Intelligent dialogue method and system

Abstract

The invention provides an intelligent dialogue method and system, and belongs to the technical field of artificial intelligence. The method comprises the steps of obtaining local feedback data based on question key data and a local knowledge base; wherein the question key data is obtained by analyzing the captured question information; performing network retrieval according to the question key data to obtain network feedback data; when it is determined that a similarity test result between the local feedback data and the network feedback data meets a preset condition, question answer data is generated based on the local feedback data and the network feedback data; and inputting the question answer data into the large language model for adjustment to obtain feedback information for answering the question information. Therefore, when the large language model is applied to some special industry fields, the data security of the local knowledge base can be fully ensured, the situation of confidential information leakage is avoided, and the purpose of analyzing and processing user problems in some special industry fields is achieved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an intelligent dialogue method, an intelligent dialogue system, a machine-readable storage medium and an electronic device. Background Art

[0002] In the application process of large language models, commonly used large language models are trained through a large amount of general knowledge text data and powerful computing resources. By parsing the questions input by the user, the corresponding answers are generated from the knowledge structure of the large language model. Although the intelligent dialogue technology based on the large language model can communicate with users, solve problems for users or assist users in production, the knowledge structure of common open source large language models does not contain or contains less knowledge in some special industry fields, so it is impossible to parse and process user problems involving some special industry fields. In addition, due to technical confidentiality and other reasons, the knowledge data held by some users cannot be made public, resulting in the inability to be directly applied to the existing large language model.

[0003] Therefore, when applying large language models to some special industry fields, how to ensure the data security of the local knowledge base in these special industry fields to avoid the leakage of confidential information while being able to parse and process user questions in these special industry fields is an issue that needs to be solved urgently. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide an intelligent dialogue method and system to at least solve the above-mentioned problem of how to ensure the data security of the local knowledge base in some special industry fields when applying a large language model to some special industry fields, so as to avoid the leakage of confidential information and to parse and process user questions in some special industry fields.

[0005] In order to achieve the above object, the present invention provides an intelligent dialogue method in a first aspect, comprising:

[0006] Based on the key data of the question and the local knowledge base, local feedback data is obtained; wherein the key data of the question is obtained by parsing the captured question information;

[0007] Conduct network search based on the key data of the question to obtain network feedback data;

[0008] When it is determined that the similarity test result between the local feedback data and the network feedback data meets the preset conditions, generating question answer data based on the local feedback data and the network feedback data;

[0009] The question answer data is input into the large language model for adjustment to obtain feedback information used to answer the question information.

[0010] Optionally, the local knowledge base is constructed based on multiple local knowledge files;

[0011] Based on the key data of the questions and the local knowledge base, local feedback data is obtained, including:

[0012] Determine where the key data for the question appears in each local knowledge file;

[0013] When the position where the question key data appears is in the title of the local knowledge file, assigning a first weight corresponding to the question information to the corresponding local knowledge file;

[0014] When the location where the question key data appears is in the body of the local knowledge file, assigning a second weight corresponding to the question information to the corresponding local knowledge file;

[0015] Determine the frequency of occurrence of the key data of the question in each local knowledge file, and assign a third weight corresponding to the question information to the corresponding local knowledge file based on the frequency of occurrence of the key data of the question in the local knowledge file; wherein the weight of the first weight is greater than the weight of the second weight, and the weight of the second weight is greater than the weight of the third weight;

[0016] Based on the first weight, the second weight, and the third weight, each local knowledge file related to the key data of the question is sorted, and each local knowledge file related to the key data of the question is parsed to extract the content containing the key data of the question in each local knowledge file;

[0017] Based on the content of each local knowledge file containing the key data of the question, local feedback data is generated.

[0018] Optionally, the network search is performed based on the key data of the question to obtain network feedback data, including:

[0019] Based on the network search engine and the key data of the questions, the network search is carried out in the public database to obtain the network feedback data;

[0020] The data format of the network feedback data is the same as the data format of the local feedback data, and the data format of the network feedback data includes a picture format and / or a text format.

