Auxiliary reading method, system and equipment based on RAG large model and storage medium

Abstract

The invention provides an auxiliary reading method, system and device based on an RAG large model, and a storage medium. The method comprises the following steps: S1, receiving original data in multiple data forms input by a user; s2, the original data are preprocessed, preprocessed data are obtained, and preprocessing comprises noise removal, word segmentation, part-of-speech tagging, cleaning and standardization processing; s3, performing data analysis and key information extraction on the preprocessed data by adopting an RAG large model to obtain structured information; s4, verifying the structured information by using a data verification algorithm, and outputting a verification result; s5, feedback information of the user on the verification result is obtained, and parameters and strategies of the RAG large model are adjusted according to the feedback information; and S6, presenting the verification result obtained in the step S4 and the optimized result fed back by the user in the step S5 in various output forms. The information acquisition efficiency can be improved, the understanding difficulty is reduced, and the personalized reading requirement of the user is met.

Description

Technical Field

[0001] The present application relates to the field of natural language processing technology, and in particular to an auxiliary reading method, system, device and storage medium based on a RAG large model. Background Art

[0002] In the era of accelerating globalization and information explosion, people need to process a large amount of multimodal data, including documents, pictures, tables, voice and video, etc. Traditional document reading methods and manual translation face many challenges in processing such complex data.

[0003] First, the traditional way of reading documents is inefficient and it is difficult to quickly and accurately obtain the required information from a large number of documents. Especially when dealing with multilingual documents, manual translation is not only time-consuming and labor-intensive, but also difficult to ensure efficiency and accuracy. In addition, directly reading the translated documents may increase the difficulty of understanding due to language differences and cultural background differences.

[0004] Secondly, multimodal data processing is also a difficult problem. Traditional OCR technology may have low accuracy and slow processing speed when processing image text. At the same time, the accuracy of audio and video automatic recognition technology needs to be improved. These problems limit the effective use of multimodal data.

[0005] In order to solve the above problems, people began to explore the use of artificial intelligence technology to assist in reading and understanding documents. With the development of artificial intelligence technology, especially the emergence of large language models, new solutions have been provided for intelligent assisted reading. However, the existing intelligent assisted reading systems still have some shortcomings. For example, some systems can only process data in a single modality and cannot process multiple data such as documents, pictures, tables, voice and video at the same time. In addition, some systems still have comprehension bias when processing specific fields or complex documents, and lack intelligent assistance and personalized experience.

[0006] In view of this, the present application proposes an auxiliary reading method, system, device and storage medium based on the RAG large model, which can improve the efficiency of information acquisition and reduce the difficulty of understanding. Summary of the invention

[0007] In order to solve the problems of low data processing efficiency, lack of intelligent assistance and personalized experience in traditional document reading, the present application provides an assisted reading method, system, device and storage medium based on the RAG large model to solve the above-mentioned technical defects.

[0008] According to the first aspect of the present application, an intelligent assisted reading method based on the RAG large model is proposed, and the method comprises the following steps:

[0009] S1. Receive original data in various data forms input by the user;

[0010] S2. Preprocessing the original data to obtain preprocessed data, wherein the preprocessing includes: noise removal, word segmentation, part-of-speech tagging, cleaning and normalization processing;

[0011] S3, using the RAG large model to perform data analysis and key information extraction on the preprocessed data to obtain structured information;

[0012] S4. Use a data verification algorithm to verify the structured information and output the verification result;

[0013] S5. Obtain user feedback on the verification results, and adjust the parameters and strategies of the RAG large model according to the feedback information;

[0014] S6. Present the verification results obtained in step S4 and the results optimized according to user feedback in step S5 in a variety of output forms.

