Activity information retrieval method and device based on large model and medium

Abstract

The invention discloses an activity information retrieval method and device based on a large model and a medium, and the method comprises the following steps: obtaining enterprise business system data, and carrying out associated field integration processing on the enterprise business system data to obtain a business field wide table; converting preset activity audio data into activity text data, and storing the activity text data to an audio data storage position in the business field wide table to obtain a to-be-processed business field wide table; based on the to-be-processed business field wide table, a business field wide table for retrieval is obtained through mapping table description analysis of field description; obtaining audio resource data through supplementary retrieval processing according to the business field wide table for retrieval; and obtaining a user query text, and determining required information data through SQL conversion based on the user query text and the audio resource data. By means of the method, the technical problems that in the prior art, activity data retrieval is low in efficiency and inaccurate are solved.

Description

Technical Field

[0001] The present application relates to the technical field of large model data retrieval, and in particular to a large model-based activity information retrieval method, device and medium. Background Art

[0002] In existing data retrieval systems, end users often face the problem of being unable to accurately describe search terms, resulting in a large deviation between search results and user needs, making it difficult to achieve efficient use of data. Especially in intranet systems, there is a lack of effective technical means for the retrieval of audio resources, making it impossible to accurately identify and retrieve audio content, resulting in a large amount of valuable audio data being difficult to fully utilize.

[0003] Due to the wide variety of information records and the complexity of data types in corporate organization or party building activities, current technical solutions mainly rely on traditional keyword matching and structured queries, which are difficult to cope with the retrieval needs of unstructured data (such as audio) and cannot solve the problem of retrieval accuracy when the user description is vague. These problems lead to low efficiency in activity data retrieval, poor user experience, and inability to fully realize the value of data. Summary of the invention

[0004] The embodiments of the present application provide a method, device and medium for activity information retrieval based on a large model, which solves the technical problems of low efficiency and inaccurate retrieval of activity data in the prior art.

[0005] In a first aspect, an embodiment of the present application provides an activity information retrieval method based on a large model, characterized in that the method includes: acquiring enterprise business system data, and performing associated field integration processing on the enterprise business system data to obtain a business field wide table; converting preset activity audio data into activity text data, and storing the activity text data in the audio data storage location in the business field wide table to obtain a business field wide table to be processed; based on the business field wide table to be processed, obtaining a business field wide table for retrieval through a mapping table description analysis of the field description; obtaining audio resource data through supplementary retrieval processing based on the business field wide table for retrieval; acquiring user query text, and determining the required information data through SQL conversion based on the user query text and the audio resource data.

[0006] In one implementation of the present application, the enterprise business system data is integrated by associated fields to obtain a wide table of business fields, specifically including: performing system business field retrieval on the enterprise business system data to obtain business associated fields; determining searchable business associated fields through retrieval capability judgment based on the business associated fields; performing retrieval quality assessment on the searchable business associated fields to obtain retrieval key fields; and obtaining a wide table of business fields through field quality combing based on the retrieval key fields.

[0007] In one implementation of the present application, preset active audio data is converted into active text data, specifically including: based on the active audio data, determining an active audio conversion model through audio type conversion preference analysis; according to the active audio conversion model, performing audio-to-text conversion on the active audio data to obtain active text data corresponding to the active audio data.

[0008] In one implementation of the present application, based on the wide table of business fields to be processed, a wide table of business fields for retrieval is obtained through analysis of the mapping table description of the field description, specifically including: based on the wide table of business fields to be processed, determining the mapping table field type through field type distinction; configuring the mapping table field type with a mapping relationship description to obtain a description mapping table; wherein the content of the description mapping table is background information on the semantics corresponding to the field and the field mapping relationship; according to the description mapping table, obtaining the wide table of business fields for retrieval through synonym entry analysis.

