Dialogue management method and system, electronic equipment, storage medium and program product

Abstract

The invention provides a dialogue management method and system, electronic equipment, a storage medium and a program product, and the method comprises the steps: carrying out the intention recognition of an input text of a user, and obtaining an intention recognition result; based on the intention recognition result, matching is conducted in an intention list with the use permission of the user; if the matching is successful, calling a first target tool associated with the intention recognition result to execute a corresponding service, and returning a service execution result; and if the matching fails, correcting the intention recognition result, calling a second target tool associated with the corrected intention recognition result to execute a corresponding service, and returning a service execution result. According to the dialogue management method provided by the invention, an intention correction mechanism is introduced, the misrecognition rate caused by multiple intentions and similar intentions is effectively reduced, and meanwhile, the accuracy of business process execution is improved, so that the overall reliability of the system and the user experience are improved, and efficient business execution in a complex environment is ensured.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a dialogue management method, system, electronic device, storage medium and program product. Background Art

[0002] With the rapid development of artificial intelligence and natural language processing technology, large language models have demonstrated excellent capabilities in understanding and generating natural language. These models are trained with massive amounts of data and are able to capture subtle semantics and contextual relationships in language, giving them a significant advantage in intent recognition tasks.

[0003] In actual applications, multiple intents may be associated with user input text, and the semantic distance between different intents is relatively close. For example, the two intents of device control and obtaining device information on the network are often confused, resulting in increased complexity in intent recognition and the risk of misidentification. Traditional intent recognition methods often rely on basic rules or shallow learning models, which make it difficult to achieve high-accuracy recognition in multiple intents and scenarios, resulting in the information obtained by users not being what they actually need, resulting in poor user experience and affecting system reliability. Summary of the invention

[0004] The present invention provides a dialogue management method, system, electronic device, storage medium and program product, which are used to solve the problem of poor reliability of the dialogue management system in the prior art.

[0005] In a first aspect, the present invention provides a dialog management method, comprising: Perform intent recognition on the user's input text and obtain intent recognition results; Based on the intention recognition result, matching is performed in a list of intentions that the user has permission to use; If the match is successful, the first target tool associated with the intention recognition result is called to execute the corresponding business, and the business execution result is returned; If the match fails, the intent recognition result is corrected, the second target tool associated with the corrected intent recognition result is called to execute the corresponding business, and the business execution result is returned.

[0006] In one embodiment, the list of intentions that the user has permission to use is determined in the following manner: Obtaining the user's identification number; Based on the identity identification number, a plurality of tools that the user has permission to use are obtained; any of the tools is associated with an intent; Based on the intent associated with each tool, a list of intents that the user has permission to use is generated.

[0007] In one embodiment, if the match is successful, the first target tool associated with the intention recognition result is called to execute the corresponding business, and the business execution result is returned, including: If the match is successful, calling the first target tool associated with the intention recognition result to execute the corresponding business according to the input text; If the execution is successful, the business execution result is returned; If the execution fails, the intention recognition result is corrected, and based on the corrected intention recognition result, a third target tool that the user has the permission to use is called to execute the corresponding business, and the business execution result is returned.

[0008] In one embodiment, if the execution is successful, the service execution result is returned, including: If the execution is successful, determining through the first target tool whether execution parameters are missing when executing the corresponding business; If the execution parameter is not missing, executing the corresponding business through the first target tool, generating a business execution result, and returning the business execution result to the user; If the execution parameters are missing, the user is prompted to input the execution parameters. After obtaining the user input parameters, the first target tool executes the corresponding business according to the user input parameters, generates a business execution result, and returns the business execution result to the user.

[0009] In one embodiment, if the match fails, the intent recognition result is corrected, the second target tool associated with the corrected intent recognition result is called to execute the corresponding business, and the business execution result is returned, including: If the match fails, the steps of performing intent recognition on the user's input text to obtain the intent recognition result are iteratively executed until the match succeeds or the preset number of iterations is reached; If the preset number of iterations is reached, insufficient authority information is returned, and the user is prompted to check the input text; If the match is successful, the second target tool associated with the latest intent recognition result is called to execute the corresponding business and return the business execution result.

