specific locations in the logic processing unit.

[0104] As a result, key logic processing units and parameters that affect the performance of automatic answering of sound clarity issues can be systematically identified and optimized, ensuring that the system can respond to users' questions more accurately and effectively.

[0105] In a possible implementation, determining the response logic implantation point map according to the response logic sequence and the logic processing parameter sequence corresponding to the logic processing unit to be optimized in step S120 includes:

[0106] Step S123, calculating according to the length corresponding to the logic processing parameter sequence and the response instruction statistics corresponding to the response logic sequence to generate the implantation point step length.

[0107] Step S124, initializing an initial coordinate sequence having the same length as the logic processing parameter sequence as a starting point sequence.

[0108] Step S125, determining the coordinates of each implantation point from the starting point sequence according to the implantation point step length and the response instruction statistics corresponding to the response logic sequence.

[0109] Step S126, updating the basic parameters of the coordinates of each implantation point in the starting point sequence to the target parameters, generating implantation point coordinate feature data corresponding to the logic processing parameter sequence, and using the implantation point coordinate feature data as the response logic implantation point map.

[0110] For example, suppose a logic processing parameter sequence contains various parameters, such as `[ResponseTime, RecognitionAccuracy, FollowUpPrompt]`.

[0111] Thus, the implantation point step size can be calculated based on the command statistics in the response logic sequence. For example, if the command related to speech clarity appears more frequently in the entire response logic sequence, the implantation point step size may be set smaller to more finely adjust the relevant parameters in the logic processing parameter sequence.

[0112] On this basis, an initial coordinate sequence can be created, which is the same length as the logic processing parameter sequence. This sequence represents the current setting of each parameter, for example, `[500ms, 85%, true]` represents the default response time, recognition accuracy threshold, and whether to enable subsequent prompts.

[0113] Next, the position of each implant point can be determined from the start point sequence based on the implant point step size and the statistics in the response logic sequence. For example, if the speech intelligibility problem requires a faster response time, the implant point coordinates may move the ResponseTime parameter position forward.

[0114] Then, the basic parameters in the starting point sequence can be updated to the target parameters. Assume that in order to improve the accuracy of speech clarity response, `RecognitionAccuracy` may be increased from 85% to 90%, and `ResponseTime` may be reduced from 500ms to 300ms. The updated implantation point coordinate feature data is `[300ms, 90%, true]`. This set of implantation point coordinate feature data constitutes the response logic implantation point map, which is used to characterize how to adjust each parameter to optimize a specific response logic.

[0115] Therefore, in step S130, determining the adjustable control nodes associated with each response instruction in the response logic sequence from the logic processing parameter sequence according to the response logic implantation point map includes:

[0116] Step S131, determining the coordinate features of each target parameter from the implantation point coordinate feature data, and acquiring the control node corresponding to the same coordinate feature from the logic processing parameter sequence according to the coordinate features of each target parameter.

[0117] Step S132: using the control nodes corresponding to the same coordinate features as adjustable control nodes associated with each response instruction in the response logic sequence.

[0118] For example, you can view the updated implant point coordinate feature data and identify the new coordinate features for each target parameter (such as `ResponseTime` is 300ms, `RecognitionAccuracy` is 90%).

[0119] Then, based on these coordinate features, the corresponding control nodes are found in the sequence of logical processing parameters. This may mean identifying the code module in the system responsible for handling response time, or the settings in the configuration file that can adjust the accuracy of speech recognition. Finally, these control nodes can be associated with the corresponding response instructions. In the previous example, the response instruction for the sound clarity problem ("Your voice is clear." or "Please check your microphone settings.") is now directly associated with a specific control node (such as the configuration of the audio processor, or the accuracy setting of the speech recognition engine).

[0120] In a possible implementation, determining the response logic implantation point map according to the response logic sequence and the logic processing parameter sequence corresponding to the logic processing unit to be optimized in step S120 includes:

[0121] Step S127, obtaining the logic processing parameter array corresponding to the logic processing unit to be optimized, mapping and transforming the logic processing parameter array to generate the logic processing parameter sequence, calculating according to the length corresponding to the logic processing parameter sequence and the response instruction statistics corresponding to the response logic sequence, and generating the implantation point step length.

[0122] Step S128, obtaining a starting point sequence, wherein the sequence quantization unit of the starting point sequence is equal to the sequence quantization unit of the logic processing parameter sequence.

[0123] Step S129, determining the coordinates of each implant point from the starting point sequence according to the implant point step size and the response instruction statistics corresponding to the response logic sequence, updating the basic parameters of the coordinates of each implant point in the starting point sequence to the target parameters, generating implant point coordinate feature data corresponding to the logic processing parameter sequence, mapping and transforming the implant point coordinate feature data according to the attributes corresponding to the logic processing parameter array, generating the implant point feature distribution corresponding to the logic processing parameter array, and using the implant point feature distribution as the response logic implant point map.