[0009] In one implementation of the present application, audio resource data is obtained through supplementary search processing based on a wide table of business fields for retrieval, specifically including: determining supplementary retrieval requirements through ES word segmentation retrieval analysis based on a wide table of business fields for retrieval; performing conversion detail analysis on the supplementary retrieval requirements to obtain supplementary retrieval fields; wherein the conversion detail analysis includes: audio-to-text process records, new text screening, keyword recognition, and keyword confidence analysis; and obtaining audio resource data through key data integration based on the supplementary retrieval fields.

[0010] In one implementation of the present application, based on the wide table of business fields for retrieval, ES word segmentation retrieval analysis is used to determine the supplementary retrieval needs, specifically including: importing the wide table of business fields for retrieval into the ES system, and performing field preprocessing on the imported wide table of business fields for retrieval to obtain word segmentation data; wherein, field preprocessing includes: word segmentation processing, stop word removal, and keyword stem extraction; based on the word segmentation data, the supplementary retrieval needs are determined through word segmentation usage status analysis; wherein, the word segmentation usage status analysis includes: word segmentation usage frequency analysis, business logic analysis, and word segmentation weight analysis.

[0011] In one implementation of the present application, based on the user query text and audio resource data, the required information data is determined through SQL conversion, specifically including: performing semantic recognition on the user query text to obtain the user search instruction; converting the user search instruction into SQL, and based on the SQL, performing resource search on the audio resource data to obtain the required information data.

[0012] In one implementation of the present application, after determining the required information data through SQL conversion based on user query text and audio resource data, the method also includes: monitoring the usage process of the required information data to obtain key node data; wherein the key node data includes: keyword occurrence frequency, text usage nodes; based on the key node data, determining the field update requirements through text association analysis; performing question-and-answer mapping updates on the field update requirements to obtain iterative update retrieval parameters.

[0013] In a second aspect, an embodiment of the present application also provides an activity information retrieval device based on a large model, characterized in that the device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can: obtain enterprise business system data, and perform associated field integration processing on the enterprise business system data to obtain a business field wide table; convert preset activity audio data into activity text data, and store the activity text data in an audio data storage location in the business field wide table to obtain a business field wide table to be processed; based on the business field wide table to be processed, obtain a business field wide table for retrieval through a mapping table description analysis of the field description; obtain audio resource data through supplementary retrieval processing based on the business field wide table for retrieval; obtain user query text, and determine the required information data through SQL conversion based on the user query text and the audio resource data.

[0014] In a third aspect, an embodiment of the present application further provides a non-volatile computer storage medium for activity information retrieval based on a large model, storing computer executable instructions, characterized in that the computer executable instructions are configured to: obtain enterprise business system data, and perform associated field integration processing on the enterprise business system data to obtain a business field wide table; convert preset activity audio data into activity text data, and store the activity text data in an audio data storage location in the business field wide table to obtain a business field wide table to be processed; based on the business field wide table to be processed, obtain a business field wide table for retrieval through a mapping table description analysis of the field description; obtain audio resource data through supplementary retrieval processing based on the business field wide table for retrieval; obtain user query text, and determine the required information data through SQL conversion based on the user query text and the audio resource data.

[0015] The embodiments of the present application provide a method, device and medium for activity information retrieval based on a big model. By constructing a search and analysis platform based on a big model and intelligently analyzing audio data, the technical problems of low efficiency and inaccurate retrieval of activity data in the prior art are solved, data isolation of each subsidiary and business unit is avoided, and accurate retrieval of activity business data is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 A flow chart of an activity information retrieval method based on a large model provided in an embodiment of the present application;

[0018] Figure 2 A schematic diagram of the internal structure of an activity information retrieval device based on a large model provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0020] The embodiments of the present application provide a method, device and medium for activity information retrieval based on a big model. By constructing a search and analysis platform based on a big model and intelligently analyzing audio data, the technical problems of low efficiency and inaccurate retrieval of activity data in the prior art are solved, data isolation of each subsidiary and business unit is avoided, and accurate retrieval of activity business data is achieved.

[0021] The technical solution proposed in the embodiments of the present application is described in detail below with reference to the accompanying drawings.