[0015] Preferably, the auxiliary reading method based on the RAG large model provided by the present application further includes:

[0016] In response to determining that the original data is an image, a deep learning model is used to perform preliminary feature extraction and classification on the image to identify major objects and scenes in the image;

[0017] and using optical character recognition technology to extract text information from images, including titles, tags, and comments;

[0018] Use object detection algorithms to further identify key elements in the image, including people, objects, and logos, and annotate the locations and attributes of key elements;

[0019] Combine text information and key elements to generate a description of the image using natural language processing techniques, and use the RAG large model to retrieve information related to the image content from external knowledge bases or related documents;

[0020] The generated image description and the retrieved information related to the image content are integrated to obtain structured information.

[0021] Preferably, the auxiliary reading method based on the RAG large model provided by the present application further includes:

[0022] In response to determining that the original data is a table, extracting row, column and cell data from the table using a table parsing technique;

[0023] Clean and normalize row, column, and cell data to obtain preprocessed data;

[0024] Apply data mining techniques to analyze preprocessed data and use natural language processing techniques to generate table descriptions;

[0025] The RAG large model is used to retrieve information related to the table content from external knowledge bases or related documents;

[0026] The generated table description and the retrieved information related to the table content are integrated to obtain structured information.

[0027] Preferably, the auxiliary reading method based on the RAG large model provided by the present application further includes:

[0028] In response to determining that the original data is a large amount of document data, using indexing technology and query optimization strategy to retrieve information related to user needs from the large amount of document data;

[0029] and using the RAG big model to retrieve information related to user needs from external knowledge bases or related documents;

[0030] Conduct content analysis and understanding of the retrieved relevant information, extract key information and themes,

[0031] and using natural language processing techniques to perform semantic and sentiment analysis on the retrieved information;

[0032] Based on user needs and data characteristics, the GPT model is combined to generate various forms of output content, including summaries, question-answer pairs, and graph recognition.

[0033] Calibrate and adjust the generated output content using RAG technology;

[0034] Integrate the calibrated and adjusted output content with the retrieved related information to obtain structured information.

[0035] Preferably, in step S1, the original data includes documents, pictures, tables, voice and video data.

[0036] Preferably, in step S2, it further includes: in response to determining that the original data is voice or video data, using voice recognition technology to recognize the original data and convert it into text content.

[0037] Preferably, in step S1, raw data in various data formats are acquired from at least one data source, and the data source includes a database, a sensor network, a user input interface or a network server.

[0038] In a second aspect, the present application provides an intelligent auxiliary reading system based on the RAG large model, the system comprising:

[0039] A data input module, configured to receive raw data in various data forms input by a user;

[0040] A preprocessing module is configured to preprocess the raw data to obtain preprocessed data, wherein the preprocessing includes: noise removal, word segmentation, part-of-speech tagging, cleaning and normalization processing;

[0041] RAG large model module, configured to use RAG large model to perform data analysis and key information extraction on pre-processed data to obtain structured information;

[0042] A verification module is configured to verify the structured information using a data verification algorithm and output a verification result;

[0043] A feedback module is configured to obtain user feedback information on the verification results and adjust the parameters and strategies of the RAG large model according to the feedback information;

[0044] The result output module is configured to present the verification results obtained by the verification module and the results optimized according to user feedback in the feedback module in multiple output forms.

[0045] In a third aspect, the present application provides a terminal device, including a processor, a memory, and a computer program stored in the memory, wherein the computer program is executed by the processor to implement any of the above-mentioned intelligent assisted reading methods based on the RAG large model.

[0046] In a fourth aspect, the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, any of the above-mentioned intelligent assisted reading methods based on the RAG large model is implemented.

[0047] Compared with the prior art, the beneficial results of the present invention are:

[0048] (1) The present invention directly uses the RAG large model to process the images and tables in the document, thereby improving the processing efficiency and accuracy.

[0049] (2) Combined with ASR and other technologies, it can realize content recognition and extraction of various data forms, helping users to quickly and accurately understand document and data content.

[0050] (3) Introducing interactive machine learning and user feedback loops to enable the system to continuously learn and optimize, improving performance and user satisfaction.

[0051] (4) Present the output results in a visual way to facilitate users to view and understand.