[0124] Therefore, in step S130, determining the adjustable control nodes associated with each response instruction in the response logic sequence from the logic processing parameter sequence according to the response logic implantation point map includes:

[0125] Step S133, determining the characteristic distribution coordinates of each target parameter from the implantation point characteristic distribution, and acquiring the control nodes corresponding to the same characteristic distribution coordinates from the logic processing parameter array according to the coordinate characteristics of each target parameter.

[0126] Step S134: using the control nodes corresponding to the same characteristic distribution coordinates as adjustable control nodes associated with each response instruction in the response logic sequence.

[0127] For example, the logic processing units to be optimized of the user question response module have been determined, and the parameter settings of each logic processing unit to be optimized are first obtained. These parameters may include thresholds, delay times, recognition accuracy, etc., to form a logic processing parameter array.

[0128] For example, for the speech analysis layer, the parameter array may contain the sampling rate of the speech signal, the noise suppression level, etc. Then, this multi-dimensional parameter array mapping can be converted into a linear sequence of logical processing parameters for subsequent operations.

[0129] On this basis, the length of the logic processing parameter sequence can be calculated, and the implantation point step length can be determined according to the statistical value of the response instruction in the response logic sequence. The step length determines the granularity of parameter adjustment. For example, if there are more user inquiries involving sound problems, the parameter adjustment step length related to this will be set smaller to achieve more precise adjustment.

[0130] Next, the starting point sequence of the current system configuration can be determined, which contains the current value of each parameter in the logical processing parameter sequence. The starting point sequence has the same quantitative unit as the logical processing parameter sequence, ensuring that parameter adjustments can be performed seamlessly.

[0131] Thus, the calculated implantation point step size can be used to determine the implantation point coordinates from the starting point sequence, that is, the target position where each parameter needs to be adjusted.

[0132] On this basis, according to the needs of the response logic sequence, the basic parameters in the starting point sequence are updated to the target parameters, such as increasing the recognition accuracy from 85% to 90%. After the update, new implant point coordinate feature data is generated, reflecting the optimized parameter settings.

[0133] The new implant point coordinate feature data is mapped and transformed according to the properties of the logic processing parameter array to generate the implant point feature distribution. The implant point feature distribution shows the distribution of the optimized parameters in the entire module as the response logic implant point map.

[0134] By implanting the feature distribution of each target parameter, the feature distribution coordinates are determined. Then, the corresponding control node is found from the logic processing parameter array based on these coordinate features. For example, if the noise suppression level of the speech analysis layer needs to be adjusted, the corresponding control node can be found in the system settings or code.

[0135] Finally, these control nodes can be associated with corresponding instructions in the response logic sequence. For example, when a user asks a question about voice clarity, the instruction "Please move closer to the microphone and speak again" may trigger a control node that adjusts the noise suppression level of the voice capture layer.

[0136] Through this series of technical steps, it is possible to optimize the "user question response module" in the cloud conference system in a targeted manner and improve the efficiency and effectiveness of its automatic answer response to sound clarity issues.

[0137] In a possible implementation, in the aforementioned embodiment, the implant point step length is generated by calculation based on the length corresponding to the logic processing parameter sequence and the response instruction statistic corresponding to the response logic sequence, including: when the length is greater than the response instruction statistic, calculating the rounded value of the ratio of the length to the response instruction statistic, and calculating the difference between the rounded value and the set value, to generate the implant point step length.

[0138] For example, this technical step will be illustrated in conjunction with the optimization of the automatic answering function in Company A's cloud conference system.

[0139] Suppose you are optimizing a user question response module to improve the quality of answers to specific questions (e.g., "Does my voice sound clear?"). Before you can optimize, you need to identify the best locations (i.e., drop-in points) in the logic processing parameter sequence to improve the auto-answering function. These drop-in points will be where you will implement the parameter adjustments.

[0140] The logic processing parameter sequence includes various parameters that affect the behavior of the user question response module. The response instruction statistics refer to the number of times the relevant response instructions appear in the system, such as how many times the user asked questions about sound clarity.

[0141] Assume that the length of the logic processing parameter sequence is 20, which means that the user question response module has 20 adjustable parameters. The response instruction statistics value is 5, which means that there are 5 response instructions related to sound clarity.

[0142] According to the given rules, when the length of the parameter sequence is greater than the response instruction statistics, it is necessary to calculate the ratio between them and round it up. For example, the ratio is 20 / 5=4. Then, calculate the difference between this rounded value and a set value. Assuming that the set value is a fixed optimization constant, such as 3, then the difference is 4-3=1. Therefore, the implantation point step is determined to be 1.