[0022] Figure 1 A flow chart of a method for retrieving activity information based on a large model provided in an embodiment of the present application. Figure 1 As shown, an activity information retrieval method based on a large model provided in an embodiment of the present application specifically includes the following steps:

[0023] Step 101: Acquire enterprise business system data, and perform associated field integration processing on the enterprise business system data to obtain a business field wide table.

[0024] In existing data retrieval systems, end users often face the problem of being unable to accurately describe search terms, resulting in a large deviation between search results and user needs, making it difficult to achieve efficient use of data. Especially in intranet systems, there is a lack of effective technical means for the retrieval of audio resources, making it impossible to accurately identify and retrieve audio content, resulting in a large amount of valuable audio data being difficult to fully utilize.

[0025] For example, in the process of party building activities in various departments of the enterprise, relevant business documents are recorded in the form of text, charts, etc., and meeting and activity records are recorded through video or audio recording. The recorded party building activity data is stored in the enterprise business system. Since the activity information between various departments and related enterprises of the enterprise is not interoperable, when performing data retrieval for party building activities, it is necessary to integrate the relevant data and obtain the business field wide table of the corresponding fields to meet the needs of data interoperability and the organization of related data.

[0026] Specifically, the enterprise business system data is searched through system business fields to obtain business-related fields. Based on the business-related fields, the searchable business-related fields are determined through search capability judgment. The search quality of the searchable business-related fields is evaluated to obtain key search fields. Based on the key search fields, a wide table of business fields is obtained through field quality combing.

[0027] In one embodiment, the data processor combines the overall business system of the enterprise to organize the interrelated fields in each business system, and then continues to evaluate each field based on whether it has retrieval capabilities, giving priority to fields with a large number of retrievals. For example, for the integration of user information, users involve many fields, but user names and user home addresses are fields with high retrieval quality. For fields such as user email addresses, the retrieval quality is relatively low, that is, there is no need to include them in the searchable fields.

[0028] Data is cleaned, sorted, summarized and converted according to the subject, and unified into high-quality data for retrieval. According to the data subject, the data is stored in a unified database.

[0029] Step 102: Convert the preset active audio data into active text data, and store the active text data into the audio data storage location in the business field wide table to obtain the business field wide table to be processed.

[0030] Since the distribution of activity audio data in the enterprise business system is complicated, it is convenient to search for it after determining the search field to obtain all the activity audio data. For the obtained activity audio data (retrieved by the search key field), it is necessary to integrate it with the business field wide table to obtain an integrated wide table containing the activity audio data and other key field activity data.

[0031] Exemplarily, the audio data of the party building activities are collected and input into a preset audio conversion model group, and the audio data of the activities are converted into text data through data conversion. The converted text data is then stored in a wide table.

[0032] Specifically, based on the activity audio data, an activity audio conversion model is determined through audio type conversion preference analysis; according to the activity audio conversion model, the activity audio data is converted into audio text to obtain activity text data corresponding to the activity audio data.

[0033] In one embodiment, for active audio data, an audio conversion model is selected based on audio characteristics. This embodiment adopts a Paraformer model to convert active audio data into active text data, which is then stored in a large wide table and stored in the same data as the original audio resource.

[0034] Step 103: Based on the to-be-processed business field wide table, the mapping table description of the field description is analyzed to obtain the retrieval business field wide table.

[0035] After obtaining the wide table of business fields to be processed, the meaning of the field data recorded in the table needs to be determined so that the user can identify the input instructions during the search process. For the meaning of the field and its mapping relationship with the field itself, this application uses a mapping table to map the user's question and answer instructions to the wide table of business fields, thereby improving the efficiency and accuracy of user instruction recognition.

[0036] For example, according to the requirements of the data table and fields, the relevant information of the large wide table is edited, and possible synonyms are entered and saved. For different, similar, and consistent word meanings, one-to-many meaning mapping is performed on different fields, and the field meanings are stored as background information.

[0037] It should be noted that the relevant background information of the mapping table is not in the large wide table.