[0052] (5) It can be widely used in document reading, information retrieval, knowledge management and other fields, and has high practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0054] Figure 1 is a flow chart of the intelligent assisted reading method implemented based on the RAG large model according to the present application;

[0055] Figure 2 It is a structural diagram of an intelligent auxiliary reading system implemented based on the RAG large model according to the present application;

[0056] Figure 3 It is a structural diagram of a computer system suitable for implementing an electronic device of an embodiment of the present application. DETAILED DESCRIPTION

[0057] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0058] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0059] Figure 1 The flowchart of the intelligent auxiliary reading method implemented based on the RAG large model of the present application is shown as follows: Figure 1 As shown, the method comprises the following steps:

[0060] S1. Receive raw data in various data formats input by the user. The raw data includes documents (such as PDF, Word, TXT, etc.), pictures (such as JPEG, PNG, etc.), tables (such as Excel), voice (such as WAV, MP3, etc.), and video (such as MP4, AVI, etc.).

[0061] The raw data in various data forms are obtained from at least one data source, which includes a database, a sensor network, a user input interface or a network server. Specifically, the user can input data into the system by uploading files, inputting URL links, recording voice or video, etc.

[0062] S2. Preprocess the original data to obtain preprocessed data, wherein the preprocessing includes: noise removal, word segmentation, part-of-speech tagging, cleaning and normalization.

[0063] Preferably, in step S2, it further includes: in response to determining that the original data is voice or video data, using speech recognition technology (Automatic Speech Recognition, ASR) to recognize the original data and convert it into text content. Support recognition of multiple languages, including Chinese, English, Japanese, Korean, etc. And perform voice error correction and punctuation addition on the recognized text to improve recognition accuracy.

[0064] Preprocess the text content output by the ASR module and the plain text content in the document to remove noise, perform word segmentation, part-of-speech tagging, etc. Use natural language processing technology to clean and normalize the preprocessed text to remove stop words, special symbols, etc. And perform keyword extraction and topic analysis on the preprocessed text to provide input for the RAG large model module.

[0065] S3. Use the RAG large model to perform data analysis and key information extraction on the preprocessed data to obtain structured information.

[0066] S4. Use a data verification algorithm to verify the structured information and output the verification result.

[0067] It should be understood that the same type of information can also be extracted from multiple data sources for comparison and verification to ensure the accuracy and consistency of the information; or for questionable or inconsistent information, manual verification and correction can be performed; or machine learning models can be used to verify and predict the extracted key information to improve the accuracy and reliability of the information. Specifically:

[0068] 1. Determine the data quality indicators to calculate, including data accuracy, completeness, consistency and timeliness. These indicators help understand the overall quality of the data.

[0069] Data accuracy: Calculate the ratio of accurate data to total data. The formula is: Data accuracy = (accurate data / total data) x 100%

[0070] Data integrity: Calculate the ratio of complete data to total data. The formula is: Data integrity = (complete data / total data) x 100%

[0071] Data consistency: Calculate the ratio of the number of consistent data to the total number of data. The formula is: Data consistency = (number of consistent data / total number of data) x 100%

[0072] Data timeliness: Calculate the ratio of the last update time to the total amount of data. The formula is: Data timeliness = (last update time / total amount of data) x 100%

[0073] 2. Use Python and the pandas library to program the data in combination with the above indicator calculations to check the data quality.

[0074] 3. Build a machine learning model, use feature columns to train the model, and predict and verify the accuracy of the data.

[0075] 4. Process the verification results. For records that have successfully been verified, you can continue with subsequent processing; for records that have failed verification, record the error information and skip the record, and then notify the user or administrator to make manual corrections.

[0076] 5. Generate a verification report containing verification results, error messages, and suggested corrective actions to monitor data quality, improve data processing procedures, or train users

[0077] Through RAG technology and in combination with external knowledge bases and related documents, we conduct in-depth analysis and understanding of pictures, tables, large amounts of documents and data, and generate accurate, rich and practical outputs. At the same time, we use technical means such as data verification, multi-source comparison, manual verification and machine learning model verification to calibrate and verify the extracted key information to ensure the accuracy and reliability of the information. Finally, the processing results are output in a structured form to meet the different needs of users.