[0143] The above calculations lead to an implantation point step size of 1, which means that an implantation point will be set for every other parameter in the logical processing parameter sequence. This way, the optimization effort can be evenly distributed and ensures comprehensive improvements to the user question response module. For example, if the parameter sequence is a linear arrangement, every other parameter will be evaluated and possibly adjusted to optimize the automatic answering function.

[0144] In summary, in this scenario, the calculation of the implantation point step can effectively identify which positions in the parameter sequence of the user question response module are most suitable for optimization operations, thereby improving the efficiency of the cloud conference system in answering specific questions.

[0145] In a possible implementation, the method further includes:

[0146] Step A110, when the length is not greater than the response instruction statistical value, determine N logical processing units to be optimized from the cloud conference interaction system framework, and obtain the target logical processing parameter sequence corresponding to the N logical processing units to be optimized, and the target length in the target logical processing parameter sequence is greater than the response instruction statistical value.

[0147] For example, in a cloud conference system optimization project, if it is found that the original logic processing parameter sequence length is not sufficient to cover all response instruction statistics, it is necessary to determine additional N logic processing units to be optimized from the cloud conference interactive system framework.

[0148] For example, in addition to the identified audio processing unit and user question response module, other modules such as the network connection manager and the data storage manager may also need to be optimized. Next, the target logic processing parameter sequences corresponding to these newly determined logic processing units are obtained to ensure that their target lengths are greater than the response instruction statistics.

[0149] Step A120, calculating based on the target length corresponding to the target logic processing parameter sequence and the response instruction statistics corresponding to the response logic sequence, to generate a target implantation point step length.

[0150] For example, the target implantation point step length can be calculated based on the length of the new target logic processing parameter sequence and the response instruction statistics of the response logic sequence. This step length determines the interval of parameter adjustment to ensure that the optimization covers all necessary logic processing units.

[0151] Step A130, obtaining a target starting point sequence, wherein the sequence quantization unit of the target starting point sequence is equal to the sequence quantization unit of the target logic processing parameter sequence.

[0152] For example, you can create a target starting point sequence containing the current system configuration, whose quantization unit (sequence length) is the same as the target logic processing parameter sequence.

[0153] Step A140, determining the coordinates of each target implantation point from the target starting point sequence according to the target implantation point step length and the response instruction statistics corresponding to the response logic sequence.

[0154] For example, using the target implantation point step length, the coordinates of each target implantation point can be determined from the target starting point sequence.

[0155] Exemplarily, the calculated target implantation point step size can be used as a reference to traverse the target starting point sequence. Starting from the starting position of the target starting point sequence, the positions of the implantation points are marked one by one according to the step size value. Each time the step size unit is moved, the point reached is regarded as an implantation point until the entire target starting point sequence is covered. For each implantation point in the target starting point sequence, its index position (i.e., implantation point coordinates) can be recorded, and the original parameter value corresponding to the index position can be identified.

[0156] Assume that the target logic processing parameter sequence contains 30 parameters, and the response instruction statistics value is 10. Assume that the calculated target implant point step is 3. Counting from the first parameter, every two parameters (because the step is 3) mark an implant point. In other words, the 1st, 4th, 7th, 10th... parameters will become the target implant points. The coordinate information will record the indexes of these implant points, for example: 0, 3, 6, 9,... and so on.

[0157] Each marked implantation point will correspond to a specific set of operations or adjustments aimed at optimizing relevant parameters to improve the performance of the user question response module.

[0158] In this way, it can be ensured that subsequent optimization operations are distributed in the entire target logic processing parameter sequence in a targeted manner, thereby improving the response efficiency and efficiency of the entire cloud conference system.

[0159] Step A150, updating the basic parameters of the coordinates of each target implantation point in the target starting point sequence to target parameters, and generating target implantation point coordinate feature data corresponding to the target logic processing parameter sequence.

[0160] For example, the basic parameters in the target starting point sequence can be updated to the target parameters to generate target implantation point coordinate feature data corresponding to the target logic processing parameter sequence.

[0161] For example, in the above scenario, it is necessary to determine the configuration status of each logical processing unit to be optimized in the current cloud conference system, that is, the target starting point sequence. These target starting point sequences contain all the basic parameters that the current operation depends on. Basic parameters may include buffer size, audio processing threshold, network delay tolerance, etc.

[0162] Then, the previously calculated target implantation point step size is used to locate the position of each implantation point from the target starting point sequence. These implantation points are the locations where parameter adjustment is scheduled.

[0163] Once the target implantation point is determined, the appropriate target parameter value needs to be selected based on the optimization goal. For example, if one implantation point is the level of noise suppression in an audio processing module, and statistics indicate that speech intelligibility needs to be improved, then the target parameter may be to increase the strength of that noise suppression. The update operation can be as simple as replacing a value or it can be more complex, such as adjusting an algorithm or rewriting a code segment.

