129 results about "Content extraction" patented technology

The invention provides a full-text retrieval method and system fusing various types of documents, and relates to the technical field of information retrieval, and the method comprises the following steps: obtaining document representation through document content extraction and structure recognition, generating a cross-modal semantic vector by using word embedding and nonlinear transformation, constructing a hierarchical index and a cross-document association graph, and obtaining a full-text retrieval result; the basic correlation score is calculated after the query request is received, and the comprehensive score of the candidate content segments is calculated based on the association graph to determine the optimal retrieval result, so that unified representation and retrieval of heterogeneous documents are realized, the cross-document retrieval precision and relevance are improved, and the processing capability of a retrieval system on complex queries is enhanced.

The invention belongs to the technical field of artificial intelligence, and discloses a knowledge extraction method and system based on semantic consistency evaluation and hybrid verifiable reward, and the method comprises the steps: constructing a training data set with evidence labeling; based on the training data set, reinforcement learning training is carried out on a pre-trained large language model, a group strategy optimization GRPO algorithm is adopted, and model output is evaluated by using a mixed reward function; and on the basis of an evaluation result of the mixed reward function, updating parameters of a large language model so as to generate structured knowledge which is correct in format, accurate in content and provided with verifiable evidence. According to the method, in reinforcement learning training, effective decoupling format, content and credibility evaluation is realized, and refined feedback is provided for the model; when the model outputs knowledge, traceable original text evidence is provided for the model, so that the credibility and the interpretability of the model are enhanced, the accuracy of outputting the JSON format by the model is improved, and the accuracy and the integrity of the model in the aspect of content extraction are enhanced.

A large model memory enhancement method based on time perception consistency feedback and optimization comprises the following steps: 1, data preprocessing: obtaining a historical interaction text generated by each agent in a target scene, and processing the text by using a preprocessing module; 2, attribute mining: extracting agent personal attributes, site attributes, logic attributes and core arguments from each generated content; and 3, keyword management and memory storage: maintaining an independent keyword historical file and a memory library file for each agent, and constructing an evolvable memory which changes along with time. 4, performing memory retrieval and context construction; 5, performing multi-dimensional consistency evaluation, and generating a comprehensive score; and 6, self-adaptive optimization is carried out, and self-feedback and self-evolution of model behaviors are realized. By means of the method, the memory ability and historical consistency of the multi-agent large language model can be effectively enhanced.

The invention provides an intelligent marketing decision analysis method and system for a real-time competitive product strategy, and the method comprises the steps: collecting the marketing content of a competitive product published on a social media platform, extracting an account number, an arriving person portrait feature, a content structure label, a platform identifier and publishing time, and constructing a standardized strategy behavior vector; performing time sequence alignment on the strategy behavior vectors and our marketing indexes, and identifying competing product strategy behavior subsets which have significant influence on the our indexes; based on the subset, generating a structured response strategy containing a recommended person type, a content structure label, a delivery platform and a time period; and performing multi-dimensional matching on the marketing resource library and the content template library, and outputting an executable marketing resource combination and scheduling instruction. According to the method, a closed loop from competitive product behavior perception and causal influence identification to response strategy automatic generation and landing execution is realized, and the real-time performance, the accuracy and the automation level of marketing response are improved.

The invention provides a demand specification generation method and system based on multi-modal understanding, and relates to the field of artificial intelligence, and the method comprises the steps: carrying out the content understanding operation of data of each modal through employing a corresponding content extraction tool, so as to generate a demand representation document after multi-modal unification; establishing a demand knowledge base according to an existing template base and a corresponding rule base, matching and comparing the demand knowledge base with the demand representation document, checking the integrity of the demand representation document, and generating a question for seeking a decision; creating an overall framework of the standard document according to the standard standard template, and performing content filling on the overall framework according to the demand representation document and an answer to the question seeking the decision to generate the standard document; and collecting feedback information for the standard document, analyzing the feedback information by using a large language model, and converting the feedback information into executable modification suggestions for the standard document. According to the method, the automation degree, the integrity and the accuracy of demand specification generation are improved.

Embodiments described herein relate to systems and methods for automatically generating content for a generative answer interface of a collaboration platform. The system receives a natural language user input identifying corresponding blocks of text or snippets using a content extraction service. A prompt is generated using the blocks of text and is used to obtain a generative response. The generative response and links to corresponding content are displayed in the generative answer interface and can be inserted into content of the collaboration platform. The systems and methods described use a network architecture that includes a prompt generation service and a set of one or more purpose-configured large language model instances (LLMs) and / or other trained classifiers or natural language processors used to provide generative responses for content collaboration platforms.