[0038] Specifically, based on the wide table of business fields to be processed, the mapping table field type is determined by distinguishing the field types; the mapping table field types are described and configured to obtain a description mapping table; wherein the content of the description mapping table is the background information of the semantics corresponding to the field and the field mapping relationship; according to the description mapping table, the business field wide table for retrieval is obtained through synonym entry analysis.

[0039] In one embodiment, the present application uses the SPACE-T model for parsing. The SPACE-T model is a multi-round table knowledge pre-training language model, which is specifically used to solve the downstream multi-round Text-to-SQL semantic parsing tasks. After analyzing the acquisition needs of the active business, the SPACE-T model determines the relevant data table and field type requirements, and improves the background information according to the needs of the data table and field, etc. After the user enters the search instruction, the user's search needs are determined through semantic recognition, and the circle of the question-and-answer process is called to query the mapping relationship. The mapping relationship is stored in the mapping table, and the circle queries the mapping table to obtain the field corresponding to the user's search needs. Therefore, in the process of perfecting and obtaining the wide table of business fields for retrieval, it is necessary to expand the fields and the background information corresponding to the fields. Similarly, the mapping table for the mapping relationship between the two also needs to perform corresponding data expansion, and then edit the relevant information of the large wide table, and enter and save any possible synonyms.

[0040] Step 104: Obtain audio resource data through supplementary search processing based on the service field wide table for search.

[0041] In the actual activity data retrieval process, since there are audio retrieval fields in the business data, it is necessary to update the retrieval fields of the audio files that are not subsequently retrieved or converted to achieve advanced retrieval of audio data and improve the retrieval efficiency of audio data in the activity business data.

[0042] For example, word segmentation search is implemented through ES. In addition to the original search fields of the audio, detailed data of the audio converted into text is added for supplementary search, thereby realizing advanced search of audio data.

[0043] Specifically, it includes: based on the wide table of business fields used for retrieval, through ES word segmentation retrieval analysis, determining the supplementary retrieval needs; conducting conversion detail analysis on the supplementary retrieval needs to obtain the supplementary retrieval fields; wherein, the conversion detail analysis includes: audio-to-text process records, new text screening, keyword recognition, keyword confidence analysis; according to the supplementary retrieval fields, audio resource data is obtained through key data integration.

[0044] In one embodiment, the input character stream is segmented into independent tokens by the tokenizer of ES (Elasticsearch). According to the characteristics of the text after audio-to-text conversion, a suitable tokenizer is selected for configuration and optimization, and the search results are sorted according to the matching degree between the document and the query token and the relevance algorithm to obtain supplementary search fields. Finally, the key data is integrated and updated into the wide table of business fields for retrieval to obtain audio resource data.

[0045] Based on the wide table of business fields for retrieval, the search supplement needs are determined through ES word segmentation retrieval analysis, specifically including: importing the wide table of business fields for retrieval into the ES system, and performing field preprocessing on the imported wide table of business fields for retrieval to obtain word segmentation data; wherein, field preprocessing includes: word segmentation processing, stop word removal, and keyword stem extraction; based on the word segmentation data, the search supplement needs are determined through word segmentation usage status analysis; wherein, word segmentation usage status analysis includes: word segmentation usage frequency analysis, business logic analysis, and word segmentation weight analysis.

[0046] Step 105: Obtain the user query text, and determine the required information data through SQL conversion based on the user query text and audio resource data.

[0047] Specifically, it includes: performing semantic recognition on user query text to obtain user search instructions; converting user search instructions into SQL, and performing resource search on audio resource data based on SQL to obtain required information data.

[0048] Exemplarily, by providing the SPACE-T model, the text input by the user is semantically recognized and converted into corresponding SQL, thereby implementing the query and ultimately returning the data required by the user.

[0049] After determining the required information data through SQL conversion based on user query text and audio resource data, the method also includes: monitoring the usage process of the required information data to obtain key node data; wherein the key node data includes: keyword occurrence frequency, text usage nodes; based on the key node data, determining the field update requirements through text association analysis; performing question-answer mapping update on the field update requirements to obtain iterative update retrieval parameters.