[0164] After the target parameters are set, the new parameter values ​​are recorded to create a new sequence containing the updated parameters of all implant points, namely the target implant point coordinate feature data. This new target implant point coordinate feature data now represents the desired state after the optimization measures are executed.

[0165] Before officially applying new parameters, it is usually necessary to test to verify the effectiveness and stability of the new configuration. After the test is passed, these new target implantation point coordinate feature data can be deployed to the cloud conference system, replacing the original configuration parameters to complete the optimization process.

[0166] Taking the cloud conference system as an example, suppose you want to improve a specific speech recognition rate. You may adjust specific parameters in the speech recognition logic processing unit (such as the noise suppression threshold). If the original noise suppression threshold parameter is 0.7, and after calculation, it is determined that it should be increased to 0.9 to achieve the best recognition effect, then 0.7 is the basic parameter and 0.9 is the target parameter. The corresponding noise suppression threshold position can be found in the target implantation point sequence and updated to 0.9, generating new implantation point coordinate feature data, and then pushing these changes to the system.

[0167] Step A160, based on the target implant point coordinate feature data, determine the target adjustable control node associated with each response instruction in the response logic sequence from the target logic processing parameter sequence, and perform optimization operations on the corresponding target adjustable control nodes based on each response instruction to generate each target customized response control node, and the positions of the target adjustable control nodes are dispersed.

[0168] For example, based on the target implantation point coordinate feature data, the target adjustable control node associated with each response instruction in the response logic sequence is determined, and the required optimization operation is performed. For example, if the speech analysis layer needs to process the questions of users with fast speaking speed more finely, the relevant control nodes can be optimized to improve the processing speed. The optimized control node positions are dispersed, ensuring the stability and flexibility of the system.

[0169] Step A170, based on the target customized response control nodes, generates enhanced logic processing units associated with each of the N logic processing units to be optimized, and based on the enhanced logic processing units associated with each of the N logic processing units to be optimized, generates a target cloud conference interaction system framework after the cloud conference response mechanism is enhanced.

[0170] For example, according to the target customized response control node, an enhanced logic processing unit associated with each of the N to-be-optimized logic processing units can be generated. Then, the cloud conference response mechanism is updated using these enhanced logic processing units to generate an enhanced target cloud conference interactive system framework. For example, by adding an advanced speech enhancement processor as an enhanced logic processing unit of the audio processing unit, the audio quality can be improved and echo and noise can be reduced.

[0171] Through the above steps, not only the specific logic processing unit was carefully optimized, but also the optimization scope was expanded to other related modules, thus comprehensively improving the automatic response performance of the cloud conference system. These changes were eventually integrated into the framework of the cloud conference interactive system, forming a more efficient and intelligent response mechanism.

[0172] In a possible implementation manner, the adjustable control node includes a control parameter to be adjusted.

[0173] Step S130 performs optimization operations on the corresponding adjustable control nodes according to the response instructions to generate customized response control nodes, including:

[0174] Step S135 , determining a current response instruction from the response instructions, and obtaining a control parameter to be adjusted corresponding to the current response instruction.

[0175] Step S136: when the control parameter to be adjusted has a decimal part, determine the initial parameter value of the setting coordinate from the decimal part corresponding to the control parameter to be adjusted.

[0176] Step S137, performing an optimization operation on the initial parameter value of the set coordinate according to the current response instruction, and generating a customized response control node corresponding to the current response instruction.

[0177] For example, the modulus operation value corresponding to the initial parameter value of the set coordinates may be calculated, and when the modulus operation value does not match the current response instruction, the sum of the initial parameter value of the set coordinates and the set optimization parameter is calculated to generate a customized response control parameter value, and the modulus operation value corresponding to the customized response control parameter value matches the current response instruction. Then, the customized response control parameter value covers the initial parameter value of the set coordinates in the mantissa corresponding to the control parameter to be adjusted, and generates a customized response control node corresponding to the current response instruction.

[0178] To better illustrate the above technical content, we continue to take the optimization of the user question response module of the cloud conference system as an example. It is necessary to optimize the adjustable control nodes related to the automatic answering performance in the user question response module. The following is a scenario description of the specific steps:

[0179] The user question response module is currently being optimized, with a particular focus on those responses to questions about voice clarity. These responses are supported by a set of adjustable control nodes, each of which has multiple control parameters.

[0180] Assume that a specific user query is being processed: "Does my voice sound clear?" The team identifies the current response instruction from the response logic sequence. Then, it obtains the control parameter to be adjusted corresponding to this response instruction, such as the sensitivity setting of the speech recognition engine.