A content extraction method and apparatus including presenting a document selection interface based on an uploading operation in an information exchange interface, the document selection interface comprising at least one document, based on a selection operation on a first document of the at least one document, presenting, in the information exchange interface, the first document and processing information of the first document in a status area corresponding to the first document, the processing information indicating a current processing stage in a process of performing content extraction on the first document and a corresponding processing status, and after the content extraction on the first document is completed, removing display of the status area, and presenting, in the information exchange interface, a document digest obtained by performing content extraction on the first document.

A system is provided.SOLUTION: A system, comprising: means for selecting or uploading a video that a user wants to view; a server for storing and analyzing video files; an artificial intelligence engine for extracting content of the analyzed video as feature vectors and identifying scenes containing inappropriate expressions; means for tagging the inappropriate scenes identified by the artificial intelligence engine and notifying a parent dashboard; means for a parent to confirm and approve modification of the inappropriate scenes; an editing module for converting the inappropriate scenes into appropriate content; and means for generating and providing an edited video to a user terminal.SELECTED DRAWING: Figure 1

The invention provides a professional process enhancement generation method, electronic equipment and a computer program product. The professional process enhancement generation method comprises the following steps: acquiring associated context information of a paragraph to be generated from a memory according to reported state information of the paragraph to be generated; constructing a cue word of a paragraph to be generated, wherein the cue word comprises the associated context information; inputting the cue word into a large language model, and enabling the large language model to generate the content of the paragraph to be generated according to the associated context information; and extracting context information according to the content of the paragraph to be generated, and updating the context information into the memory. According to the method, the stability of the hierarchical structure of the paragraph of the generated content can be ensured, the style consistency and the logic coherence in long text generation are improved, and the quality of a vertical professional field report generated by a large language model is improved.

The invention provides a contract content extraction and integration method and system and a storage medium, and the method comprises the steps: obtaining a contract file uploaded by a user through an OA system, recognizing the context features of the contract file, and dynamically triggering an analysis strategy adaptive to the contract file based on the context features; performing content extraction on the contract file by using an analysis strategy to generate a preliminary analysis result containing core content and risk reminding; obtaining the complexity of the contract file and the confidence coefficient of the initial analysis result, and determining an auditing path of the initial analysis result based on the complexity and the confidence coefficient; and auditing the preliminary analysis result by using the auditing path to generate auditing feedback information, and correcting the preliminary analysis result in real time based on the auditing feedback information to generate an optimized analysis result. According to the method and the device, intelligent and dynamic analysis and auditing guidance of the contract file can be realized in an OA system environment, the efficiency and the accuracy of contract auditing are effectively improved, and adaptive adjustment of an analysis strategy and an auditing path is realized.

The invention discloses an automobile data recorder video recovery and recombination method and device, and relates to the technical field of video data recovery, and the method comprises the steps: extracting a reference file to obtain a packaging format, a coding format and a corresponding video feature parameter set; scanning the damaged automobile data recorder, and extracting a decoding parameter set of a to-be-repaired video and a frame set of the to-be-repaired video from the scanned content based on the packaging format and the coding format of the reference file; decoding the video frames in the frame set and identifying burning timestamp information through OCR (Optical Character Recognition) in sequence to obtain an ascending-sequence video frame set; performing parameter updating on the video characteristic parameter set by using the decoding parameter set of the to-be-repaired video to obtain an updated video characteristic parameter set; and based on the updated video characteristic parameter set, repackaging the ascending-order video frame set to generate a video file. According to the method, the frame reordering driven by the OCR timestamp is utilized, so that the problems of discontinuous restored video playing pictures and out-of-order time are effectively solved.

This invention provides a method and system for extracting content from long documents based on dynamic segmentation and multi-model optimization. The method includes: identifying chapter boundaries and paragraph boundaries of the long document to be extracted; selecting combinations of pending segmentation nodes according to preset filtering rules, and segmenting the long document into different sets of segmented documents based on different combinations of pending segmentation nodes; calculating the boundary recognition accuracy of different sets of segmented documents, and calculating the semantic unit integrity of different sets of segmented documents, and taking the set of segmented documents that achieves the best combination of boundary recognition accuracy and semantic unit integrity as the optimal set of segmented documents; calling multiple AI large model interfaces, and for each segmented document, evaluating the output of each AI large model call in multiple dimensions, taking the one with the highest evaluation score as the final result of content extraction for the current segmented document, and integrating the final results of content extraction from all segmented documents included in the optimal set of segmented documents to obtain the long document content extraction result.

