20 results about "Document filtering" patented technology

The invention discloses a PDF (Portable Document Format) document structured extraction system based on a multi-modal language model, belongs to the technical field of document processing and optical character recognition, and aims at solving the technical problem of how to improve the existing OCR (Optical Character Recognition) technology to improve the analysis capability of a complex document structure and improve the recognition precision of handwritten forms and other non-standard fonts. According to the technical scheme, the system adopts a layered decoupling architecture and comprises an input layer, a preprocessing layer, a reasoning layer, an output layer and a monitoring and fault-tolerant module; wherein the output layer is used for multi-source data access and path management to realize a local file system or S3 cloud storage; the preprocessing layer is used for invalid document filtering and visual feature extraction; the reasoning layer is used for multi-modal model interaction and content processing; the output layer is used for outputting a content aggregation result; and the monitoring and fault-tolerant module is used for realizing real-time state monitoring, resource consumption analysis and exception handling.

1. Name of the product in this design: Graphical User Interface for a Multimodal Knowledge Management AI Knowledge Base Platform for Computers. 2. Purpose of this design: To present the graphical user interface of the AIcoach knowledge base on the display screen of a computer device, enabling the input, management, and retrieval of scripts and document knowledge. 3. The key design features of this product are the interface design in each view. 4. Images or photos that best illustrate the design points: Interface change state diagram 4. 5. Purpose of the graphical user interface: The graphical user interface is used for the operation and information interaction of the AIcoach knowledge base, including: login portal, script / document filtering and retrieval, script editing and publishing, knowledge import processing and prompts, online customer service dialogue and source tracing, and brand platform attribute configuration management. 6. Other situations requiring explanation are illustrated in the diagrams: Clicking "Login" in the main view will redirect to the interface change state diagram 1; clicking "View Knowledge Base Management Terminal" in the interface change state diagram 1 will redirect to the interface change state diagram 2; clicking "Go to Client" in the interface change state diagram 2 will redirect to the interface change state diagram 3; clicking "Ask a Question" in the interface change state diagram 3 will redirect to the interface change state diagram 4.

This invention discloses a code security scanning method, comprising: project file filtering: filtering project files in a source code repository according to a preset filtering strategy to generate a set of files to be scanned; scan scheduling: scheduling a first scanning tool and / or a second scanning tool to perform code security scanning on the set of files to be scanned according to the scan mode selected by the user, and obtaining a set of scan alert results, wherein the first scanning tool is a static scanning tool based on regular expressions or abstract syntax trees, and the second scanning tool is a deep scanning tool based on graph analysis; result standardization: deduplicating the scan alert result set to obtain a standardized scan alert dataset; report generation: generating a standardized scan report based on the standardized scan alert dataset. By scheduling the corresponding scanning tool to perform code security scanning on the files to be scanned according to the scan mode selected by the user, multiple scanning modes are provided to improve scanning accuracy.

The invention provides a file access system and method, electronic equipment, a storage medium and a program product, and belongs to the technical field of computers, and the system comprises a sending module which creates a file object corresponding to an access request based on the access request of an application program, and sends the file object to a file filtering driving module, the access request comprises an identifier of a to-be-accessed file; the file filtering driving module is used for determining a target cache region accessed by the access request according to the permission type of the application program, the identifier of the to-be-accessed file and a first section of object pointer and a second section of object pointer associated with a file object in the stream context information; and the encryption and decryption module is used for determining an encryption and decryption strategy for the to-be-accessed file in the target cache region based on the permission type of the application program and the type of the access request. The method and the device are used for solving the problem of how to stably and conveniently realize secure encryption and decryption caching of enterprise data.

The invention discloses an RAG document filtering method, device and system and a storage medium. The RAG document filtering method provided by the embodiment of the invention comprises the following steps: deducing candidate answers according to user query and each candidate document through a first language model agent to form triple information consisting of the user query, the candidate documents and the candidate answers; performing correlation evaluation on the triple information through a second language model agent to obtain a correlation score corresponding to each candidate document; carrying out filtering processing on the plurality of candidate documents based on the correlation scores to obtain related documents; and through the third language model agent, generating a related answer corresponding to the user query according to the related document. Thus, by introducing a retrieval enhancement generation framework based on multi-agent filtering and utilizing cooperation among a plurality of language model agents, the retrieved document is precisely filtered and scored, so that the retrieval document quality and the answer accuracy of the RAG system are remarkably improved.

The invention provides a method and device for constructing a file filter, and relates to the technical field of data processing. The problems that effective integration of resources and optimization of I O performance cannot be achieved due to the fact that the used filter is not matched with the requirement of the current application scene are solved. Obtaining input parameter information of a to-be-constructed file filter, wherein the input parameter information comprises one or more of the following information: the maximum file quantity supported by a target folder, the file quantity currently stored in the target folder, the available storage space corresponding to the target folder, and the file quantity currently stored in a data file system; based on the input parameter information, determining performance parameters of each type of file filter in multiple types of file filters; determining a target type of the to-be-constructed file filter at least based on the performance parameters of each type of file filter; and constructing a file filter of the target type.