[0181] In the control parameters to be adjusted, notice that there is a decimal part. For example, the sensitivity setting may be a floating point number, such as 2.5 (where 0.5 is the decimal part). Then, the decimal part of 2.5, 0.5, can be used as the initial parameter value for setting the coordinates.

[0182] Next, the modulo value of the initial parameter value of 0.5 can be calculated. Assume that the base of the modulo operation is 1, so 0.5 % 1 = 0.5. If the modulo value does not match the expected result (in this scenario, it may be required that the modulo value is 0 to ensure the highest sensitivity), then adjustments are required. The initial parameter value can be added to the set optimization parameter. Assuming the set optimization parameter is 0.5, the new custom reply control parameter value is calculated: 0.5 + 0.5 = 1.0. This newly calculated modulo value of the custom reply control parameter value of 1.0 (1.0 % 1) matches the expectation of the current reply instruction because 1.0 % 1 = 0.

[0183] Finally, the original mantissa can be overwritten with the calculated custom response control parameter value of 1.0, generating a new custom response control node with a sensitivity setting of 3.0 (2 + 1.0). This means that when the user asks about voice clarity, the new sensitivity setting will be used to improve the accuracy of speech recognition, thereby improving the efficiency of the response.

[0184] Through this series of optimization operations, control parameters can be adjusted in a targeted manner, allowing the cloud conference system to respond more accurately to specific types of user inquiries and enhance user experience.

[0185] In a possible implementation manner, after step S140, the method further includes:

[0186] Step S150, obtaining a cloud conference response test request, wherein the cloud conference response test request carries a cloud conference interactive system framework and a test response script library after the cloud conference response mechanism is enhanced.

[0187] Step S160, performing AI semantic analysis on the test response script library, generating a test response logic sequence, and determining a test enhancement logic processing unit from the cloud conference interaction system framework after the cloud conference response mechanism is enhanced.

[0188] Step S170, determining a test response logic implantation point map according to the test response logic sequence and the test logic processing parameter sequence corresponding to the test enhancement logic processing unit.

[0189] Step S180, determining each test response control node from the test logic processing parameter sequence according to the test response logic implantation point map.

[0190] Step S190, extracting response instructions according to each test response control node, generating each target response instruction, determining response instruction features according to each target response instruction, decoding the response instruction features, and generating a response extraction script library.

[0191] Step S110: When the response extraction script library matches the test response script library, response test success information is generated, and the response test success information is returned to the test server corresponding to the cloud conference response test request.

[0192] For example, suppose that an enhanced logic processing unit has been generated by customizing and improving the logic processing unit to be optimized and integrated into the cloud conference interaction system framework. Now, it is necessary to verify whether these changes effectively improve the response mechanism.

[0193] Here are the specific steps on how to implement the testing and validation process:

[0194] First, a cloud conference response test request including the latest cloud conference interactive system framework and a test response script library may be received from the test server. The test response script library includes various expected user queries and corresponding ideal responses.

[0195] Then, AI semantic analysis can be performed on the test response scripts in the test response script library to create a test response logic sequence that reflects the relationship between the user query and the expected response. In this process, which enhanced logic processing units will participate in the test are also identified.

[0196] Next, the test response logic implantation point map can be determined based on the test response logic sequence and the related test logic processing parameter sequence. This test response logic implantation point map can be used to clarify the implantation point location of each logic processing unit in the test. Using the test response logic implantation point map, all test response control nodes can be determined from the test logic processing parameter sequence.

[0197] According to the test response control node, the response instruction of each test response control node is extracted and the target response instruction is generated. Each target response instruction has specific features, which are decoded to generate a response extraction script library. If the response extraction script library matches the original test response script library, indicating that the response mechanism works properly, a response test success message is generated.

[0198] The response test success information is sent back to the test server that initiated the test request, confirming that the enhancement of the cloud conference response mechanism is successful.

[0199] During the test, the test response parameter values ​​of each test response control node can be checked to ensure that they meet the preset conditions, thereby ensuring the accuracy of the test.

[0200] Through the above process, the effect of the enhanced response mechanism in the cloud conference system can be systematically verified to ensure that the newly added functions can not only correctly respond to the preset queries, but also maintain compatibility with the old system. This complete testing and verification process is an important part of software development. It ensures product quality and provides a stable and reliable system for end users.

[0201] The test response control node in each of the test response control nodes includes a test response parameter value.

[0202] Step S190 may include:

[0203] Step S191, when the test response parameter value has a decimal part, determine the test response parameter value of the set coordinates from the decimal part corresponding to the test response parameter value.

[0204] Step S192, extracting definition information of a target response instruction corresponding to the test response parameter value according to the test response parameter value of the set coordinate.