[0021] Optionally, before determining that the similarity check result between the local feedback data and the network feedback data meets a preset condition, the method further includes:

[0022] A similarity check is performed on the local feedback data and the network feedback data. If the similarity check result does not meet the preset conditions, the network feedback data is updated and adjusted until the similarity check result meets the preset conditions.

[0023] Optionally, if the similarity check result does not meet the preset condition, the network feedback data is updated and adjusted, including:

[0024] If the similarity test result does not meet the preset conditions, the network feedback data will be screened for citations to trace the initial release time and initial release channel of the network feedback data;

[0025] Conduct reliability analysis on the initial release channels of network feedback data to obtain channel reliability analysis data;

[0026] If the initial release time of the network feedback data is earlier than the collection time of the local feedback data and / or the channel reliability analysis data is less than the preset reliability data threshold, the network search is performed again to generate new network feedback data;

[0027] If the initial release time of the network feedback data is later than the collection time of the local feedback data and the channel reliability analysis data is not less than the preset reliability data threshold, data enhancement is performed on the network feedback data and an enhancement label is added to the network feedback data.

[0028] Optionally, after obtaining the network feedback data, the method further includes:

[0029] Update the local knowledge base based on network feedback data.

[0030] Optionally, the question information may be in text and / or voice form;

[0031] The parsing rules for the above question information include:

[0032] Determine the form of the question information, recognize and convert the question information in voice form to obtain the corresponding question information in text form;

[0033] Perform text analysis on the question information in text form to obtain key data of the question.

[0034] A second aspect of the present invention provides an intelligent dialogue system, comprising:

[0035] A local feedback data obtaining module is used to obtain local feedback data based on question key data and a local knowledge base; wherein the question key data is obtained by parsing the captured question information;

[0036] A network feedback data acquisition module is used to perform network search based on the key data of the question and obtain network feedback data;

[0037] A question answer data generating module, configured to generate question answer data based on the local feedback data and the network feedback data when it is determined that the similarity test result between the local feedback data and the network feedback data meets a preset condition;

[0038] The model analysis module is used to input the question answer data into the large language model for adjustment to obtain feedback information for answering the question information.

[0039] In a third aspect of the present invention, a machine-readable storage medium is provided, on which instructions are stored, and when the instructions are executed by a processor, the processor is configured to execute the above-mentioned intelligent dialogue method.

[0040] In a fourth aspect of the present invention, an electronic device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the intelligent dialogue method when executing the computer program.

[0041] Through the above technical solution, an intelligent dialogue method and system are provided. According to the key data of the question and the local knowledge base, local feedback data is obtained, and network retrieval is performed according to the key data of the question to obtain network feedback data. When it is determined that the similarity test results between the local feedback data and the network feedback data meet the preset conditions, question answer data are generated according to the local feedback data and the network feedback data, and the question answer data are input into the large language model for adjustment to obtain feedback information for answering the question information. The method and system use the local knowledge base generated based on the local knowledge file and the network search engine to generate feedback information corresponding to the question information. The data of the local knowledge base is only stored locally, which can fully guarantee the data security of the local knowledge base and avoid the leakage of confidential information. And the conversion of the question answer data to the feedback information is completed by the large language model, which can help users to obtain the desired results more easily and more easily. Therefore, when the large language model is applied to some special industry fields, the data security of the local knowledge base can be fully guaranteed to avoid the leakage of confidential information, thereby achieving the purpose of parsing and processing user questions in some special industry fields.

[0042] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0044] Figure 1 is a flow chart of an intelligent dialogue method provided by one embodiment of the present invention;

[0045] Figure 2is a block diagram of an intelligent dialogue system provided by one embodiment of the present invention;

[0046] Figure 3 It is a schematic diagram of the structure of an electronic device provided by a preferred embodiment of the present invention.

[0047] Description of Reference Numerals

[0048] 10 - electronic device, 100 - processor, 101 - memory, 102 - computer program. DETAILED DESCRIPTION

[0049] The specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present invention, and is not used to limit the present invention.