[0010] In one embodiment, performing intent recognition on the user's input text to obtain an intent recognition result includes: Get the user's input text; The input text is input into a pre-trained large language model to obtain an intent recognition result output by the large language model.

[0011] In a second aspect, the present invention further provides a dialogue management system, comprising: The intention recognition module is used to recognize the intention of the user's input text and obtain the intention recognition result; An intention matching module, used for matching, based on the intention recognition result, a list of intentions that the user has permission to use; A first feedback module is used to call the first target tool associated with the intention recognition result to execute the corresponding business if the match is successful, and return the business execution result; The second feedback module is used to correct the intention recognition result if the match fails, call the second target tool associated with the corrected intention recognition result to execute the corresponding business, and return the business execution result.

[0012] In a third aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above-described dialogue management methods when executing the program.

[0013] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of any of the above-mentioned dialog management methods are implemented.

[0014] In a fifth aspect, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by the processor, it implements the steps of any of the above-mentioned dialog management methods.

[0015] The dialogue management method, system, electronic device, storage medium and program product provided by the present invention determine the user's intention based on the user's input text, and then match the recognized intention with the intention list that the user has the permission to use, thereby ensuring that the user has the authority required to execute the recognized intention, and automatically calling the tool associated with the recognized intention to execute the corresponding business. In addition, in the event of a matching failure, a correction mechanism is introduced to dynamically adjust the intent recognition result and reassign it to other tools associated with the recognized intent to execute the corresponding business, and finally accurately return the business execution result actually required by the user. The use of this correction mechanism effectively reduces the misrecognition rate caused by multiple intentions and similar intentions, while improving the accuracy of business process execution, thereby improving the overall reliability of the system and user experience, and ensuring efficient business execution in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 This is one of the flow charts of the dialogue management method provided by the present invention.

[0018] Figure 2 This is the second flow chart of the dialogue management method provided by the present invention.

[0019] Figure 3 It is a structural diagram of the dialogue management system provided by the present invention.

[0020] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The terms "first", "second", etc. in the present invention are used to distinguish similar objects, rather than to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention can be implemented in an order other than those illustrated or described herein.

[0023] Combine the following Figure 1-Figure 4 The present invention describes the conversation management method, system, electronic device, storage medium and program product provided by the present invention.

[0024] It should be noted that the dialogue management method provided in the embodiment of the present invention is implemented based on the dialogue management system. The dialogue management method provided by the present invention has a high degree of multi-scenario adaptability and can be extended to a variety of different application scenarios to meet the intention recognition and execution requirements in different environments. The wide applicability enables the system to be deployed in a variety of industries and applications, enhancing its market competitiveness and application value. This embodiment is suitable for a variety of property management application scenarios such as equipment control, equipment query and business analysis. The dialogue management method first performs a deep multi-intention analysis on the user input text, and can efficiently and accurately parse complex and semantically similar user instructions. A dynamic permission matching mechanism is also designed to verify user permissions in real time to ensure the legality and security of the operation, and an intelligent error handling and correction mechanism is introduced. The intention is automatically re-analyzed within a limited number of retries. By cyclically optimizing the user interaction process, it can provide timely feedback and guide the user to re-enter when the intention recognition or tool execution fails, significantly reducing the misrecognition rate and improving the reliability of the system and user experience.

[0025] Therefore, the embodiment of the present invention describes the dialog management method by taking the dialog management system in the dialog management system as an execution subject as an example.

[0026] Combination Figure 1 , Figure 1 This is one of the flow charts of the dialogue management method provided by the present invention.

[0027] like Figure 1 As shown, the method includes the following: Step 101: Perform intent recognition on the user's input text to obtain an intent recognition result; Step 102: Based on the intention recognition result, matching is performed in the intention list that the user has permission to use; Step 103: If the match is successful, the first target tool associated with the intention recognition result is called to execute the corresponding business, and the business execution result is returned; Step 104: If the match fails, the intent recognition result is corrected, the second target tool associated with the corrected intent recognition result is called to execute the corresponding business, and the business execution result is returned.

