Multi-level derivation engine analysis method and system for AI work card audio text

Abstract

The invention provides a multi-level derivation engine analysis method for an AI work card audio text. The method comprises the following steps: configuring analysis models of different service types; acquiring an audio text acquired and transferred by the AI work card; according to the business type, selecting a corresponding analysis model to analyze the audio text; an analysis task is established and executed, and different working threads are responsible for task scheduling, task timeout management and task state derivation in parallel; and obtaining an analysis result of the audio text. According to the method, the multi-stage derivation state is adopted in the analysis process, the coupling is reduced, and the state consistency is enhanced; the extraction efficiency of customer information is effectively improved, and manual intervention is reduced; in practical application, the accuracy of information extraction is improved by at least 30%, and the customer insight ability of the consultant is enhanced.

Description

Technical Field

[0001] The present invention relates to the field of audio text analysis, and in particular to a multi-level derivation engine analysis method and system for AI work badge audio text. Background Art

[0002] In the real estate sales process, after the salesperson converts the recording into text through the AI ​​work badge, how to effectively analyze and extract the customer's effective information so that the consultant can quickly grasp the customer's needs and increase the transaction rate has become an urgent problem to be solved. The current challenges include: 1. The text information is complex and it is difficult to quickly extract effective information. 2. The lack of a flexible model selection mechanism makes it impossible to perform efficient analysis on different types of information. 3. Relying only on regular expressions or a single model can easily cause inaccurate information extraction and affect decision-making. 4. The total task relies on hard-coded and long transactions to update the task status, making it difficult to ensure the final consistency of the task status. Summary of the invention

[0003] In response to the problems existing in the prior art, a multi-level deduction engine analysis method and system for AI work badge audio text is provided. By combining the rule engine and the big model, multi-level speech analysis is realized to extract the hit information from the audio text.

[0004] The first aspect of the present invention proposes a multi-level derivation engine analysis method for an AI work badge audio text, comprising:

[0005] Configure analysis models for different business types;

[0006] Get the audio text collected and transcribed by AI ID card;

[0007] Select the corresponding analysis model according to the business type to analyze the audio text;

[0008] Establish and execute analysis tasks, including task scheduling, task timeout management, and task status derivation through different worker threads in parallel;

[0009] Get the analysis results of the audio text.

[0010] As a preferred solution, the business types of the analysis model include sales talk, demand generation, risk control and labeling; the analysis model includes speech points under the corresponding business type, and each speech point is configured with a corresponding rule engine algorithm or large model algorithm according to needs.

[0011] As a preferred solution, the selecting a corresponding analysis model according to the business type to analyze the audio text includes: selecting one or multiple analysis models at the same time to analyze the audio text in parallel.

[0012] As a preferred solution, the process of establishing an analysis task includes:

[0013] Establishing a main task table; the main task table stores all main tasks and corresponding task states;

[0014] Create a corresponding number of main tasks according to the number of audio texts to be analyzed;

[0015] Create a corresponding number of subtasks under the corresponding main task according to the number of selected analysis models;

[0016] Initialize the main task and subtask status to be analyzed.

[0017] As a preferred solution, when executing an analysis task, each subtask uses a single large model or rule engine algorithm under the analysis model to perform NLP text analysis.

[0018] As a preferred solution, the task scheduling process includes:

[0019] Scan the main tasks to be analyzed in the main task table;

[0020] For each main task to be analyzed, if the audio-text transcription is completed, NLP analysis will be initiated to the manufacturer according to the subtask. If the initiation is successful, the status of the subtask and the main task will be updated to analyzing. If the initiation fails, the status of the subtask and the main task will be updated to analysis failure. If the audio-text transcription fails, the status of the subtask and the main task will be updated to analysis failure.

[0021] As a preferred solution, the task timeout management includes:

[0022] Scan the main tasks to be analyzed in the main task table;

[0023] Query the execution time of the main task to be analyzed. If the execution time is greater than the threshold, update the status of the subtask and main task to analysis failure.

[0024] As a preferred solution, the task state derivation includes:

[0025] Scan the main tasks to be analyzed in the main task table;

[0026] Query the status of all subtasks under the main task, and update the status of the main task according to the status of the subtasks.