[0050] Figure 1 FIG. 1 is a flow chart of an intelligent dialogue method provided by an embodiment of the present invention. Figure 1 As shown, an embodiment of the present invention provides an intelligent dialogue method, including:

[0051] S110: obtaining local feedback data based on the question key data and the local knowledge base; wherein the question key data is obtained by parsing the captured question information;

[0052] In some implementations of the present embodiment, the local knowledge base is constructed based on a plurality of local knowledge files; the local feedback data is obtained based on the question key data and the local knowledge base, including: determining the position where the question key data appears in each local knowledge file; when the question key data appears in the title of the local knowledge file, assigning a first weight corresponding to the question information to the corresponding local knowledge file; when the question key data appears in the body of the local knowledge file, assigning a second weight corresponding to the question information to the corresponding local knowledge file; determining the frequency of occurrence of the question key data in each local knowledge file, and assigning a third weight corresponding to the question information to the corresponding local knowledge file based on the frequency of occurrence of the question key data in the local knowledge file; wherein the weight of the first weight is greater than the weight of the second weight, and the weight of the second weight is greater than the weight of the third weight; sorting the local knowledge files related to the question key data based on the first weight, the second weight and the third weight, and parsing the local knowledge files related to the question key data to extract the content containing the question key data in each local knowledge file; generating local feedback data based on the content containing the question key data in each local knowledge file.

[0053] Specifically, the local knowledge base is searched according to the key data of the question to obtain local feedback data. The local knowledge base is built based on the existing local knowledge files. There are two main forms of knowledge files. The first is a structured document form, such as plain text documents, electronic publishing documents, and portable file format documents. The second is an unstructured data form, such as electronic icons and databases. After obtaining the key data of the question, local feedback data is generated based on the matching degree between the key data of the question and the local knowledge base. Specifically, the position where the key data of the question appears in the local knowledge file is determined. When the key data of the question appears in the title of the local knowledge file, the local knowledge file is assigned a first weight corresponding to the current question information. When the key data of the question appears in the body of the local knowledge file, the local knowledge file is assigned a second weight corresponding to the current question information. Then, the frequency of occurrence of the key data of the question in the local knowledge file is determined, and the local knowledge file is assigned a third weight corresponding to the current question information based on the frequency. The weights of the first weight, the second weight, and the third weight gradually decrease. Then, based on the first weight, the second weight, and the third weight, several local knowledge files related to the key data of the question are sorted. Then, these local knowledge files related to the key data of the question are parsed, the part containing the key data of the question is extracted, and the language is reorganized to form local feedback data. Reorganizing the language based on the existing text belongs to the prior art in this field and will not be repeated here.

[0054] Among them, the method of sorting several local knowledge files related to the key data of the question based on the first weight, the second weight and the third weight includes: for each local knowledge file, all the weights of the local knowledge file (at least including one or more of the first weight, the second weight and the third weight) are summed up to obtain the corresponding total weight, and then sorting is performed based on the total weight of each local knowledge file.

[0055] S120: Perform network search based on the key data of the question to obtain network feedback data;

[0056] In some implementations of the present embodiment, the above-mentioned network search based on the question key data to obtain the network feedback data includes: based on the network search engine and the question key data, a network search is performed in a public database to obtain the network feedback data; wherein the data format of the network feedback data is the same as the data format of the local feedback data, and the data format of the network feedback data includes picture format and / or text format.