[0205] For example, the modulo operation value corresponding to the custom response control parameter value of the set coordinate may be calculated, and the modulo operation value may be used as the definition information of the target response instruction corresponding to the test response parameter value.

[0206] Exemplarily, by creating multiple test response control nodes, each test response control node contains specific parameter values ​​for testing. For example, there may be a test response control node specifically used to test speech recognition rate and response effect under different noise levels. These test response parameter values ​​may be similar to actual control parameter values, but are set to specific values ​​to trigger system responses to evaluate their performance.

[0207] Assuming a test response parameter value is 2.5, you can notice that this value has a tail part 0.5. Depending on the test design, the tail part may represent a specific test scenario or condition. For example, 0.5 may represent a test condition in a low noise environment.

[0208] Next, a modulo operation is performed on the test response parameter value (i.e., the mantissa) of the set coordinates. If the base is 1, then the modulo value is 0.5 % 1 = 0.5. This modulo value is used to define the target response command. For example, a modulo value of 0.5 might correspond to a predefined response command, such as "Please adjust your microphone for improved clarity."

[0209] Through the above steps, the corresponding target response instruction can be extracted according to the test response parameter value, and it can be verified whether the optimization effect of the automatic response function works as expected.

[0210] In a possible implementation, the method further includes:

[0211] Step B110, obtaining target response script data and the cloud conference interaction system framework to be optimized, performing AI semantic analysis on the target response script data, and generating a response logic sequence.

[0212] Step B120, determining the logic processing unit to be upgraded from the cloud conference interactive system framework to be optimized, and determining the response logic implantation point map according to the response logic sequence and the control parameter sequence corresponding to the logic processing unit to be upgraded.

[0213] Step B130, determining the control nodes to be optimized associated with each response instruction in the response logic sequence from the control parameter sequence according to the response logic implantation point map, and performing upgrade operations on the corresponding control nodes to be optimized according to each response instruction to generate each customized response control node.

[0214] Step B140: generating an upgraded logic processing unit corresponding to the logic processing unit to be upgraded according to each customized response control node, and obtaining a cloud conference interactive system framework with an enhanced response mechanism according to the upgraded logic processing unit.

[0215] For example, in the above scenario, assume that user feedback and usage data are collected, and based on this information, a series of improved target response script data is written. At the same time, the existing cloud conference interaction system framework is analyzed to determine that it needs to be optimized to better serve users.

[0216] The NLP model can be used to perform AI semantic analysis on the target answer script data to understand the intent and corresponding operations of each target answer script data. This step results in a logical sequence of answers that clearly indicates the logical steps that need to be performed to meet user needs.

[0217] Next, you can review the existing cloud conference system framework and identify those logical processing units that need to be upgraded. For example, you may find that the audio processing unit needs to be upgraded to improve sound clarity.

[0218] Combine the logic processing unit to be upgraded and the response logic sequence to create a response logic implantation point map. This response logic implantation point map depicts which control parameters need to be adjusted and the location of these parameters in the logic processing unit. According to the response logic implantation point map, determine the control nodes to be optimized associated with the response instruction. Then, perform the necessary upgrade operations on these control nodes, such as adjusting parameter values ​​or replacing algorithms. For example, if the noise suppression algorithm needs to be improved, new digital filter parameters may be added or replaced with a more advanced algorithm model. After completing the customized upgrade of the control node, these changes are integrated back into the logic processing unit to form an upgraded logic processing unit. Subsequently, these upgraded units are integrated into the cloud conference interactive system framework to enhance the system response mechanism.

[0219] Through these steps, key components in the cloud conference system can be optimized to provide faster and more accurate user responses, thereby enhancing the overall user experience. This systematic approach allows the team to iteratively improve the product and continuously adjust the optimization direction based on user feedback and business needs.

[0220] Figure 2 The hardware structure of the cloud conference system 100 provided in the embodiment of the present application for implementing the above-mentioned artificial intelligence-based cloud conference virtual response method is shown, as shown in FIG. Figure 2 As shown, the cloud conference system 100 may include a processor 110 , a machine-readable storage medium 120 , a bus 130 , and a communication unit 140 .

[0221] In an alternative embodiment, the cloud conference system 100 can be a single server or a server group. The server group can be centralized or distributed (for example, the cloud conference system 100 can be a distributed system). In an alternative embodiment, the cloud conference system 100 can be local or remote. For example, the cloud conference system 100 can access information and / or data stored in a machine-readable storage medium 120 via a network. For another example, the cloud conference system 100 can be directly connected to the machine-readable storage medium 120 to access the stored information and / or data. In an alternative embodiment, the cloud conference system 100 can be implemented on a cloud platform. By way of example only, the cloud platform can include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, the like, or any combination thereof.