The invention provides a system security protection method and device, electronic equipment and a storage medium, and relates to the technical field of network security. The method comprises the following steps: monitoring a client based on a configuration rule of a preset virtual machine control structure, and determining an event type and an instruction for triggering an event when monitoring that the client triggers a virtual machine exit event; determining a reason for triggering the virtual machine to exit the event based on the event type, the instruction and the configuration rule; when the reason is that the client executes the system call entry instruction, redirecting the read-write operation of the instruction on the target register to a preset corresponding shadow register; when the read-write operation is to read a system call kernel entry address, determining a file operation request of the client based on a system call number and parameters associated with a system call entry instruction; a file path of the file operation request is redirected to a file filtering module in a kernel mode of the client, and the control right of the central processing unit is returned to the client, so that the client sends the file operation request to the kernel mode of the client.

The invention provides an off-line file automatic encryption synchronization method and system, and the method comprises the steps: intercepting an I / O request through a kernel mode file filter driver, and initializing monitoring when a write release event is captured; performing stability judgment based on a sliding window on the monitored object, calculating a discrete difference fluctuation ratio, and confirming that writing is complete after convergence; calculating the local information entropy of the file to generate a weighted feature sensing fingerprint, and performing duplicate removal through Hamming distance comparison; generating a dynamic key based on the system epoch time to perform streaming encryption; only uploading the missing ciphertext data block according to a missing bitmap fed back by the server; and the server side decrypts the restored file and generates a chained hash log for solidification storage. According to the method, the file writing state can be accurately judged, data damage is prevented, and efficient duplicate removal, incremental transmission and whole-process credible traceability are achieved.

The invention provides a similar document filtering method and device based on condition awareness, equipment and a medium, and the method comprises the steps: obtaining user question information, and determining text capture condition information according to the user question information and a preset condition rule; matching the text capture condition information with metadata corresponding to each text block in a pre-established knowledge base, and screening out successfully matched text blocks; the text blocks are obtained by dividing text contents obtained by identifying documents in a pre-established document library; the metadata is obtained by analyzing a document directory configured by a document in a document library through the condition rule; and constructing a text block list for displaying the successfully matched text blocks. According to the method, similar documents can be reasonably distinguished, the question answering complexity is reduced, and the answering precision is improved.

The present disclosure relates to an artificial intelligence (AI)-powered computer-implemented system (100) for connecting school leaders with informative resources that are aligned with industry-specific educational standards. The system (100) comprising a user device (102). The system (100) also comprising a document filtering module (104) configured to analyses and filter documents based on relevance to the received user inputs. The system (100) also comprising an intent recognition module (106) configured to interpret the user inputs to identify the user's intent and specific requirements. The system (100) also comprising a cloud-based application programming interface (API) module (114) and a page ranking module (116). The system (100) also comprising an artificial intelligence-based processing assembly (118) further having a document searching module (120), a text extraction module (122), and a summarizer (124). The system (100) also comprising an implementation, tracking, and evaluation module (ITEM) (126) configured to enable the users to apply best practices by providing recommendations, tracking progress, and evaluating outcomes.

The application discloses a multi-modal document reordering method and system based on answer quality driving and a medium, relates to the technical field of artificial intelligence, and inputs a query question into a rearrangement large model to output a document rearrangement result; the training process of the rearrangement large model is as follows: an initial candidate document set corresponding to the query question is acquired through a multi-modal retriever and is taken as the input of a rearranger, and through screening and rearrangement, a rearranged document subset and a thinking process are generated; the document subset and the thinking process are input into a sandbox, the sandbox is internally deployed with a frozen parameter answer model, is used for parallel processing of multiple different rearrangement results of the same question, and generates corresponding multiple answers; a format consistency reward is set by using the output of the rearranger, a document filtering reward is set by using the rearranged document subset, and an answer quality reward is set by using the similarity between the answers and standard answers, so as to optimize the parameters of the rearranger; and the reordering method improves the document sorting quality and the question and answer accuracy.

This disclosure provides a document analysis method, apparatus, device, and storage medium. This disclosure relates to the field of big data, particularly to the technical fields of large models and data filtering. The specific implementation scheme is as follows: A large model is used to analyze decision reference information to determine a document filtering method and the filtering parameters required for that method; based on the document filtering method and the filtering parameters, a coarse screening of a mixed document set is performed to obtain a first document set; finely screened documents are selected from the first document set; related documents of the finely screened documents are selected from the mixed document set to obtain a document set to be described, with the finely screened documents as the core; and descriptive information for the document set to be described is generated based on the large model. In this embodiment, the large model can flexibly determine the document filtering method, filter out the required document set to be described, and provide rich descriptions.