Systems and methods for processing natural language text using a graph obtain a natural language text and a query text, and parse that the natural language text into the plurality of text units, associating each with a graph node, and removing information leak text units from the plurality of text units. Connecting relationship between at least two of the remaining set of the plurality of text units are determined and associated with a graph edge between graph nodes. Based on the probabilistic relations between each graph node and the query text, graph node restart probabilities are determined for one or more of the graph nodes. The graph nodes that can be ranked.

The invention discloses a material standardization method and device, and relates to the technical field of computers. A specific embodiment of the method comprises the steps of obtaining material description content, and determining at least one input type corresponding to the material description content; aiming at each input type, determining type content of the material description content aiming at the input type; extracting a type text corresponding to the type content; combining the type texts corresponding to the input types to generate a combined description text corresponding to the material description content; extracting at least one attribute key value pair from the combined description text; wherein the attribute key value pair comprises an attribute name and an attribute value; and according to the at least one attribute key value pair, matching a standard material corresponding to the material description content from a standard material library. According to the embodiment, the material description content can be subjected to standardization processing, the problem of material identification errors is reduced, and the product quality and equipment safety of enterprises are guaranteed.

The application discloses a kind of progressive multimodal semantic alignment's teaching video content extraction method, belong to video abstract generation technical field, including: sampling education video, obtain video frame sequence and corresponding sentence sequence;Video frame sequence is input into visual feature extraction model to obtain video frame feature representation, sentence sequence is input into text feature extraction model to obtain sentence feature representation;Establish progressive multimodal semantic alignment model;The video frame sequence, sentence sequence, video frame feature representation and sentence feature representation obtained by the teaching video to be extracted are input into progressive multimodal semantic alignment model, obtain fine-grained video frame abstract set and fine-grained text abstract set, that is, the abstract content extracted from teaching video.The method can effectively utilize the semantic association between multi-modal, generate more accurate, smooth teaching video abstract content.

The application relates to the technical field of big data analysis, and particularly discloses a content extraction and delivery method based on multi-modal business order videos, which comprises the following steps: calling a preset domain name-brand mapping library to output a preliminary brand identification or extract to-be-verified brand information and obtain directional multi-modal features; determining a core brand based on the preliminary brand identification or the directional multi-modal features, the to-be-verified brand information and a brand vector library, and performing standardization processing to obtain a standardized brand identification; based on the standardized brand identification, combining platform characteristics and target user portrait data of the business order video, calling a delivery strategy library to generate a directional delivery strategy of an adaptive brand type, and pushing the business order video to a target user group according to the directional delivery strategy, while monitoring delivery effect indexes in real time to obtain monitoring data, adjusting current multi-dimensional weights according to the monitoring data, and generating an effect report; the brand identification accuracy in the business order video is improved, and the delivery strategy can be continuously optimized.

The network threat knowledge graph construction method and related equipment provided by the embodiments of the present application comprise: obtaining report image data and report text data from a same network attack report, and respectively performing classification mapping on the report image data and the report text data based on a threat intelligence type set to obtain image categories and text categories; performing content extraction on the report image data based on the image categories to obtain image entity relationship data, and performing content extraction on the report text data based on the text categories to obtain text entity relationship data; performing alignment verification on the image entity relationship data and the text entity relationship data belonging to a same network threat type to obtain an alignment verification result; when there is a difference between the image and text data in a representation entity data verification group, performing multi-dimensional conflict processing to obtain target entity relationship data, and constructing a network threat knowledge graph, thereby effectively improving the data accuracy of network threat knowledge graph construction.

This invention discloses a method for intelligent routing and unified adaptation of large language models. It defines all access details of the model through a declarative configuration file, allowing new models to be added without writing any code, achieving zero-code access for large language models. It provides a completely consistent calling interface for upstream applications, shielding the heterogeneity of all downstream large language models and constructing a unified request and response abstraction layer. Through a strategy engine and JSON path technology, it accurately handles complex streaming responses, including thought content, with intelligent parsing and content extraction. It supports dynamic configuration and intelligent strategy hot updates. It achieves intelligent model routing, dynamically selecting the optimal large language model instance. Simultaneously, it provides enterprise-level governance capabilities, integrating circuit breaking, degradation, rate limiting, and monitoring functions to ensure stability for model calls. This comprehensively improves system maintainability, scalability, and user experience consistency.