[0028] Specifically, when users need to handle various property management services such as equipment control, equipment query and business analysis, they can enter natural language instructions in the interactive page of the dialogue management system. The dialogue management system receives the user's input text through the user interaction module. The dialogue management system is a very useful tool in property management and even more application scenarios. It can enhance user experience, simplify operating procedures, and improve management efficiency.

[0029] For property management services involving equipment control, the dialogue management system can help users control equipment through natural language commands, such as adjusting air-conditioning temperature, turning lights on and off, starting or shutting down security systems, etc. This interaction method is more intuitive and convenient than traditional buttons or touch screens. For property management services involving equipment query, users may need to query equipment status, fault history, maintenance records and other information. The dialogue management system can understand the user's query request and provide corresponding information or perform related operations. For property management services involving business analysis, the dialogue management system can generate real-time property business reports for users, analyze potential business problems, etc., to help users understand property operations and optimize service processes.

[0030] While accepting the user's input text, the dialogue management system also reads the current user's permission information. Different users, that is, different property managers, may have different property management powers because they are responsible for different businesses. Therefore, the system will pre-set the permission information of each property manager. The specific permission information refers to the tools that the property manager has the authority to use, and each tool can perform a business, and a business corresponds to an intent. Therefore, each tool is associated with an intent. Therefore, based on the current user's permission information, you can get a list of intents that the current user has permission to use.

[0031] When the dialog management system obtains the list of intents that the current user has permission to use, it will also simultaneously use the model's semantic understanding capabilities to identify the intent of the user's input text and obtain the intent recognition result output by the model.

[0032] Furthermore, the dialog management system matches the intent list that the user has permission to use based on the intent recognition result. Specifically, the intent list that the user has permission to use includes multiple target intents, and the intent list is generated by the name and description of each target intent. Therefore, in the matching process, the intent recognition result output by the model needs to be matched with the name of each target intent in the intent list. A strict match is required to be considered a successful match, otherwise the match fails. For example, the intent recognition result output by the model is "Business Analysis", but there is only "Business Analysis" in the intent list, which means that the intent recognition result output by the model is not output in the predetermined format and cannot be matched to the "Business Analysis" in the intent list, then the match is considered a failure. If the intent recognition result output by the model is "Business Analysis" and can be matched to the "Business Analysis" in the intent list, then the match is considered a success.

[0033] In one case, when the intent recognition result successfully identifies a target intent in the intent list, the dialog management system calls the first target tool associated with the intent recognition result to execute the business it is responsible for, generates the business execution result through the first target tool, and feeds back the business execution result to the user through the interactive page.

[0034] In another case, when the intent recognition result cannot identify any target intent in the intent list, it may be that the model does not output the result in the predetermined format, or it may be that the model recognition is wrong. At this time, the intent correction mechanism is introduced. The dialogue management system iteratively corrects the intent recognition result within a limited number of times until the corrected intent recognition result is successfully matched in the intent list. The second target tool associated with the corrected intent recognition result is called to execute the business it is responsible for, and the business execution result is generated through the second target tool. The business execution result is fed back to the user through the interactive page.

[0035] The dialogue management method provided by the present invention determines the user intention based on the user's input text, and then matches the recognized intention with the intention list that the user has the permission to use, thereby ensuring that the user has the authority required to execute the recognized intention, and automatically calls the tool associated with the recognized intention to execute the corresponding business. In addition, in the event of a matching failure, a correction mechanism is introduced to dynamically adjust the intent recognition result and reassign it to other tools associated with the recognized intent to execute the corresponding business, and finally accurately returns the business execution result actually required by the user. The use of this correction mechanism effectively reduces the misrecognition rate caused by multiple intentions and similar intentions, while improving the accuracy of business process execution, thereby improving the overall reliability of the system and user experience, and ensuring efficient business execution in complex environments.