[0078] Continue to refer Figure 1 The intelligent assisted reading method based on the RAG large model provided in this application also includes the following steps:

[0079] S5. Obtain user feedback on the verification results, and adjust the parameters and strategies of the RAG large model based on the feedback information.

[0080] Specifically, users can like or dislike the results generated by the system, and submit some feedback on errors that they do not like, including wrong answers, grammatical errors, illegal, sensitive information, or other suggestions. Based on user feedback, the system can automatically adjust the parameters and strategies of the RAG large model module to optimize system performance. Then, the user feedback information is analyzed and summarized to provide a basis for further improvement of the system.

[0081] S6. Present the verification results obtained in step S4 and the results optimized according to user feedback in step S5 in multiple output forms. The output results support multiple output forms, including text, charts, knowledge graphs, etc. Users can choose different output forms according to their needs for easy viewing and understanding.

[0082] In a specific embodiment, in response to determining that the original data is an image, a deep learning model (such as a convolutional neural network (CNN)) is used to perform preliminary feature extraction and classification on the image to identify the main objects and scenes in the image.

[0083] And use optical character recognition technology (Optical Character Recognition, OCR) to extract text information in the image, the text information includes: title, label and annotation.

[0084] Use object detection algorithms (such as YOLO, Faster R-CNN, etc.) to further identify key elements in the image, including people, objects, and logos, and annotate the location and attributes of key elements. Perform semantic analysis on the extracted text information to understand its meaning and context.

[0085] Combining text information and key elements, natural language processing technology is used to generate image descriptions, and the RAG large model is used to retrieve information related to the image content from external knowledge bases or related documents to calibrate and supplement the descriptions, thereby improving the accuracy and richness of the descriptions.

[0086] The generated image description and the retrieved information related to the image content are integrated to obtain structured information. The output can include basic information of the image (such as file name, size, format, etc.), descriptive text, related links or references, etc.

[0087] In a specific embodiment, in response to determining that the original data is a table, a table parsing technique is used to extract row, column and cell data from the table.

[0088] Clean, format and normalize row, column and cell data to ensure data accuracy and consistency and obtain preprocessed data.

[0089] Apply data mining techniques (such as cluster analysis, association rule mining, etc.) to analyze preprocessed data to discover patterns and trends in the data. Use natural language processing technology to generate concise text descriptions of the table and summarize the key information and findings in the table.

[0090] The RAG big model is used to retrieve information related to the table content from external knowledge bases or related documents to calibrate and verify the analysis results. Questionable or inconsistent data are manually checked or further analyzed to ensure the accuracy of the information.

[0091] The generated table description and the retrieved information related to the table content are integrated to obtain structured information. The output structured information may include basic information of the table (such as title, number of rows and columns, data source, etc.), key data points, data trends, related links or references, etc.

[0092] In a specific embodiment, in response to determining that the original data is a large amount of document data, indexing technology and query optimization strategy are used to retrieve information related to user needs from the large amount of document data.

[0093] And use the RAG big model to retrieve information related to user needs from external knowledge bases or related documents to improve the accuracy and comprehensiveness of the retrieval.

[0094] Conduct content analysis and understanding of the retrieved relevant information, extract key information and themes,

[0095] And use natural language processing technology to perform semantic analysis and sentiment analysis on the retrieved information to gain a deep understanding of the meaning and context of documents and data.

[0096] According to user needs and data characteristics, the GPT model (or other generative language models) is combined to generate various forms of output content, including: summary, question-answer pairs, and graph recognition.

[0097] Use RAG technology to calibrate and adjust the generated output content to ensure the accuracy and usefulness of the information.