[0057] In some implementations of this embodiment, when the data form of the network feedback data is in the form of a picture, it is necessary to perform picture recognition conversion on the network feedback data in the form of a picture to obtain the corresponding text information. Wherein, the step of performing picture recognition conversion on the network feedback data in the form of a picture includes: inputting the network feedback data in the form of a picture into a pre-trained text picture recognition model to perform recognition processing on the picture area containing text information to obtain a recognition result. When there is a picture area containing text information in the recognition result, the picture area containing text information is selected as the target picture; wherein the text picture recognition model is obtained by modeling and training based on a deep neural network algorithm and / or a machine learning algorithm. Determine the text information area and the picture information area in each target picture. Extract the picture text information from the text information area, wherein the extraction of the picture text information is based on extracting text lines; identify the picture body from the picture information area; determine whether the picture text information and the picture body match; if it is determined to match, determine that the picture text information is the text information corresponding to the network feedback data in the form of a picture. Thus, the picture recognition of the network feedback data in the form of a picture is completed.

[0058] S130: When it is determined that the similarity test result between the local feedback data and the network feedback data meets a preset condition, generating question answer data based on the local feedback data and the network feedback data;

[0059] Specifically, after obtaining accurate and reliable local feedback data and network feedback data, the local feedback data and the network feedback data are combined to generate question answer data. The method for generating question answer data is prior art and will not be described in detail here.

[0060] S140: Input the question answer data into the large language model for adjustment to obtain feedback information for answering the question information.

[0061] Among them, the large language model makes oral adjustments or accuracy adjustments to the question and answer data.

[0062] Specifically, the question answer data generated based on the local feedback data and the network feedback data are input into the large language model, and the question answer data is adjusted using the large language model to generate feedback information for answering the question information. The feedback information can correspond to the answer to the question information.

[0063] In the above implementation process, the method obtains local feedback data according to the key data of the question and the local knowledge base, and performs network retrieval according to the key data of the question to obtain network feedback data. When it is determined that the similarity test results between the local feedback data and the network feedback data meet the preset conditions, the question answer data is generated according to the local feedback data and the network feedback data, and the question answer data is input into the large language model for adjustment to obtain feedback information for answering the question information. The method uses the local knowledge base generated based on the local knowledge file and the network search engine to generate feedback information corresponding to the question information. The data of the local knowledge base is only stored locally, which can fully guarantee the data security of the local knowledge base and avoid the leakage of confidential information. And the conversion of the question answer data to the feedback information is completed by the large language model, which can help users to obtain the desired results more easily and more easily. Therefore, when the large language model is applied to some special industry fields, the data security of the local knowledge base can be fully guaranteed to avoid the leakage of confidential information, thereby achieving the purpose of parsing and processing user questions in some special industry fields.

[0064] In some implementations of this embodiment, before determining that the similarity check result between the local feedback data and the network feedback data meets the preset conditions, the method also includes: performing a similarity check on the local feedback data and the network feedback data, and if the similarity check result does not meet the preset conditions, updating and adjusting the network feedback data until the similarity check result meets the preset conditions.

[0065] Specifically, a similarity check is performed on the local feedback data and the network feedback data. When the similarity check result does not meet the preset conditions, a reference screening is performed on the network feedback data, and the network feedback data is corrected or strengthened according to the reference screening result, and a similarity check is performed again until the similarity check result meets the preset conditions.

[0066] In some implementations of the present embodiment, if the similarity check result does not meet the preset conditions, the network feedback data is updated and adjusted, including: if the similarity check result does not meet the preset conditions, the network feedback data is referenced and screened to trace the initial release time and initial release channel of the network feedback data; reliability analysis is performed on the initial release channel of the network feedback data to obtain channel reliability analysis data; if the initial release time of the network feedback data is earlier than the collection time of the local feedback data and / or the channel reliability analysis data is less than the preset reliability data threshold, network retrieval is performed again to generate new network feedback data; if the initial release time of the network feedback data is later than the collection time of the local feedback data and the channel reliability analysis data is not less than the preset reliability data threshold, data enhancement is performed on the network feedback data, and an enhancement label is added to the network feedback data.