The application discloses a kind of based on the system and method for diagnosing abnormal delay of business message of multi-source data, it is related to business message analysis technical field, the present application is from at least two independent physical channel real-time acquisition business message stream, message content is parsed, and the business identifier generated by information source end is extracted;Analysis data quality evaluation index of each physical channel;Confirm reference channel based on data quality evaluation index;The message describing the same market event in different physical channels is matched and aligned, and the completeness index, jump point time difference and continuity difference index are analyzed;The completeness index, jump point time difference and continuity difference index are combined into hybrid feature pair;Known channel abnormal type label is injected into hybrid feature pair for classification training;The hybrid feature pair generated based on real-time running data is screened candidate abnormal channel;Positioning main abnormal channel leading to abnormality;And generate diagnosis report.The accuracy of data analysis is improved.

This invention discloses an AI-based personalized workplace learning system and method for enterprises, comprising: an enterprise profile construction module, which collects multi-source data and constructs a digital profile of the enterprise through multimodal semantic analysis; a dynamic learning system generation module, which automatically generates a customized workplace learning system based on the enterprise profile using a large language model combined with thought chain and retrieval enhancement generation technology; a multimodal content extraction and AIGC production module, which reconstructs knowledge resources in its proprietary copyright content library into lightweight workplace knowledge units through a RAG and multi-agent collaborative architecture; and a personalized recommendation engine, which adopts an improved DSSM dual-tower model, introduces a dynamic attention feature cross-layer, and combines multi-source user features and enterprise context to achieve accurate push notifications. This invention achieves dynamic matching of the learning system with enterprise strategy, large-scale reconstruction of knowledge assets, and deep personalized recommendation, effectively solving the problems of rigid training systems and single recommendation dimensions in existing technologies.

Disclosed are systems and methods that provide a decision-intelligence (DI)-based, computerized framework for deterministically identifying and extracting content from network resources, and generating focused content based therefrom for delivery to electronic users. The framework enables real-time customization of extraction tasks, which can cause tailored, accurate results that are contextually relevant and tied into the purpose of the extraction task. This data can then be compiled and / or leveraged to generate content campaigns that can target specific sets of users, geographies, time periods, trends and the like. The framework can leverage a large language model (LLM) to seamlessly extract relevant information from network resources, which enables the generation and execution of extraction requests that can be dynamically executed and updated, which can enable the framework to “drill-down” on contextual and / or topical aspects of categories of data.

The embodiment of the invention relates to a video plot information generation method and device, equipment and a storage medium. The method comprises the following steps: firstly, obtaining evaluation data of a target video; clustering processing is carried out on film review data in the evaluation data to obtain film review data under multiple types, and evaluation theme content which corresponds to the type and is related to the plot of the target video is extracted from the film review data under the same type; generating model cue words based on the evaluation subject contents corresponding to the multiple types, inputting the model cue words into a plot information generation model, and driving the plot information generation module to generate first plot information of the target video based on the model cue words; and further, storing the first plot information of the target video into a retrieval knowledge base. According to the embodiment of the invention, clustering processing and theme content extraction are carried out on the film review data of the target video, so that the wonderful plot information of the target video is generated by utilizing various types of evaluation theme contents.

This application relates to the field of data extraction technology, specifically to a method for extracting content from unstructured reports based on knowledge graphs. This method includes: extracting candidate columns from a two-dimensional image by converting its layout features into a one-dimensional signal; constructing a coordinate mapping relationship that conforms to physical monotonicity constraints using an ordinal-preserving regression algorithm, and quantifying the degree of local stretching as a local coordinate scaling factor; dynamically adjusting the matching cost using the local coordinate scaling factor, and searching for the global optimal solution using a dynamic programming algorithm; updating the original standard field width ratio to obtain the final width ratio, regenerating the baseline coordinate sequence, and triggering the entire process calculation. This application aims to introduce a nonlinear mapping and deformation weighting mechanism constrained by monotonicity, solving the frequent column misalignment problem in existing linear template matching techniques when processing non-uniformly distorted reports.

The present specification provides a text content extraction method and device, wherein the text content extraction method comprises: performing character recognition processing on a target image, and determining a mark image containing character labels according to a processing result; inputting the mark image containing the character labels into an image processing model for processing, and obtaining a reference image containing predicted lines output by the image processing model; determining mark information corresponding to the character labels in the mark image and line information corresponding to the predicted lines in the reference image; calculating connection parameters according to the mark information and the line information, and constructing a target relationship graph corresponding to the target image based on the connection parameters; and extracting text content in the target image by using the target relationship graph, thereby improving the text completeness degree of text content extraction in the target image, and improving the efficiency and accuracy of text extraction.