[0036] In some embodiments, based on step 101, performing intent recognition on the user's input text to obtain an intent recognition result includes: Get the user's input text; The input text is input into a pre-trained large language model to obtain an intent recognition result output by the large language model.

[0037] Specifically, the dialog management system receives the user's input text through the user interaction module. This process can also be extended to receiving the voice information sent by the user, performing voice recognition on the voice information, generating the input text, and then performing subsequent analysis.

[0038] After obtaining the user's input text, the dialogue management system will preprocess the input text, which may include removing irrelevant characters, unifying the format, segmenting words, and converting it into a representation that the model can understand.

[0039] Furthermore, the dialog management system inputs the input text into a pre-trained large language model to obtain the intent recognition result output by the large language model. This large language model can be a neural network based on deep learning, such as the Transformer architecture, which can understand and generate natural language. Of course, it is not limited to this, but advanced large language models will be used first for accurate intent recognition. When training a large language model, a large number of input text samples will be used, classified and labeled to obtain the intent label of each text sample, and then pre-trained according to the training sample data and its label data to obtain a pre-trained large language model. The model can parse multiple intents that may be contained in the user input, and finally output the intent that best matches the user input.

[0040] The embodiment of the present invention utilizes an advanced large language model to pre-identify and analyze multiple intents during the user input stage, and can efficiently parse multiple semantically similar intents contained in the user input. Compared with traditional rule-based or shallow learning models, deep semantic understanding significantly improves the accuracy and robustness of intent recognition, especially when processing complex and diverse user instructions.

[0041] In some embodiments, based on step 102, the list of intentions for which the user has permission to use is determined in the following manner: Obtaining the user's identification number; Based on the identity identification number, a plurality of tools that the user has permission to use are obtained; any of the tools is associated with an intent; Based on the intent associated with each tool, a list of intents that the user has permission to use is generated.

[0042] Specifically, each property manager has a dedicated account on the dialogue management system. The property manager needs to log in to the dialogue management system through the account and obtain business execution feedback through natural language instructions. When the current user sends a natural language instruction, the dialogue management system can obtain the current user's identity identification number.

[0043] Furthermore, the dialogue management system obtains multiple tools that the current user has permission to use based on the current user's identity identification number. As described above, any tool is associated with an intent. Therefore, the dialogue management system can generate a list of intents that the current user has permission to use based on the intent associated with each tool.

[0044] The embodiment of the present invention obtains the user's identity identification number, and based on this identity identification number, identifies the tools that the current user has the right to use and their associated intentions, and generates a list of intentions that the current user has the right to use, and then can match the identified intentions in the intention list to ensure that the user has the authority required to execute the identified intentions, thereby improving the system's service targeting and user experience, ensuring that users can access functions that match their authority during the interaction process, and enhancing the security and efficiency of the system.

[0045] In some embodiments, based on step 103, if the match is successful, the first target tool associated with the intention recognition result is called to execute the corresponding business, and the business execution result is returned, including: If the match is successful, calling the first target tool associated with the intention recognition result to execute the corresponding business according to the input text; If the execution is successful, the business execution result is returned; If the execution fails, the intention recognition result is corrected, and based on the corrected intention recognition result, a third target tool that the user has the permission to use is called to execute the corresponding business, and the business execution result is returned.

[0046] Specifically, if the intent recognition result successfully identifies a target intent in the intent list, the dialog management system calls the first target tool associated with the intent recognition result to execute the corresponding business according to the input text.

[0047] When the tool executes the corresponding business, errors may occur during the execution process, resulting in execution failure. Most of the time, the parameters captured from the input text do not conform to the established format of the business to be executed by the tool. For example, in the device query, some other numbers are captured from the user's input as the room number, but there is actually no such room number. At this time, the intent correction mechanism will also be introduced.

[0048] In one case, the first target tool successfully executes the corresponding business, the dialog management system generates a business execution result, and feeds back the business execution result to the user through an interactive page.