[0027] As a preferred solution, the updating of the status of the main task according to the status of the subtask specifically includes:

[0028] If there is a subtask under the main task with the status of analysis failure, the status of the main task will be updated to analysis failure;

[0029] If the status of all subtasks under the main task is analysis success, the status of the main task is updated to analysis success;

[0030] If there is a subtask in the main task with the status of "Analysis in Progress" and the status of other subtasks is not "Analysis Failed", the status of the main task remains "Analysis in Progress".

[0031] The second aspect of the present invention proposes a multi-level deduction engine analysis system for AI work badge audio text, including a memory and a processor, wherein the memory stores a computer program corresponding to the multi-level deduction engine analysis method for AI work badge audio text as described in the first aspect and can be loaded and executed by the processor.

[0032] Compared with the prior art, the beneficial effects of adopting the above technical solution are:

[0033] (1) The analysis process derives the state at multiple levels, reducing coupling and enhancing state consistency

[0034] (2) Effectively improve the efficiency of extracting customer information and reduce manual intervention.

[0035] (3) In actual applications, the accuracy of information extraction is improved by at least 30%, enhancing consultants’ customer insight capabilities.

[0036] (4) Form an extensible rule engine framework to facilitate the introduction and upgrading of subsequent models. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a flow chart of the multi-level derivation engine analysis method for the AI ​​work badge audio text proposed by the present invention.

[0038] Figure 2 This is a task scheduling flowchart proposed in an embodiment of the present invention.

[0039] Figure 3 This is a flowchart of task timeout management proposed in an embodiment of the present invention.

[0040] Figure 4 This is a flowchart of task state derivation proposed in an embodiment of the present invention.

[0041] Figure 5 A system schematic diagram of an embodiment of the present application.

[0042] Figure 6 A schematic diagram of the structure of a computer system suitable for implementing the embodiments of the present application. DETAILED DESCRIPTION

[0043] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0044] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0045] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0046] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0047] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In the current real estate sales process, property consultants will record the conversation content through AI work badges, and then transcribe the recording into text, and then analyze and refine the text so that the property consultant can focus on the key information of the customer. However, the current analysis method has problems such as difficulty in quickly extracting effective information due to the complexity of actual audio text information; lack of flexible model selection mechanism, resulting in the inability to perform efficient analysis on different types of information. Based on this, the embodiment of the present invention proposes a multi-level derivation engine analysis method for AI work badge audio text, please refer to Figure 1 , the specific plan is as follows:

[0049] Step 1: Configure analysis models for different business types.

[0050] Before the actual analysis, this embodiment pre-creates analysis models for different business types and configures dimensions and speech points. When performing analysis, select the corresponding analysis model to execute the analysis process. Specifically, in order to better classify and count the results of text analysis, this embodiment configures different dimensions under each analysis model, and also configures different speech points under each dimension. For example: under 1 model, there are 6 dimensions, 1 dimension and 2 speech points, a total of 12 speech points. In the subsequent analysis process, the hit situation is given as follows:

[0051] Dimension 1: (speech point 1×, speech point 2√) hit 1 speech,

[0052] Dimension 2: (Speech Point 1√, Speech Point 2√) hit 2 speeches,

[0053] Dimension 3: (Speech Point 1√, Speech Point 2√) hit 2 speeches,

[0054] Dimension 4: (1× speech point, 2× speech point) hit 0 speech points,

[0055] Subsequent business can count the results of sales speech in the model. In one embodiment, the analysis model includes four business types: sales talk, demand mining, risk control and labeling. Different types of models implement different business services, specifically:

[0056] (1) Sales presentation

[0057] The dimensions and speech points configured by the operation personnel under this type of model should follow the speech rules of the "sales talk" type or the large model prompt words. The goal is to count the amount of information conveyed by the sales staff when explaining to the customer. In subsequent business, this type of model will appear in the display of dimensions and speech execution rate.

[0058] (2) Digging for demand

[0059] The dimensions and speech points configured by the operation personnel under this type of model should follow the "demand mining" type speech rules or large model prompts. The goal is to count the number of needs and intentions that sales personnel ask customers about when explaining to them.

[0060] (3) Risk Control

[0061] The dimensions and points of speech configured by the operation personnel under this type of model should follow the "risk control" type speech rules or large model prompts. The goal is to count whether the sales staff touches on some sensitive information of the project configuration when explaining to customers.