The invention discloses an electroencephalogram signal enhanced intelligent video abstraction method and system, relates to the technical field of video content understanding and electroencephalogram emotion analysis, and aims to solve the technical problems that in the prior art, video abstraction cannot reflect subjective emotion experience of viewers, electroencephalogram and video time axes are difficult to align, and heterogeneous data are difficult to fuse. Acquiring an electroencephalogram signal of a viewer in a video watching process, synchronizing the electroencephalogram signal with a video clip, and extracting and generating a structured video abstract based on multi-modal content; frequency domain features, time domain features, time frequency features and spatial correlation features are obtained according to the electroencephalogram signals, and electroencephalogram feature visual representation is generated; and fusing the video content abstract, the video key event timestamp and the electroencephalogram feature visual representation, deducing the emotional state and aligning the emotional state with the video event, and outputting a personalized video abstract reflecting the real emotional experience of a viewer. The method can be used for application scenes such as intelligent video abstraction, video platform recommendation, immersive content understanding and emotion calculation.

The invention provides a heterogeneous file standardization processing method and device based on an OCR and a large model, and relates to the technical field of intelligent document paragraption.The method comprises the steps that a to-be-processed file is obtained, content extraction is conducted on the basis of a file name suffix of the to-be-processed file, and extracted content is obtained; generating a dynamic cue word based on an anchor point variable by utilizing the extracted content, a pre-constructed self-defined file type library and a pre-generated dynamic cue word template; and inputting the dynamic cue word into a large model for analysis to obtain a standardized analysis result in a preset format. According to the heterogeneous file standardization processing method and device based on the OCR and the large model, the new file type can be expanded without modifying codes, and the structured data are stably output, so that the adaptability, the accuracy and the processing efficiency are remarkably improved.

The invention discloses a method for writing translated characters in cad file AI translation, and relates to the technical field of cad file processing. The method comprises the following steps: S1, structured analysis and content extraction: reading a cad source file, analyzing the internal structure of the cad source file, and extracting all character entities and all associated graphic attributes thereof; s2, AI multi-dimensional content classification: performing language classification, entity type association classification and technical semantic recognition on the extracted character strings; and S3, processing and generating branch path contents. According to the method, through the structured analysis and accurate attribute binding write-back process, characters and attributes of the characters can be comprehensively and accurately extracted, accurate write-back is conducted after translation, all non-text graphic attributes are reserved completely, the problems that in a traditional method, information extraction is incomplete, attributes are lost and the like are effectively solved, the accuracy and format consistency of a cad file after translation are guaranteed, and the user experience is improved. And the requirements on high quality of cad file translation in the CAD application field of various industries such as engineering design, manufacturing industry and electronic information are met.

This invention relates to the field of contract management technology in the construction and real estate industry, and particularly to an intelligent element extraction method and system for contracts in the construction and real estate sector. The invention includes: standardizing the format and extracting the content of contract documents to obtain extracted files; defining semantic feature rules for the extracted files to obtain a graph structure; performing divide-and-conquer parallel processing on the graph structure to obtain processed subtasks; and performing deep validation on the processed subtasks to obtain structured data. This invention, through the formal representation of graph theory structures, enables rules to handle nested, branching, and looping logic, thereby significantly improving the accuracy and coverage of extraction and reducing missed and false extractions.

A system and method for efficiently extracting and categorizing business information from online sources is disclosed. The system comprises a web crawler that obtains company domains from a database and collects depth-1 URLs from company homepages. A classification model, utilizing a fine-tuned BERT architecture, predicts which URLs contain relevant information for generating tags. A content extractor then extracts content from these predicted URLs using one or more modules. Finally, a large language model (LLM) processes the extracted content and generates tags using custom prompts designed for each tag category. These prompts are tailored to the nature of the extracted content, enhancing the context provided to the LLM. This multi-stage approach addresses challenges in processing large-scale, unstructured business data from diverse web sources, potentially offering improved efficiency, scalability, and accuracy in automated business intelligence gathering.

Described herein are examples of a system comprising a processor and memory storing instructions to execute a predictive multi-modal retrieval subsystem configured to analyze digital content, extract metadata, predict information needs, and retrieve relevant content; a multi-document viewing subsystem configured to index documents, establish relationships, and present aggregated content in a unified interface; a data management subsystem configured to organize files, generate folder structures, and provide interactive navigation; a file summary LLM configured to process documents and generate summaries with metadata; and a user context blob configured to maintain user context data across sessions. The system includes a method for receiving digital content, processing through OCR to extract text, analyzing with the file summary LLM to generate metadata and summaries, predicting information needs by simulating task progression and identifying related documents, and presenting retrieved content through a unified interface.