[0049] In another case, an error occurs when the first target tool executes the corresponding business. The dialog management system iteratively corrects the intent recognition result within a limited number of times until the corrected intent recognition result is successfully matched in the intent list, and calls the third target tool associated with the corrected intent recognition result to execute the business it is responsible for, generates the business execution result through the third target tool, and feeds back the business execution result to the user through the interactive page.

[0050] In the event of a matching failure, the embodiment of the present invention introduces a correction mechanism to dynamically adjust the intent recognition results and reassign them to other tools associated with the recognized intent to execute the corresponding business, ultimately accurately returning the business execution results actually required by the user. The use of this correction mechanism effectively reduces the misrecognition rate caused by multiple intents and similar intents, while improving the accuracy of business process execution, thereby improving the overall reliability of the system and user experience, and ensuring efficient business execution in complex environments.

[0051] According to the above content, if the execution is successful, the business execution result is returned, including: If the execution is successful, determining through the first target tool whether execution parameters are missing when executing the corresponding business; If the execution parameter is not missing, executing the corresponding business through the first target tool, generating a business execution result, and returning the business execution result to the user; If the execution parameters are missing, the user is prompted to input the execution parameters. After obtaining the user input parameters, the first target tool executes the corresponding business according to the user input parameters, generates a business execution result, and returns the business execution result to the user.

[0052] Specifically, if the first target tool successfully executes the corresponding business, the dialogue management system determines whether the execution parameters are missing when executing the corresponding business through the first target tool. For example, when the current user wants to output a business analysis report through the dialogue management system, the tool requires the user to further input specific time parameters, and then generates the business analysis report within this time based on the input time parameters. For another example, when the current user wants to query certain information on the Internet through the dialogue management system, the tool does not require the user to further input relevant specific parameters, and can directly call the interface to query.

[0053] In one case, if the first target tool determines that there is no lack of execution parameters when executing its corresponding business, the dialog management system executes its corresponding business through the first target tool, generates a business execution result, and feeds back the business execution result to the user through an interactive page.

[0054] In another case, if the first target tool determines that the execution parameters are missing when executing its corresponding business, the dialogue management system prompts the user to enter the execution parameters required to execute the current business through a dialogue inquiry method. The user will enter the required execution parameters in the interactive page according to the system prompt. After obtaining the user input parameters, the dialogue management system executes the corresponding business according to the user input parameters and the user's original input text through the first target tool, generates the business execution results, and feeds back the business execution results to the user through the interactive page.

[0055] The embodiment of the present invention ensures the smooth execution of business processes by intelligently determining whether necessary parameters are missing during business execution and guiding users to input when necessary parameters are missing, thereby improving the system's automated processing capabilities and user interaction efficiency, achieving accurate return of business execution results, and optimizing the user interaction experience as a whole.

[0056] In some embodiments, based on step 104, if the match fails, the intent recognition result is corrected, the second target tool associated with the corrected intent recognition result is called to execute the corresponding business, and the business execution result is returned, including: If the match fails, the steps of performing intent recognition on the user's input text to obtain the intent recognition result are iteratively executed until the match succeeds or the preset number of iterations is reached; If the preset number of iterations is reached, insufficient authority information is returned, and the user is prompted to check the input text; If the match is successful, the second target tool associated with the latest intent recognition result is called to execute the corresponding business and return the business execution result.

[0057] Specifically, when the intent recognition result cannot identify any target intent in the intent list, the dialogue management system starts the intent correction mechanism, iteratively executes the steps of performing intent recognition on the user's input text, and obtains the intent recognition result until a match is successful or a preset number of iterations is reached.

[0058] If the preset number of iterations is reached, the dialogue management system will provide the user with insufficient authority feedback through the interactive interface and prompt the user to check the input text.

[0059] If the intent recognition result successfully matches the target intent in the intent list, the dialogue management system calls the second target tool associated with the latest intent recognition result to execute the corresponding business, generate the business execution result, and feedback the business execution result to the user through the interactive page.