[0067] Specifically, because the number of results of network searches is huge and the accuracy of network search results varies, network feedback data is prone to errors. The local knowledge base is constructed based on knowledge files that are manually screened and determined, and the accuracy of its content is fully guaranteed. However, considering that there is a delay in the update of the local knowledge base, it may happen that the local feedback data lags behind the network feedback data. In order to fully obtain accurate results and provide feedback to users, after obtaining local feedback data and network feedback data, this method performs a similarity test on the local feedback data and the network feedback data, and compares the local feedback data with the network feedback data. If the difference between the network feedback data and the local feedback data is too large, or even contradictory, the network feedback data needs to be updated. Specifically, when the similarity test result does not meet the preset conditions, it means that the accuracy of the network feedback data may be poor. At this time, the network feedback data is first referenced and screened, that is, the historical release time and channel of the network feedback data are traced. If the time of the initial release of the network feedback data is earlier than the collection time of the local feedback data, it means that the current network feedback data may be lagging. At this time, the network feedback data needs to be updated, that is, the network search is performed again to generate new network feedback data; if the reliability of the release channel of the network feedback data is poor when it is first released, it is also necessary to search the network again to generate new network feedback data; if the release time of the network feedback data is later than the collection time of the local feedback data when it is first released, and the reliability of its release channel is strong, it means that the current network feedback data may be the new content after the update and iteration of the local feedback data. At this time, the network feedback data is strengthened, that is, the current network feedback data is repeated again. When the network feedback data is strengthened, an enhanced label is added to the network feedback data. This can effectively avoid the situation where the network feedback data lags behind the local feedback data, causing a large error in the final result, and can also use more up-to-date network feedback data when the local feedback data lags.

[0068] In some implementations of this embodiment, the method of performing similarity check on local feedback data and network feedback data includes: before performing similarity check on local feedback data and network feedback data, first normalizing the data format of local feedback data and network feedback data to obtain a data form that can be directly used for comparison. Wherein, when the data form of the network feedback data is in the form of an image, it is necessary to first identify the image to obtain the text information contained in the image.

[0069] In some implementations of this embodiment, after obtaining the network feedback data, the method further includes: updating the local knowledge base according to the network feedback data.

[0070] Specifically, the method can enrich and update the local knowledge base according to the results of the network search, so as to avoid serious lag caused by the long-term non-update of the data of the local knowledge base.

[0071] The step of updating the local knowledge base includes: receiving an update command for the local knowledge base, and identifying whether the update command is the latest update command based on the latest update data of the local knowledge base; wherein the update command for the local knowledge base is generated based on network feedback data; if the update command is determined to be the latest update command, then based on the network feedback data, determining the data to be changed and metadata information related to the data to be changed; based on the metadata information related to the data to be changed, determining the corresponding change position in the database, and updating the local knowledge base based on the data to be changed. This can avoid repeated operations when the local knowledge base is updated and synchronized, thereby improving the data update synchronization speed of the local knowledge base.

[0072] In some implementations of this embodiment, the local knowledge base may also be updated manually.

[0073] In some implementations of this embodiment, the form of the above-mentioned question information includes text form and / or voice form; the analysis rules of the above-mentioned question information include: judging the form of the question information, identifying and converting the question information in voice form to obtain the corresponding question information in text form; performing text analysis on the question information in text form to obtain key data of the question.

[0074] Specifically, the question information sent by the user is obtained, and the question information is analyzed to obtain the key data of the question. Among them, the question information sent by the user can be in the form of text or voice. For the question information in text form, it can be directly analyzed. For the question information in voice form, it is necessary to first identify it and convert it into text form, and then analyze it. The specific analysis tools belong to the conventional technology in this field and will not be repeated here. The key data of the question refers to the core content in the question information that can reflect the user's true intention. For example, when the question information sent by the user is "What are the specific parameters of a certain device?", the corresponding key data of the question should be a certain device and specific parameters.

[0075] Figure 2 FIG. 1 is a block diagram of an intelligent dialogue system provided by an embodiment of the present invention. Figure 2 As shown, an embodiment of the present invention provides an intelligent dialogue system, including:

[0076] A local feedback data obtaining module is used to obtain local feedback data based on question key data and a local knowledge base; wherein the question key data is obtained by parsing the captured question information;

[0077] A network feedback data acquisition module is used to perform network search based on the key data of the question and obtain network feedback data;

[0078] A question answer data generating module, configured to generate question answer data based on the local feedback data and the network feedback data when it is determined that the similarity test result between the local feedback data and the network feedback data meets a preset condition;

[0079] The model analysis module is used to input the question answer data into the large language model for adjustment to obtain feedback information for answering the question information.