The invention discloses a two-stage document filtering and robust fine tuning method based on a graph attention network, and relates to the technical field of natural language processing and deep learning. The problems that an existing retrieval enhancement generation system is difficult to accurately identify useful documents in a mixed document environment and is easily interfered by anti-fact information are solved. The method comprises the following steps of: constructing a multi-class document pool comprising correct documents, anti-fact documents, noise documents and irrelevant documents, and constructing a semantic graph on a paragraph level; the method comprises the following steps: sequentially filtering irrelevant documents and noise documents by adopting a two-stage graph attention network to obtain a reference document set, constructing a document discrimination training sample and a question and answer training sample based on the reference document set, and combining the document discrimination training sample and the question and answer training sample into joint fine tuning data to perform instruction fine tuning on a large language model. And the model has a document reliability discrimination capability and keeps stable output in a mixed document situation, so that the factuality and credibility of a retrieval enhancement generation system in an anti-fact attack and noise environment are improved.

A two-stage document filtering and robust fine-tuning method based on graph attention networks is proposed, relating to the fields of natural language processing and deep learning. It addresses the problems of existing retrieval augmentation generation systems struggling to accurately identify useful documents in mixed document environments and being susceptible to counterfactual information interference. This method constructs a multi-class document pool including correct documents, counterfactual documents, noisy documents, and irrelevant documents, and builds a semantic graph at the paragraph level. A two-stage graph attention network is used to sequentially filter irrelevant and noisy documents to obtain a reference document set. Based on this reference document set, document discrimination training samples and question-answering training samples are constructed. These two sets are combined into joint fine-tuning data to fine-tune a large language model, enabling the model to possess document reliability discrimination capabilities and maintain robust output in mixed document scenarios. This improves the factual accuracy and credibility of the retrieval augmentation generation system under counterfactual attacks and noisy environments.

The invention discloses a kernel-based virus searching and killing method, device and equipment and a storage medium, and relates to the technical field of information security, the kernel-based virus searching and killing method comprises the following steps: determining a target file operation on a target file, and determining a first target processing function corresponding to the target file operation; the first target processing function is a function which is configured through a file filtering drive of a kernel and corresponds to the target file operation; identifying whether the target file operation is a virus attack or not through a first target processing function, and if the identification fails, identifying whether the target file operation is the virus attack or not through a target antivirus engine; and if the target file operation is identified as a virus attack, intercepting the target file operation, generating virus attack risk prompt information, obtaining confirmation information generated by a user based on the virus attack risk prompt information, and if the confirmation information represents that interception is agreed, stopping execution of the target file operation, and generating a corresponding processing log. According to the invention, efficient and accurate virus searching and killing of the file can be realized.

The invention relates to a power unstructured file use control method and device, computer equipment, a computer readable storage medium and a computer program product. The method is applied to an electric power trusted data space, and comprises the following steps: determining an electric power unstructured file, file information and a user main body, and obtaining file use requirement information; based on the file information and the file use requirement information, determining a use control scheme for the power unstructured file; intercepting a use operation for the power unstructured file according to the file filter driver; determining a user permission corresponding to the use operation and a permission requirement corresponding to the use operation in the use control scheme; under the condition that the user permission meets the permission requirement, the use operation is allowed; and prohibiting the use operation under the condition that the user permission does not meet the permission requirement. By adopting the method, the power unstructured file circulation based on the power trusted data space can be realized on the basis of ensuring the data security.

Various embodiments of methods, systems and computer program products described herein are directed to a Filter Engine. The Filter Engine includes a supervisor component and one or more File Input / Output (“I / O”) filters. The Filter Engine instantiates one or more file input / output (“file I / O”) filter modules that correspond with a file system, each respective file I / O module assigned to a different type of file system operation. A first file I / O filter module is assigned to a first type of file system operation. Prior to execution of an instance of a first type of file system operation: (i) a first file I / O filter module detects a first call for the instance of the first type of file system operation and determines whether to provide access to file system decoy data to a source of the first call.

This application discloses a method, apparatus, and electronic device for acquiring motion information. Relating to the field of artificial intelligence, the method includes: querying M source documents in a database and generating N query statements based on the requirements for acquiring target motion information; performing query operations on the M source documents using each query statement to obtain a set of N source documents; performing tag matching operations on each source document in the N source document sets to obtain tag matching results, and performing document filtering operations on the N source document sets based on the tag matching results to obtain P target documents; splitting the P target documents according to the document structure of each target document to obtain multiple text blocks, and identifying the multiple text blocks as motion information of the target motion. This application solves the problem of low accuracy in determining motion information based on knowledge documents, literature, and other information collected from the internet in related technologies.

The invention discloses a data recovery method, a storage medium and equipment, a file filter driver is combined with an interception program in a shutdown process to realize a system recovery function, and the file filter driver is adopted to monitor an operation behavior of disk data without influencing operation of other processes in a system, so that the system recovery efficiency is improved. According to the method and the device, the data of the whole disk does not need to be backed up, the performance loss of the operating system is reduced, meanwhile, the disk data is restored through the stored operation record by monitoring and recording the operation of the disk data in the shutdown process, the restoration process of the device is realized from the disk data level, and the reliability of the device is improved. And the whole operating system does not need to be restored, so that the data restoration efficiency is improved.