[0062] (4) Tags

[0063] The dimensions and speech points configured by operators under this type of model should follow the "label" type speech regularization or large model prompt words, with the goal of extracting customer portraits.

[0064] The speech points under each analysis model can be configured with rule engine algorithms or large model algorithms according to needs to facilitate speech matching of audio texts. Among them, the rule engine algorithm can include regular expression and keyword rule engines. On the basis of retaining the rule engine, introducing a large model to analyze and refine audio texts can ensure cost-effectiveness and system scalability.

[0065] In practical applications, the analysis model type can be expanded as needed to meet different analysis requirements.

[0066] Step 2: Obtain the audio text collected and transcribed by the AI ​​work badge.

[0067] As the source of analysis data for this embodiment, the AI ​​work badge can realize functions such as collecting recording data and transcribing recordings into text in actual scenarios. The AI ​​work badge can be implemented using existing products, which will not be described in detail here. In other embodiments, audio text data can also be obtained by other means. The focus of the present invention is on the analysis of audio text, and there is no restriction on the means of obtaining audio text.

[0068] Step 3: Select the corresponding analysis model to analyze the audio text according to the business type.

[0069] After obtaining the audio text, the analysis model of the corresponding business type can be selected for analysis according to actual needs. In practical applications, different analysis models can be used to analyze the same audio text, that is, one or more analysis models can be selected to perform the analysis task.

[0070] Step 4: Establish and execute the analysis task, in which different working threads are responsible for task scheduling, task timeout management, and task status derivation in parallel.

[0071] When analyzing an audio text, a main task is created for each audio text. Take an audio text as an example to illustrate. Based on the selected analysis model, an analysis task is created to analyze the audio text. When creating an analysis task, different business types of analysis models will create different subtasks, that is, each time an analysis model is selected, a corresponding analysis subtask needs to be created. The specific creation process is as follows:

[0072] (1) Establish a main task table. The main task table stores all main tasks and corresponding task status.

[0073] (2) Create a corresponding number of main tasks according to the number of audio texts to be analyzed;

[0074] (3) Create a corresponding number of subtasks under the corresponding main task according to the selected analysis model;

[0075] For example, for audio text 1, the selected model is: [sales talk, demand generation, risk control]. At this time, a main task is established: the main task of text 1, under which: subtask 1: sales talk subtask; subtask 2: demand generation subtask; subtask 3: risk control subtask.

[0076] After the task is created, the main task and subtask status must be initialized to pending analysis. It should be noted that in this embodiment, the task status is divided into four states: pending analysis, analyzing, analysis completed, and analysis failed. The recording status is divided into two states: text transcription completed and text transcription failed.

[0077] After creating the analysis task, each subtask uses the single large model or rule engine algorithm configured under the model to perform NLP text analysis, and updates the status of the subtask according to the analysis results. The status of the main task is derived from the status of the subtask. By establishing parallel tasks under the multi-level deduction engine, an extensible task deduction framework is formed, which facilitates the introduction and upgrade of subsequent models.

[0078] In traditional analysis solutions, after each subtask is completed, it is necessary to check whether other subtasks are completed in the same transaction, which increases the complexity of the system and increases the probability of inconsistent states between subtasks and main tasks. Based on this, the embodiment of the present application uses different working threads to implement the management and control of the task state life cycle during the analysis process. The threads do not interfere with each other, mainly including the management and control of task scheduling, task timeout management, and task state derivation, as follows:

[0079] (I) For task scheduling:

[0080] Please refer to Figure 2 ,Task scheduling includes the implementation of the initiation scheduling of analysis tasks. First, scan the main task table to determine the main task whose status is to be analyzed. Then judge the audio text status. If the transcription of the recording is completed, the subtask under the main task will initiate NLP analysis to the corresponding NLP manufacturer according to the configured analysis algorithm. After the initiation is successful, the status of the subtask and the main task will be updated to analyzing respectively; if the initiation fails, the status of the subtask will be updated to analysis failure; if the transcription of the recording fails, the status of the subtask and the main task will be updated to analysis failure respectively.

[0081] (II) Task overtime management:

[0082] Please refer to Figure 3,Task timeout management is mainly used to force the timed analysis tasks to be closed to avoid continuous resource occupation. First, scan the main task table to confirm the status of the main task to be analyzed. Then, query the execution time of the main task. If the execution time is greater than the threshold (for example, 20 minutes), update the subtask and main task status to analysis failure, and broadcast the analysis failure message to other business parties, and the analysis task process is completed.