[0222] The machine-readable storage medium 120 may store data and / or instructions. In an alternative embodiment, the machine-readable storage medium 120 may store data obtained from an external terminal. In an alternative embodiment, the machine-readable storage medium 120 may store data and / or instructions used by the cloud conference system 100 to execute or use to complete the exemplary methods described in this application. In an alternative embodiment, the machine-readable storage medium 120 may include a mass storage device, a removable storage device, a volatile read-write memory, a read-only memory, or the like, or any combination thereof. Exemplary mass storage devices may include magnetic disks, optical disks, solid-state disks, or the like. Exemplary removable storage devices may include flash drives, floppy disks, optical disks, memory cards, compressed disks, magnetic tapes, or the like.

[0223] During the specific implementation process, multiple processors 110 execute computer executable instructions stored in the machine-readable storage medium 120, so that the processor 110 can execute the cloud conference virtual response method based on artificial intelligence in the above method embodiment. The processor 110, the machine-readable storage medium 120 and the communication unit 140 are connected through the bus 130, and the processor 110 can be used to control the sending and receiving actions of the communication unit 140.

[0224] The specific implementation process of the processor 110 can refer to the various method embodiments executed by the above-mentioned cloud conference system 100. The implementation principles and technical effects are similar, and this embodiment will not be repeated here.

[0225] In addition, an embodiment of the present application also provides a readable storage medium, in which computer executable instructions are preset. When the processor executes the computer executable instructions, the above-mentioned cloud conference virtual response method based on artificial intelligence is implemented.

[0226] Similarly, it should be noted that in order to simplify the description disclosed in this application, thereby helping to understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present application, multiple features are sometimes combined into one embodiment, drawings, or descriptions thereof. Similarly, it should be noted that in order to simplify the description disclosed in this application, thereby helping to understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present application, multiple features are sometimes combined into one embodiment, drawings, or descriptions thereof.

Claims

1. A cloud conference virtual answering method based on artificial intelligence, characterized in that: The method comprises: Obtaining the optimized cloud conference interactive system framework and the preset automatic answering script library, performing artificial intelligence AI semantic analysis on the automatic answering script library, and generating an answering logic sequence; Determine the logic processing unit to be optimized from the cloud conference interactive system framework, and determine the response logic implantation point map according to the response logic sequence and the logic processing parameter sequence corresponding to the logic processing unit to be optimized; According to the response logic implantation point map, the adjustable control nodes associated with each response instruction in the response logic sequence are determined from the logic processing parameter sequence, and optimization operations are performed on the corresponding adjustable control nodes according to each response instruction to generate each customized response control node, the positions of the adjustable control nodes determined by the response logic implantation point map are dispersed, and the adjustable control nodes include an audio input processor, a speech recognition engine, and an automatic response generator, wherein, for the audio input processor, the optimization operation includes adjusting the gain or noise elimination setting of the audio input processor, for the speech recognition engine, the optimization operation includes increasing the accuracy threshold of the speech recognition engine, and for the automatic response generator, the optimization operation includes programming and adding judgment logic, the adjustable control node includes control parameters to be adjusted, and the control parameters to be adjusted include: the gain or noise elimination setting of the audio input processor, the accuracy threshold of the speech recognition engine, and the judgment logic of the automatic response generator; Based on the various customized response control nodes, an enhanced logic processing unit corresponding to the logic processing unit to be optimized is generated, and based on the enhanced logic processing unit, a cloud conference interactive system framework with an enhanced cloud conference response mechanism is generated. The cloud conference interactive system framework with an enhanced cloud conference response mechanism is used to interactively process any cloud conference request data to be responded to.

2. The method for virtual answering of cloud conference based on artificial intelligence according to claim 1 is characterized in that: The step of determining the logic processing unit to be optimized from the cloud conference interaction system framework includes: Obtaining the number of target logic units in the cloud conference interactive system framework, wherein the target logic unit refers to a functional layer in the cloud conference interactive system framework that is responsible for processing data streams and command sequences; A hierarchical index value is calculated based on the total number of logical layers and preset target parameters, and a logic processing unit to be optimized for integrating a virtual response function is determined in a cloud-based interactive system framework according to the hierarchical index value.