[0060] For example, the system records the current number of retries and determines whether it is less than the preset maximum number of retries (such as 5 times). If the number of retries is less than 5: return to the "large model pre-intention analysis" step and re-analyze the intention. If the number of retries reaches 5: prompt the user that the authority is insufficient and terminate the process.

[0061] Similarly, through the above-mentioned intention correction mechanism, the intention correction mechanism introduced when the tool fails to execute can be determined, as follows.

[0062] If an error occurs when the first target tool executes the corresponding business, the dialog management system updates the intent list by deleting the target intent that is successfully matched by the intent recognition result in the intent list to ensure that invalid or unmatched intents are not used.

[0063] Furthermore, the dialog management system iteratively executes the steps of performing intent recognition on the user's input text to obtain intent recognition results until a match is successful or a preset number of iterations is reached.

[0064] If the preset number of iterations is reached, the dialogue management system prompts the user to check the input text through the interactive interface.

[0065] If the intent recognition result successfully matches the target intent in the updated intent list, the dialogue management system calls the third target tool associated with the latest intent recognition result to execute the corresponding business. If the third target tool successfully executes the corresponding business, the dialogue management system generates the business execution result and feeds back the business execution result to the user through the interactive page.

[0066] The embodiment of the present invention introduces an intelligent error detection and correction process based on the retry limit. When the intent recognition or operation execution fails, the system can automatically clear the current intent analysis results, record the number of retries, and re-analyze the intent until it succeeds or reaches the maximum number of retries, effectively reducing the misrecognition rate caused by multiple intentions and similar intentions, improving the overall reliability of the system and user experience, and ensuring efficient execution in complex environments. At the same time, by designing a loop mechanism, the system can promptly feedback problems and guide users to re-enter when intent recognition or execution fails, thereby continuously optimizing the user interaction experience, improving user satisfaction, reducing the frustration caused by misoperation or misrecognition, and enhancing the intelligence and humanity of the system.

[0067] Combination Figure 2 , Figure 2 This is the second flow chart of the dialogue management method provided by the present invention. Figure 2 The entire process of dialogue management of this embodiment is described.

[0068] Initialize the dialog management system.

[0069] After initialization, the dialog management system reads the user permissions and input text, and obtains a list of intents that the current user has permission to use through the user permissions.

[0070] Furthermore, the dialogue management system calls the large language model to perform open intent analysis to obtain intent recognition results.

[0071] Furthermore, the dialog management system matches the intent list that the user has permission to use based on the intent recognition result.

[0072] When matching the intent recognition results, if the match is successful, the dialog management system will call the first target tool associated with the intent recognition result to perform the corresponding business and return the business execution result; if the match fails, the dialog management system will correct the intent recognition result within a preset number of times until the match is successful or the preset number of iterations is reached. If the preset number of iterations is reached, it will return insufficient authority information and prompt the user to check the input text. If the match is successful, it will call the second target tool associated with the latest intent recognition result to execute the corresponding business and return the business execution result.

[0073] When the first target tool executes a business, if the execution is successful, the dialogue management system determines through the first target tool whether the execution parameters are missing when executing the corresponding business. If the execution parameters are not missing, the business operation is executed directly. If the execution parameters are missing, the user is prompted to enter the required execution parameters through dialogue polling, and the business operation is further executed according to the execution parameters through the first target tool; if the execution fails, the dialogue management system first deletes the previously matched target intent in the intent list to update the intent list, and corrects the intent recognition result through the latest intent list within a preset number of times until the match is successful or the preset number of iterations is reached. If the preset number of iterations is reached, the user is prompted to check the input text. If the execution is successful, the third target tool associated with the latest intent recognition result is called to execute the corresponding business and return the business execution result.

[0074] The dialog management system provided by the present invention is described below. The dialog management system described below and the dialog management method described above can be referenced to each other.

[0075] Reference Figure 3 , Figure 3 It is a structural diagram of the dialogue management system provided by the present invention.