[0083] (III) For task status derivation:

[0084] Please refer to Figure 4 In this embodiment, the main task status is derived from the subtask status during the analysis process to ensure the probability that the subtask and main task status are consistent. First, the main task table is scanned to confirm that the status is the main task under analysis. Then the status of all subtasks under the main task is queried, and the status of the main task is updated according to the status of the subtask. In the main task status derivation process, the following three states are mainly included:

[0085] (1) Analysis failure: As long as there is a subtask that fails the analysis, the status of the main task is updated to analysis failure regardless of the status of other subtasks.

[0086] (2) Analysis success: When all subtasks are analyzed successfully, the main task is updated to analysis success.

[0087] (3) Analyzing: As long as there is a subtask in the analyzing state and the other subtasks are not in the analysis failed state, the main task status will not be updated and will remain in the analyzing state.

[0088] Step 5: Get the analysis result of the audio text.

[0089] According to the control of the task status life cycle during the analysis process, when the analysis is successful, the corresponding business model hit status can be obtained. Take 1 model as an example. Under this model, there are 6 dimensions, 1 dimension and 2 speech points, a total of 12 speech points. In the subsequent analysis process, the hit status is given as follows:

[0090] Dimension 1: (speech point 1×, speech point 2√) hit 1 speech,

[0091] Dimension 2: (Speech Point 1√, Speech Point 2√) hit 2 speeches,

[0092] Dimension 3: (Speech Point 1√, Speech Point 2√) hit 2 speeches,

[0093] Dimension 4: (1× speech point, 2× speech point) hit 0 speech points,

[0094] After the analysis is successful, the hit status of the corresponding model's speech points and the text snippet content of the specific hit speech points will be returned. For example: hit result: dimension hit rate: 3 / 4 = 0.75; speech hit rate: 5 / 12 = 0.42.

[0095] As for the case of using regular expressions to configure speech points, take the speech points configured with regular expressions as an example:

[0096] "(?: we|our company|our company|Longfor Company|Longfor Group|Longfor)\s*(?: already|have|have|experience|have experienced)\s*(?:[\done two three four five six seven eight nine ten hundred thousand trillion]+years(?: of)?|several years|many years|decades|decades|several years)\s*(?: history|development process|development history|development|growth process|growth history)(?: already)?"

[0097] When a property consultant introduces a similar phrase such as "Longfor Group has many years of history", the phrase is hit. The corresponding regular analysis service will return the model ID information, the phrase configuration information, and the hit result.

[0098] The multi-level derivation engine analysis method of the AI ​​work card audio text proposed in the present invention has the following advantages:

[0099] (1) Through the multi-level derivation analysis process, the state is deduced at multiple levels, the coupling is reduced and the state consistency is enhanced.

[0100] (2) Effectively improve the CPU utilization efficiency of the task.

[0101] (3) Single task status log can be queried

[0102] Please refer to Figure 5 The embodiment of the present invention also proposes a multi-level deduction engine analysis system for AI work badge audio text, including a memory 101 and a processor 102, wherein the memory 101 stores a computer program corresponding to the multi-level deduction engine analysis method for AI work badge audio text as described above, which can be loaded and executed by the processor 102.

[0103] Figure 6 A schematic diagram of the structure of a computer system suitable for implementing a multi-level deduction engine analysis system of an embodiment of the present application is shown.

[0104] It should be noted that Figure 6 The computer system 200 of the multi-level deduction engine analysis system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0105] like Figure 6As shown, the computer system 200 includes a central processing unit (CPU) 201, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 202 or the program loaded from the storage part 208 to the random access memory (RAM) 203, such as executing the method described in the above embodiment. In the RAM 203, various programs and data required for system operation are also stored. The CPU 201, ROM 202 and RAM 203 are connected to each other through a bus 204. An input / output (I / O) interface 205 is also connected to the bus 204.

[0106] The following components are connected to the I / O interface 205: an input section 206 including a keyboard, a mouse, etc.; an output section 207 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 208 including a hard disk, etc.; and a communication section 209 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 209 performs communication processing via a network such as the Internet. A drive 210 is also connected to the I / O interface 205 as needed. A removable medium 211, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 210 as needed so that a computer program read therefrom is installed into the storage section 208 as needed.