[0050] This application can help enterprises develop better in the following scenarios by building a search and analysis platform based on a big model: audio retrieval, breaking the data isolation of each subsidiary and business unit, providing audio data connection at the company level, and providing retrieval capabilities to achieve data sharing and traceability; key business scenarios, by connecting business lines, integrating relevant feature data of business lines together, and using the text2Sql processing of the big model to provide the ability to generate SQL from text, and achieve accurate retrieval of business data.

[0051] The above is an embodiment of the method proposed in this application. Based on the same inventive concept, the embodiment of this application also provides an activity information retrieval device based on a large model, whose structure is as follows: Figure 2 shown.

[0052] Figure 2 The internal structure diagram of an activity information retrieval device based on a large model provided in an embodiment of the present application. Figure 2 As shown, the device includes:

[0053] at least one processor 201;

[0054] and, a memory 202 communicatively connected to the at least one processor;

[0055] The memory 202 stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor 201 to enable at least one processor 201 to:

[0056] The enterprise business system data is obtained, and the associated fields of the enterprise business system data are integrated to obtain a business field wide table; the preset active audio data is converted into active text data, and the active text data is stored in the audio data storage location in the business field wide table to obtain a business field wide table to be processed; based on the business field wide table to be processed, a mapping table description analysis of the field description is performed to obtain a business field wide table for retrieval; based on the business field wide table for retrieval, audio resource data is obtained through supplementary retrieval processing; the user query text is obtained, and based on the user query text and the audio resource data, the required information data is determined through SQL conversion.

[0057] Some embodiments of the present application provide corresponding Figure 1 A non-volatile computer storage medium for large model-based active information retrieval, storing computer executable instructions, wherein the computer executable instructions are set as follows:

[0058] The enterprise business system data is obtained, and the associated fields of the enterprise business system data are integrated to obtain a business field wide table; the preset active audio data is converted into active text data, and the active text data is stored in the audio data storage location in the business field wide table to obtain a business field wide table to be processed; based on the business field wide table to be processed, a mapping table description analysis of the field description is performed to obtain a business field wide table for retrieval; based on the business field wide table for retrieval, audio resource data is obtained through supplementary retrieval processing; the user query text is obtained, and based on the user query text and the audio resource data, the required information data is determined through SQL conversion.

[0059] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the IoT device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0060] The system and medium provided in the embodiments of the present application correspond one-to-one to the method. Therefore, the system and medium also have similar beneficial technical effects to the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the system and medium will not be repeated here.

[0061] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0062] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program 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 computer program instructions. These computer program instructions can be provided to a processor 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 of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. 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.

[0063] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0064] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing 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.

[0065] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0066] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0067] Computer readable media include permanent and non-permanent, removable and non-removable media that 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 magnetic 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.

[0068] 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.

[0069] 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. A method for activity information retrieval based on a large model, characterized in that: The method comprises: Acquire enterprise business system data, and perform associated field integration processing on the enterprise business system data to obtain a business field wide table; Converting preset active audio data into active text data, and storing the active text data in the audio data storage location in the service field wide table to obtain a service field wide table to be processed; Based on the to-be-processed service field wide table, a retrieval service field wide table is obtained by analyzing the mapping table description of the field description; According to the service field wide table for retrieval, audio resource data is obtained through supplementary retrieval processing; The user query text is obtained, and based on the user query text and the audio resource data, the required information data is determined through SQL conversion.

2. The method for retrieving activity information based on a large model according to claim 1, characterized in that: The enterprise business system data is subjected to associated field integration processing to obtain a business field wide table, specifically including: Performing a system business field search on the enterprise business system data to obtain business-related fields; Based on the business-related fields, determining the searchable business-related fields through search capability judgment; Performing a search quality assessment on the searchable business-related fields to obtain search key fields; According to the search key fields, the business field wide table is obtained by combing the field quality.