[0076] The dialogue management system comprises: The intention recognition module 310 is used to perform intention recognition on the user's input text and obtain an intention recognition result.

[0077] The intention matching module 320 is used to match the intention list that the user has permission to use based on the intention recognition result.

[0078] The first feedback module 330 is used to call the first target tool associated with the intention recognition result to execute the corresponding business and return the business execution result if the match is successful.

[0079] The second feedback module 340 is used to correct the intention recognition result if the match fails, call the second target tool associated with the corrected intention recognition result to execute the corresponding business, and return the business execution result.

[0080] The dialogue management system provided by the present invention determines the user's intention based on the user's input text, and then matches the recognized intention with the intention list that the user has the permission to use, thereby ensuring that the user has the authority required to execute the recognized intention, and automatically calls the tool associated with the recognized intention to execute the corresponding business. In addition, in the event of a matching failure, a correction mechanism is introduced to dynamically adjust the intent recognition result and reassign it to other tools associated with the recognized intent to execute the corresponding business, and finally accurately returns the business execution result actually required by the user. The use of this correction mechanism effectively reduces the misrecognition rate caused by multiple intentions and similar intentions, while improving the accuracy of business process execution, thereby improving the overall reliability of the system and user experience, and ensuring efficient business execution in complex environments.

[0081] Furthermore, the intention recognition module 310 is also used for: Get the user's input text; The input text is input into a pre-trained large language model to obtain an intent recognition result output by the large language model.

[0082] Furthermore, the dialogue management system is also used for: Obtaining the user's identification number; Based on the identity identification number, a plurality of tools that the user has permission to use are obtained; any of the tools is associated with an intent; Based on the intent associated with each tool, a list of intents that the user has permission to use is generated.

[0083] Furthermore, the first feedback module 330 is further configured to: If the match is successful, calling the first target tool associated with the intention recognition result to execute the corresponding business according to the input text; If the execution is successful, the business execution result is returned; If the execution fails, the intention recognition result is corrected, and based on the corrected intention recognition result, a third target tool that the user has the permission to use is called to execute the corresponding business, and the business execution result is returned.

[0084] Furthermore, the first feedback module 330 is further configured to: If the execution is successful, determining through the first target tool whether execution parameters are missing when executing the corresponding business; If the execution parameter is not missing, executing the corresponding business through the first target tool, generating a business execution result, and returning the business execution result to the user; If the execution parameters are missing, the user is prompted to input the execution parameters. After obtaining the user input parameters, the first target tool executes the corresponding business according to the user input parameters, generates a business execution result, and returns the business execution result to the user.

[0085] Furthermore, the second feedback module 340 is further configured to: If the match fails, the steps of performing intent recognition on the user's input text to obtain the intent recognition result are iteratively executed until the match succeeds or the preset number of iterations is reached; If the preset number of iterations is reached, insufficient authority information is returned, and the user is prompted to check the input text; If the match is successful, the second target tool associated with the latest intent recognition result is called to execute the corresponding business and return the business execution result.

[0086] It should be noted that the dialogue management system provided by the present invention can execute the dialogue management method described in any of the above embodiments during specific operation, which will not be elaborated in this embodiment.

[0087] Figure 4 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the dialogue management method, which includes: performing intent recognition on the user's input text to obtain an intent recognition result; based on the intent recognition result, matching in the intent list that the user has permission to use; if the match is successful, calling the first target tool associated with the intent recognition result to execute the corresponding business, and returning the business execution result; if the match fails, correcting the intent recognition result, calling the second target tool associated with the corrected intent recognition result to execute the corresponding business, and returning the business execution result.

[0088] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0089] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the dialogue management method provided by the above-mentioned embodiments, and the method includes: performing intent recognition on the user's input text to obtain an intent recognition result; based on the intent recognition result, matching in a list of intents that the user has permission to use; if the match is successful, calling the first target tool associated with the intent recognition result to execute the corresponding business, and returning the business execution result; if the match fails, correcting the intent recognition result, calling the second target tool associated with the corrected intent recognition result to execute the corresponding business, and returning the business execution result.