[0107] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 209, and / or installed from a removable medium 211. When the computer program is executed by a central processing unit (CPU) 201, various functions defined in the system of the present application are executed.

[0108] It should be noted that the computer-readable medium shown in the embodiment of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, - but not limited to - an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by an instruction execution system, device or device or used in combination with it. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0109] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0110] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.

[0111] As another aspect, the present application also provides a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of a computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the communication-based method described in the above embodiment.

[0112] As another aspect, the present application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the communication method described in the above embodiment.

[0113] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0114] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the implementation methods of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the implementation methods of the present application.

[0115] For those skilled in the art, the specific meanings of the above terms in the present invention can be understood in specific situations; the drawings in the embodiments are used to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0116] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A multi-level derivation engine analysis method for AI work card audio text, characterized in that: include: Configure analysis models for different business types; Get the audio text collected and transcribed by AI ID card; Select the corresponding analysis model according to the business type to analyze the audio text; Establish and execute analysis tasks, including task scheduling, task timeout management, and task status derivation through different worker threads in parallel; Get the analysis results of the audio text.

2. The multi-level derivation engine analysis method of AI work card audio text according to claim 1 is characterized in that: The business types of the analysis model include sales talk, demand generation, risk control and labeling; the analysis model includes speech points under the corresponding business type, and each speech point is configured with a corresponding rule engine algorithm or a large model algorithm according to the needs.

3. The multi-level derivation engine analysis method of the AI ​​work card audio text according to claim 1 or 2 is characterized in that: The selecting a corresponding analysis model according to the business type to analyze the audio text includes: selecting one or simultaneously selecting multiple analysis models to analyze the audio text in parallel.

4. The multi-level derivation engine analysis method of AI work card audio text according to claim 1 is characterized in that: The process of establishing an analysis task includes: Establishing a main task table; the main task table stores all main tasks and corresponding task states; Create a corresponding number of main tasks according to the number of audio texts to be analyzed; Create a corresponding number of subtasks under the corresponding main task according to the number of selected analysis models; Initialize the main task and subtask status to be analyzed.

5. The multi-level derivation engine analysis method of AI work card audio text according to claim 1 is characterized in that: When performing analysis tasks, each subtask uses a single large model or rule engine algorithm under the analysis model to perform NLP text analysis.

6. The multi-level derivation engine analysis method of AI work card audio text according to claim 1 is characterized in that: The task scheduling process includes: Scan the main tasks to be analyzed in the main task table; For each main task to be analyzed, if the audio-text transcription is completed, NLP analysis will be initiated to the manufacturer according to the subtask. If the initiation is successful, the status of the subtask and the main task will be updated to analyzing. If the initiation fails, the status of the subtask and the main task will be updated to analysis failure. If the audio-text transcription fails, the status of the subtask and the main task will be updated to analysis failure.

7. The multi-level derivation engine analysis method of AI work card audio text according to claim 1 is characterized in that: The task timeout management includes: Scan the main tasks to be analyzed in the main task table; Query the execution time of the main task to be analyzed. If the execution time is greater than the threshold, update the status of the subtask and main task to analysis failure.

8. The multi-level derivation engine analysis method of AI work card audio text according to claim 1 is characterized in that: The task state derivation includes: Scan the main tasks to be analyzed in the main task table; Query the status of all subtasks under the main task, and update the status of the main task according to the status of the subtasks.

9. The multi-level derivation engine analysis method of AI work card audio text according to claim 8 is characterized in that: The updating of the status of the main task according to the status of the subtask specifically includes: If there is a subtask under the main task with the status of analysis failure, the status of the main task will be updated to analysis failure; If the status of all subtasks under the main task is analysis success, the status of the main task is updated to analysis success; If there is a subtask in the main task with the status of "Analysis in Progress" and the status of other subtasks is not "Analysis Failed", the status of the main task remains "Analysis in Progress".

10. A multi-level derivation engine analysis system for AI work badge audio text, characterized in that: It includes a memory and a processor, wherein the memory stores a computer program corresponding to the multi-level deduction engine analysis method for the AI ​​work badge audio text as described in any one of claims 1 to 9, which can be loaded by the processor and executed.

Images