[0080] Specifically, the system obtains local feedback data based on the key data of the question and the local knowledge base, and performs network retrieval based on the key data of the question to obtain network feedback data. When it is determined that the similarity test results between the local feedback data and the network feedback data meet the preset conditions, the question answer data is generated based on the local feedback data and the network feedback data, and the question answer data is input into the large language model for adjustment to obtain feedback information for answering the question information. The system uses the local knowledge base generated based on the local knowledge file and the network search engine to generate feedback information corresponding to the question information. The data of the local knowledge base is only stored locally, which can fully guarantee the data security of the local knowledge base and avoid the leakage of confidential information. And the conversion of the question answer data to the feedback information is completed by the large language model, which can help users to obtain the desired results more easily and more easily. Therefore, when the large language model is applied to some special industry fields, the data security of the local knowledge base can be fully guaranteed to avoid the leakage of confidential information, thereby achieving the purpose of parsing and processing user questions in some special industry fields.

[0081] The embodiment of the present invention further provides a machine-readable storage medium, on which instructions are stored. When the instructions are executed by the processor 100, the processor 100 is configured to execute the above-mentioned intelligent dialogue method.

[0082] Machine-readable storage media include permanent and non-permanent, removable and non-removable media, and can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0083] An embodiment of the present invention further provides an electronic device 10, which includes a memory 101, a processor 100, and a computer program 102 stored in the memory 101 and executable on the processor 100. When the processor 100 executes the computer program 102, the above-mentioned intelligent dialogue method is implemented.

[0084] like Figure 3 FIG. 1 is a schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, the electronic device 10 of this embodiment includes: a processor 100, a memory 101, and a computer program 102 stored in the memory 101 and executable on the processor 100. When the processor 100 executes the computer program 102, the steps in the above method embodiment are implemented. Alternatively, when the processor 100 executes the computer program 102, the functions of each module / unit in the above device embodiment are implemented.

[0085] Exemplarily, the computer program 102 may be divided into one or more modules / units, one or more modules / units are stored in the memory 101, and are executed by the processor 100 to complete the present invention. One or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 102 in the electronic device 10. For example, the computer program 102 may be divided into a local feedback data acquisition module, a network feedback data acquisition module, a question answer data generation module, and a model analysis module.

[0086] The electronic device 10 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The electronic device 10 may include, but is not limited to, a processor 100 and a memory 101. Those skilled in the art will appreciate that Figure 3 It is only an example of the electronic device 10 and does not constitute a limitation of the electronic device 10. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0087] The processor 100 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0088] The memory 101 may be an internal storage unit of the electronic device 10, such as a hard disk or memory of the electronic device 10. The memory 101 may also be an external storage device of the electronic device 10, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 10. Further, the memory 101 may also include both an internal storage unit of the electronic device 10 and an external storage device. The memory 101 is used to store computer programs and other programs and data required by the electronic device 10. The memory 101 may also be used to temporarily store data that has been output or is to be output.

[0089] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0090] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program 102 products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program 102 product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0091] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program 102 products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by the computer program 102 instructions. These computer program 102 instructions can be provided to a processor 100 of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor 100 of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0092] These computer program 102 instructions may also be stored in a computer readable memory 101 that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory 101 produce an article of manufacture including an instruction device that implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0093] These computer program 102 instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0094] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0095] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. An intelligent dialogue method, characterized in that: include: Based on the question key data and the local knowledge base, local feedback data is obtained; wherein the question key data is obtained by parsing the captured question information; Conduct network search based on the key data of the question to obtain network feedback data; When it is determined that the similarity test result between the local feedback data and the network feedback data meets the preset conditions, generating question answer data based on the local feedback data and the network feedback data; The question answer data is input into a large language model for adjustment to obtain feedback information for answering the question information.