[0090] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the dialogue management method provided by the above-mentioned embodiments, the method comprising: performing intent recognition on the user's input text to obtain an intent recognition result; based on the intent recognition result, matching in a list of intents that the user has permission to use; if the match is successful, calling the first target tool associated with the intent recognition result to execute the corresponding business, and returning the business execution result; if the match fails, correcting the intent recognition result, calling the second target tool associated with the corrected intent recognition result to execute the corresponding business, and returning the business execution result.

[0091] The system embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. Those of ordinary skill in the art may understand and implement the present embodiment without creative effort.

[0092] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dialogue management method, characterized in that: include: Perform intent recognition on the user's input text and obtain intent recognition results; Based on the intention recognition result, matching is performed in a list of intentions that the user has permission to use; If the match is successful, the first target tool associated with the intention recognition result is called to execute the corresponding business, and the business execution result is returned; If the match fails, the intent recognition result is corrected, the second target tool associated with the corrected intent recognition result is called to execute the corresponding business, and the business execution result is returned.

2. The method for managing a conversation according to claim 1, characterized in that: The list of intents that the user has permission to use is determined in the following way: Obtaining the user's identification number; Based on the identity identification number, a plurality of tools that the user has permission to use are obtained; any of the tools is associated with an intent; Based on the intent associated with each tool, a list of intents that the user has permission to use is generated.

3. The method for managing a conversation according to claim 1, characterized in that: If the match is successful, the first target tool associated with the intention recognition result is called to execute the corresponding business, and the business execution result is returned, including: If the match is successful, calling the first target tool associated with the intention recognition result to execute the corresponding business according to the input text; If the execution is successful, the business execution result is returned; If the execution fails, the intention recognition result is corrected, and based on the corrected intention recognition result, a third target tool that the user has the permission to use is called to execute the corresponding business, and the business execution result is returned.

4. The method for managing a conversation according to claim 3, characterized in that: If the execution is successful, the business execution result is returned, including: If the execution is successful, determining through the first target tool whether execution parameters are missing when executing the corresponding business; If the execution parameter is not missing, executing the corresponding business through the first target tool, generating a business execution result, and returning the business execution result to the user; If the execution parameters are missing, the user is prompted to input the execution parameters. After obtaining the user input parameters, the first target tool executes the corresponding business according to the user input parameters, generates a business execution result, and returns the business execution result to the user.

5. The method for managing a conversation according to claim 4, characterized in that: If the match fails, the intent recognition result is corrected, the second target tool associated with the corrected intent recognition result is called to execute the corresponding business, and the business execution result is returned, including: If the match fails, the steps of performing intent recognition on the user's input text to obtain the intent recognition result are iteratively executed until the match succeeds or the preset number of iterations is reached; If the preset number of iterations is reached, insufficient authority information is returned, and the user is prompted to check the input text; If the match is successful, the second target tool associated with the latest intent recognition result is called to execute the corresponding business and return the business execution result.

6. The method for managing a conversation according to any one of claims 1 to 5, characterized in that: The performing intent recognition on the user's input text to obtain the intent recognition result includes: Get the user's input text; The input text is input into a pre-trained large language model to obtain an intent recognition result output by the large language model.

7. A dialogue management system, characterized in that: include: The intention recognition module is used to recognize the intention of the user's input text and obtain the intention recognition result; An intention matching module, used for matching, based on the intention recognition result, a list of intentions that the user has permission to use; A first feedback module is used to call the first target tool associated with the intention recognition result to execute the corresponding business if the match is successful, and return the business execution result; The second feedback module is used to correct the intention recognition result if the match fails, call the second target tool associated with the corrected intention recognition result to execute the corresponding business, and return the business execution result.

8. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the dialogue management method according to any one of claims 1 to 6 are implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the dialogue management method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by a processor, the steps of the dialogue management method according to any one of claims 1 to 6 are implemented.

Images