3. The activity information retrieval method based on a large model according to claim 1 is characterized in that: Converting preset active audio data into active text data, specifically including: Based on the activity audio data, determining an activity audio conversion model through audio type conversion preference analysis; According to the activity audio conversion model, the activity audio data is converted into audio text to obtain activity text data corresponding to the activity audio data.

4. The method for retrieving activity information based on a large model according to claim 1, characterized in that: Based on the to-be-processed service field wide table, a retrieval service field wide table is obtained through analysis of the mapping table description of the field description, specifically including: Based on the wide table of business fields to be processed, determining the field type of the mapping table by distinguishing the field types; Performing a mapping relationship description configuration on the mapping table field type to obtain a description mapping table; wherein the content of the description mapping table is the background information of the semantics corresponding to the field and the mapping relationship of the field; According to the description mapping table, the business field wide table for retrieval is obtained through synonym entry analysis.

5. The method for activity information retrieval based on a large model according to claim 1, characterized in that: According to the service field wide table for retrieval, audio resource data is obtained through supplementary retrieval processing, specifically including: Based on the wide table of business fields for retrieval, the search supplementary requirements are determined through ES word segmentation search analysis; Performing conversion details analysis on the supplementary search requirements to obtain supplementary search fields; wherein the conversion details analysis includes: audio-to-text process records, newly added text screening, keyword recognition, and keyword confidence analysis; The audio resource data is obtained by integrating key data according to the supplementary search field.

6. The method for activity information retrieval based on a large model according to claim 5, characterized in that: Based on the wide table of business fields for retrieval, the ES word segmentation retrieval analysis is performed to determine the retrieval supplementary requirements, including: Importing the retrieval business field wide table into the ES system, and performing field preprocessing on the imported retrieval business field wide table to obtain word segmentation data; wherein the field preprocessing includes: word segmentation processing, stop word removal, and keyword stem extraction; Based on the word segmentation data, the search supplement demand is determined through word segmentation usage status analysis; wherein the word segmentation usage status analysis includes: word segmentation usage frequency analysis, business logic analysis, and word segmentation weight analysis.

7. The method for activity information retrieval based on a large model according to claim 1, characterized in that: Based on the user query text and the audio resource data, the required information data is determined through SQL conversion, specifically including: Performing semantic recognition on the user query text to obtain a user search instruction; The user search instruction is converted into SQL, and based on the SQL, the audio resource data is searched for resources to obtain the required information data.

8. The method for activity information retrieval based on a large model according to claim 1, characterized in that: After determining the required information data through SQL conversion based on the user query text and the audio resource data, the method further includes: Monitor the usage process of the required information data to obtain key node data; wherein the key node data includes: keyword occurrence frequency and text usage nodes; Based on the key node data, determine the field update requirements through text association analysis; The question-answer mapping is updated for the field update requirement to obtain iterative update retrieval parameters.

9. An activity information retrieval device based on a large model, characterized in that: The device comprises: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to: Acquire enterprise business system data, and perform associated field integration processing on the enterprise business system data to obtain a business field wide table; Converting preset active audio data into active text data, and storing the active text data in the audio data storage location in the service field wide table to obtain a service field wide table to be processed; Based on the to-be-processed service field wide table, a retrieval service field wide table is obtained by analyzing the mapping table description of the field description; According to the service field wide table for retrieval, audio resource data is obtained through supplementary retrieval processing; The user query text is obtained, and based on the user query text and the audio resource data, the required information data is determined through SQL conversion.

10. A non-volatile computer storage medium for large model-based activity information retrieval, storing computer executable instructions, characterized in that: The computer executable instructions are configured to: Acquire enterprise business system data, and perform associated field integration processing on the enterprise business system data to obtain a business field wide table; Converting preset active audio data into active text data, and storing the active text data in the audio data storage location in the service field wide table to obtain a service field wide table to be processed; Based on the to-be-processed service field wide table, a retrieval service field wide table is obtained by analyzing the mapping table description of the field description; According to the service field wide table for retrieval, audio resource data is obtained through supplementary retrieval processing; The user query text is obtained, and based on the user query text and the audio resource data, the required information data is determined through SQL conversion.

Images