2. The intelligent dialogue method according to claim 1, characterized in that: The local knowledge base is constructed based on multiple local knowledge files; The local feedback data is obtained based on the question key data and the local knowledge base, including: Determine where the key data for the question appears in each local knowledge file; When the position where the question key data appears is in the title of the local knowledge file, assigning a first weight corresponding to the question information to the corresponding local knowledge file; When the location where the question key data appears is in the body of the local knowledge file, assigning a second weight corresponding to the question information to the corresponding local knowledge file; Determine the frequency of occurrence of the key data of the question in each local knowledge file, and assign a third weight corresponding to the question information to the corresponding local knowledge file based on the frequency of occurrence of the key data of the question in the local knowledge file; wherein the weight of the first weight is greater than the weight of the second weight, and the weight of the second weight is greater than the weight of the third weight; Based on the first weight, the second weight, and the third weight, each local knowledge file related to the key data of the question is sorted, and each local knowledge file related to the key data of the question is parsed to extract the content containing the key data of the question in each local knowledge file; Based on the content of each local knowledge file containing the key data of the question, local feedback data is generated.

3. The intelligent dialogue method according to claim 1, characterized in that: The network search is performed based on the key data of the question to obtain network feedback data, including: Based on the network search engine and the key data of the questions, the network search is carried out in the public database to obtain the network feedback data; The data format of the network feedback data is the same as the data format of the local feedback data, and the data format of the network feedback data includes a picture format and / or a text format.

4. The intelligent dialogue method according to claim 1, characterized in that: Before determining that the similarity test result between the local feedback data and the network feedback data meets a preset condition, the method further includes: A similarity check is performed on the local feedback data and the network feedback data. If the similarity check result does not meet the preset conditions, the network feedback data is updated and adjusted until the similarity check result meets the preset conditions.

5. The intelligent dialogue method according to claim 4, characterized in that: If the similarity test result does not meet the preset conditions, the network feedback data is updated and adjusted, including: If the similarity test result does not meet the preset conditions, the network feedback data will be screened for citations to trace the initial release time and initial release channel of the network feedback data; Conduct reliability analysis on the initial release channels of network feedback data to obtain channel reliability analysis data; If the initial release time of the network feedback data is earlier than the collection time of the local feedback data and / or the channel reliability analysis data is less than the preset reliability data threshold, the network search is performed again to generate new network feedback data; If the initial release time of the network feedback data is later than the collection time of the local feedback data and the channel reliability analysis data is not less than the preset reliability data threshold, data enhancement is performed on the network feedback data and an enhancement label is added to the network feedback data.

6. The intelligent dialogue method according to claim 1, characterized in that: After obtaining the network feedback data, the method further includes: Update the local knowledge base based on network feedback data.

7. The intelligent dialogue method according to claim 1, characterized in that: The question information may be in the form of text and / or voice; The analysis rules of the question information include: Determine the form of the question information, recognize and convert the question information in voice form to obtain the corresponding question information in text form; Perform text analysis on the question information in text form to obtain key data of the question.

8. An intelligent dialogue system, characterized in that: include: A local feedback data obtaining module is used to obtain local feedback data based on question key data and a local knowledge base; wherein the question key data is obtained by parsing the captured question information; A network feedback data acquisition module is used to perform network search based on the key data of the question and obtain network feedback data; A question answer data generating module, configured to generate question answer data based on the local feedback data and the network feedback data when it is determined that the similarity test result between the local feedback data and the network feedback data meets a preset condition; The model analysis module is used to input the question answer data into the large language model for adjustment to obtain feedback information for answering the question information.

9. A machine-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by a processor, the processor is configured to execute the intelligent dialogue method as described in any one of claims 1 to 7.

10. 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 intelligent dialogue method described in any one of claims 1 to 7 is implemented.

Images