3. The method for virtual answering of cloud conference based on artificial intelligence according to claim 1 is characterized in that: The step of determining a response logic implantation point map according to the response logic sequence and the logic processing parameter sequence corresponding to the logic processing unit to be optimized includes: Calculate according to the length corresponding to the logic processing parameter sequence and the response instruction statistics corresponding to the response logic sequence to generate the implantation point step length; Initialize an initial coordinate sequence with the same length as the logic processing parameter sequence as the starting point sequence; Determine the coordinates of each implantation point from the starting point sequence according to the implantation point step length and the response instruction statistics corresponding to the response logic sequence; Updating the basic parameters of the coordinates of each implantation point in the starting point sequence to the target parameters, generating implantation point coordinate feature data corresponding to the logic processing parameter sequence, and using the implantation point coordinate feature data as the response logic implantation point map; The step of determining, from the logic processing parameter sequence according to the response logic implantation point map, the adjustable control nodes associated with each response instruction in the response logic sequence comprises: Determining the coordinate features of each target parameter from the implantation point coordinate feature data, and acquiring the control node corresponding to the same coordinate feature from the logic processing parameter sequence according to the coordinate features of each target parameter; The control nodes corresponding to the same coordinate features are used as adjustable control nodes associated with each response instruction in the response logic sequence.

4. The method for virtual answering of cloud conference based on artificial intelligence according to claim 3 is characterized in that: The step of calculating according to the length corresponding to the logic processing parameter sequence and the response instruction statistics corresponding to the response logic sequence to generate the implantation point step length includes: When the length corresponding to the logic processing parameter sequence is greater than the response instruction statistical value, the rounded value of the ratio of the length corresponding to the logic processing parameter sequence to the response instruction statistical value is calculated, and the difference between the rounded value and the set value is calculated to generate the implantation point step.

5. The method for virtual answering of cloud conference based on artificial intelligence according to claim 1 is characterized in that: The performing of optimization operations on the corresponding adjustable control nodes according to the response instructions to generate each customized response control node includes: Determine a current response instruction from the response instructions, and obtain a control parameter to be adjusted corresponding to the current response instruction; When the control parameter to be adjusted has a mantissa part, determining the initial parameter value of the set coordinate from the mantissa part corresponding to the control parameter to be adjusted, wherein the mantissa part is used to represent the remainder of the floating point number after removing the integer part; Performing an optimization operation on the initial parameter value of the set coordinate according to the current response instruction to generate a customized response control node corresponding to the current response instruction; The performing of an optimization operation on the initial parameter value of the set coordinate according to the current response instruction to generate a customized response control node corresponding to the current response instruction includes: Calculating a modulo operation value corresponding to the initial parameter value of the set coordinates, and when the modulo operation value does not match the current response instruction, calculating the sum of the initial parameter value of the set coordinates and the set optimization parameter to generate a customized response control parameter value, and the modulo operation value corresponding to the customized response control parameter value matches the current response instruction; The customized response control node corresponding to the current response instruction is generated based on the customized response control parameter value covering the initial parameter value of the set coordinate in the mantissa part corresponding to the control parameter to be adjusted.

6. The method for virtual answering of cloud conference based on artificial intelligence according to claim 1 is characterized in that: After generating the enhanced logic processing unit corresponding to the to-be-optimized logic processing unit according to each customized response control node, and generating the cloud conference interactive system framework with enhanced cloud conference response mechanism according to the enhanced logic processing unit, the method further includes: Obtain a cloud conference response test request, where the cloud conference response test request carries a cloud conference interactive system framework and a test response script library after the cloud conference response mechanism is enhanced; Performing AI semantic analysis on the test response script library to generate a test response logic sequence, and determining a test enhancement logic processing unit from the cloud conference interaction system framework enhanced by the cloud conference response mechanism; Determine a test response logic implantation point map according to the test response logic sequence and the test logic processing parameter sequence corresponding to the test enhancement logic processing unit; Determining each test response control node from the test logic processing parameter sequence according to the test response logic implantation point map; Extracting response instructions according to each test response control node, generating each target response instruction, determining response instruction features according to each target response instruction, decoding the response instruction features, and generating a response extraction script library; When the response extraction script library matches the test response script library, a response test success message is generated, and the response test success message is returned to the test server corresponding to the cloud conference response test request; The test response control node in each of the test response control nodes includes a test response parameter value; The extracting response instructions according to each test response control node to generate each target response instruction includes: When the test response parameter value has a mantissa part, determining the test response parameter value of the set coordinate from the mantissa part corresponding to the test response parameter value, wherein the mantissa part is used to represent the remainder of the floating point number after removing the integer part; Extracting definition information of a target response instruction corresponding to the test response parameter value according to the test response parameter value of the set coordinate; The step of extracting definition information of a target response instruction corresponding to the test response parameter value according to the customized response control parameter value of the set coordinates includes: The modulo operation value corresponding to the custom response control parameter value of the set coordinate is calculated, and the modulo operation value is used as the definition information of the target response instruction corresponding to the test response parameter value.

7. A cloud conference system, characterized in that: The cloud conference system includes a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores machine-executable instructions, and the machine-executable instructions are loaded and executed by the processor to implement the artificial intelligence-based cloud conference virtual response method of any one of claims 1-6.