[0098] The calibrated and adjusted output content and the retrieved related information are integrated to obtain structured information. The output structured information can include basic information of the document (such as title, author, publication date, etc.), key information points, abstracts, question-answer pairs, knowledge graphs, etc. According to user needs, the output can be presented in different formats, such as text files, HTML pages, PDF reports, etc.

[0099] Further references Figure 2 As an implementation of the above method, in the second aspect, the present application provides a structural diagram of an embodiment of an intelligent auxiliary reading system 200 implemented based on the RAG large model, which can be specifically applied to various electronic devices. The system 200 includes the following modules:

[0100] The data input module 210 is configured to receive raw data in various data forms input by the user;

[0101] The preprocessing module 220 is configured to preprocess the raw data to obtain preprocessed data, wherein the preprocessing includes: noise removal, word segmentation, part-of-speech tagging, cleaning and normalization processing;

[0102] RAG large model module 230, configured to perform data analysis and key information extraction on pre-processed data using the RAG large model to obtain structured information;

[0103] The verification module 240 is configured to verify the structured information using a data verification algorithm and output a verification result;

[0104] A feedback module 250 is configured to obtain user feedback information on the verification result and adjust the parameters and strategies of the RAG large model according to the feedback information;

[0105] The result output module 260 is configured to present the verification results obtained by the verification module 240 and the results optimized according to the user feedback in the feedback module 250 in a variety of output forms.

[0106] In a specific embodiment, the system also provides an interface for users to interact with the system, text and video input, audio recording input, etc. User-defined instructions, users can create their own instructions, and then locate the big model, including role description, concept cognition, dimension type, etc., so that users can communicate with the big model and let the model find the answers users want more quickly and accurately.

[0107] The intelligent auxiliary reading system based on the RAG large model provided by the present invention can directly use the RAG large model to process pictures and tables in documents, and combine OCR (Optical Character Recognition), ASR (Automatic Speech Recognition) and other technologies to realize content recognition and extraction of documents and pictures, tables, voice, video and other data contained therein, while introducing interactive machine learning and user feedback loops to help users quickly and accurately understand document and data content, and continuously optimize system performance.

[0108] In a third aspect, the present application provides a terminal device, including a processor, a memory, and a computer program stored in the memory, wherein the computer program is executed by the processor to implement any of the above-mentioned intelligent assisted reading methods based on the RAG large model.

[0109] In a fourth aspect, the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, any of the above-mentioned intelligent assisted reading methods based on the RAG large model is implemented.

[0110] Reference below Figure 3 , which shows a schematic diagram of the structure of a computer system 300 of a terminal device or server suitable for implementing an embodiment of the present application. Figure 3The terminal device or server shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0111] like Figure 3 As shown, the computer system 300 includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage part 308 into a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the computer system 300 are also stored. The CPU 301, the ROM 302, and the RAM 303 are connected to each other through a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0112] The following components are connected to the I / O interface 305: an input section 306 including a keyboard, a mouse, etc.; an output section 307 including a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed, so that a computer program read therefrom is installed into the storage section 308 as needed.

[0113] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure 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 the network through the communication part 309, and / or installed from the removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, the above functions defined in the method of the present application are executed. It should be noted that the computer-readable medium described in the present application can be a computer-readable signal medium or a computer-readable medium or any combination of the above two. The computer-readable medium can 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 media may include, but are not limited to, an electrical connection with one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or 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 medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, an apparatus or a device. 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, which carries a computer-readable program code. 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. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable medium, which may send, propagate, or transmit a program for use by or in combination with an instruction execution system, an apparatus or a device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0114] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as C or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0115] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the 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 square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square 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 square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart 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.

[0116] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above invention concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

Claims

1. An intelligent assisted reading method based on the RAG large model, characterized in that: The method comprises the following steps: S1. Receive original data in various data forms input by the user; S2, preprocessing the original data to obtain preprocessed data, wherein the preprocessing includes: noise removal, word segmentation, part-of-speech tagging, cleaning and normalization processing; S3, using the RAG large model to perform data analysis and key information extraction on the pre-processed data to obtain structured information; S4. Verify the structured information using a data verification algorithm and output a verification result; S5. Obtain user feedback information on the verification result, and adjust the parameters and strategies of the RAG large model according to the feedback information; S6. Present the verification result obtained in step S4 and the result optimized according to user feedback in step S5 in a variety of output forms.

2. The intelligent assisted reading method based on the RAG large model according to claim 1 is characterized in that: Also includes: In response to determining that the original data is an image, performing preliminary feature extraction and classification on the image using a deep learning model to identify major objects and scenes in the image; and extracting text information from the image using optical character recognition technology, the text information including: title, label and annotation; Using an object detection algorithm to further identify key elements in the image, the key elements include: people, objects and signs, and marking the positions and attributes of the key elements; Combining the text information and key elements, using natural language processing technology to generate a description of the image, and using the RAG large model to retrieve information related to the image content from an external knowledge base or related documents; The generated description of the image and the retrieved information related to the image content are integrated to obtain structured information.

3. The intelligent assisted reading method based on the RAG large model according to claim 1 is characterized in that: Also includes: In response to determining that the original data is a table, extracting row, column and cell data from the table using a table parsing technique; Cleaning and normalizing the row, column and cell data to obtain preprocessed data; Applying data mining technology to analyze the preprocessed data and using natural language processing technology to generate a description of the table; Retrieving information related to the table content from an external knowledge base or related documents using the RAG big model; The generated description of the table and the retrieved information related to the table content are integrated to obtain structured information.

4. The intelligent assisted reading method based on the RAG large model according to claim 1 is characterized in that: Also includes: In response to determining that the original data is a large amount of document data, using indexing technology and query optimization strategy to retrieve information related to user needs from the large amount of document data; and using the RAG big model to retrieve information related to the user needs from an external knowledge base or related documents; Conduct content analysis and understanding of the retrieved relevant information, extract key information and topics, and use natural language processing technology to perform semantic analysis and sentiment analysis on the retrieved information; Based on user needs and data characteristics, the GPT model is combined to generate various forms of output content, including summary, question-answer pair, and graph recognition; Calibrate and adjust the generated output content using RAG technology; Integrate the calibrated and adjusted output content with the retrieved related information to obtain structured information.

5. The intelligent assisted reading method based on the RAG large model according to claim 1 is characterized in that: Also includes: In step S1, the original data includes documents, pictures, tables, voice and video data.

6. The intelligent assisted reading method based on the RAG large model according to claim 1 is characterized in that: In step S2, it also includes: in response to determining that the original data is voice or video data, using voice recognition technology to recognize the original data and convert it into text content.

7. The intelligent assisted reading method based on the RAG large model according to claim 1 is characterized in that: In step S1 , raw data in various data formats are acquired from at least one data source, wherein the data source includes a database, a sensor network, a user input interface, or a network server.

8. An intelligent auxiliary reading system based on the RAG large model, characterized in that: The system comprises: A data input module, configured to receive raw data in various data forms input by a user; A preprocessing module is configured to preprocess the raw data to obtain preprocessed data, wherein the preprocessing includes: noise removal, word segmentation, part-of-speech tagging, cleaning and normalization processing; A RAG large model module, configured to use the RAG large model to perform data analysis and key information extraction on the pre-processed data to obtain structured information; A verification module, configured to verify the structured information using a data verification algorithm and output a verification result; A feedback module, configured to obtain user feedback information on the verification result, and adjust the parameters and strategies of the RAG large model according to the feedback information; The result output module is configured to present the verification results obtained by the verification module and the results optimized according to user feedback in the feedback module in multiple output forms.

9. A terminal device, characterized in that: It comprises a processor, a memory and a computer program stored in the memory, wherein the computer program is executed by the processor to implement the intelligent assisted reading method based on the RAG large model as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, wherein a computer program is stored in the medium, and when the computer program is executed by a processor, the intelligent assisted reading method based on the RAG large model as described in any one of claims 1 to 7 is implemented.

Images