73 results about "Document level" patented technology

The invention provides an end-side adaptive document structure understanding method, which comprises the following steps of: uniformly rendering and normalizing a to-be-analyzed document, and outputting a page-level pixel grid and basic metadata; executing lightweight layout analysis and region classification to obtain a bounding box, a reading sequence and a region type label of each region in the page; each document area is routed to a corresponding special analysis channel for parallel analysis, and each analysis channel outputs a structured intermediate result and confidence; performing consistency verification and completion reasoning on intermediate results output by each channel, and generating a traceable verification evidence chain for low-confidence fragments; all channel results after verification are fused into a unified document-level structured output; and for a new document type or a continuous low-confidence mode, starting an adaptive process of a parameter efficient fine tuning technology to generate a channel-level increment weight packet, and updating model parameters of an analysis channel. The method has the beneficial effect that parallel accurate analysis of different elements such as tables, formulas, texts and the like can be realized.

The invention discloses an event causal identification method and device based on an iterative graph neural network, equipment and a medium, and relates to the technical field of artificial intelligence and machine learning, and the method comprises the steps: carrying out the sentence coding and event extraction of an input text, and obtaining a sentence embedding and event mention result; then, sentence-level embedding and document-level embedding of the event are generated using a multi-granularity context awareness mechanism. Then, constructing an initial event causal graph structure, and encoding the initial event causal graph structure to obtain graph embedding; and finally, dynamically updating an event causal graph structure by combining sentence-level embedding, document-level embedding and graph embedding through an iterative graph optimization mechanism, and realizing accurate identification of the event causal relationship. Through a multi-granularity context perception mechanism and an iterative graph optimization mechanism, local and global context information is effectively integrated, the accuracy and robustness of document-level event causal relationship recognition are improved, and the method is particularly excellent in performance when processing long texts and cross-sentence causal relationships and can better adapt to complex document structures.

The invention provides an entity pair guided scientific and technical literature document level relation extraction method and system, and the method comprises the steps: carrying out the entity recognition of an input scientific and technical literature document, and obtaining an entity set in the scientific and technical literature document; based on an entity pair pre-screening mechanism of multiple sampling and similarity verification, screening out a candidate entity pair set from all possible entity pairs of the entity set; then generating enhanced relation description fusing corresponding entity type information and relation semantics between the entity pairs; based on a pre-constructed relation semantic knowledge base, a double-layer filtering mechanism is adopted, and a corresponding fine screening candidate relation set is retrieved for each enhanced relation description; and guiding the large language model to perform triple fact judgment by using detailed semantic description of the candidate relationship to obtain an output result. According to the method, high-precision relation extraction is ensured, and meanwhile, the calculation overhead of long text processing is remarkably reduced, so that scientific and technical literature document-level relation extraction is more accurate and efficient.

A graphical, hierarchical document stream browser and environment for semantic (e.g. framing) and performance data analysis and interactive visualization integrates three scales: entities (competitive), entity (diachronic), and document (linguistic). The document level includes annotation and computational linguistics facilities; the entity level has calendrical and time-series focus. All levels emphasize deep linkage and network (i.e. connective / relational space) view of objects, with user-configurable connectivity. Large language model (LLM) integrations provide synthetic advisories, public opinions, reports, plot insights, comparisons; traditional natural language processing techniques and neural models are also employed. A smart plot system includes a “plot cart” and interpreter with an analysis snippet library. Graph structure may arise via adjustable blending or perceptual optimization of canned attribute-related distance functions or via link-induction query language with deep “semantic stored procedure” subexpressions, or feed into graph neural network-style inference for predictions. Most non-LLM ongoing computational load is client-side, using precomputed hierarchical summary files.

The invention discloses a document-level event argument extraction method based on an enhanced AMR graph. The method comprises the following steps: 1, splicing a document text, an event type and a role type into an input text according to a predetermined label; 2, the text feature tensor is converted into a text feature tensor through an encoder; 3, selecting candidate arguments, and inputting the candidate arguments into a scoring device to calculate scores; 4, selecting the first N candidate arguments with the highest score to enhance the AMR graph; 5, extracting word node features in the graph by using the graph convolutional network; and 6, obtaining an event argument and a role set thereof through a role classifier in combination with the text features and the word node features. According to the method, the event type and role information key information required by an event argument extraction task lacked in the AMR graph are enhanced; besides, the sensitivity of the graph structure to noise interference contained in candidate arguments is reduced by limiting the information flow direction in the graph, so that the enhanced AMR graph can keep the integrity of the semantic relationship between entities in the structure dynamic adjustment process, and the accuracy of current event argument extraction can be improved.

The invention relates to the technical field of artificial intelligence, and discloses a credible RAG-based hybrid expert routing method and system, and the method comprises the steps: obtaining metadata of a document associated with an input text to generate a document-level credible RAG vector; obtaining a hidden representation of each token according to the input text and the text content of the associated document, obtaining a credible vector corresponding to each token based on the document-level credible RAG vector, and splicing the hidden representation of each token and the corresponding credible vector to generate an enhanced input vector of each token; taking the enhanced input vector as an input of a gating network, calculating an expert score and selecting at least one expert network for weighted fusion to generate a final output; through the method, the whole routing decision process of organically fusing the credible hierarchical information of the external knowledge into the hybrid expert model is displayed, the spanning from semantic driving to semantic-credible cooperative driving is realized, and the routing accuracy is remarkably improved.

The invention relates to the field of natural language processing (NLP), in particular to a document level relation extraction technology. A traditional relation extraction method has the problems of insufficient local semantics, insufficient global semantic modeling, difficulty in reasoning complex relations and the like when processing long documents, cross-sentence relations and long-distance dependence. In order to solve the technical problem, the invention provides a document-level relation extraction method based on multi-level feature collaborative modeling. The method comprises the following steps: acquiring context semantic representation of a document by utilizing a pre-training language model, and constructing entity representation through dynamic aggregation of multiple mentions of an entity; neighborhood fine-grained interaction features between entity pairs are captured in combination with a local interaction convolution module, and key information contexts related to entity relation inference are extracted through a global attention mechanism. Furthermore, a multi-level stacked feature fusion structure is designed, and progressive collaborative modeling of local semantics and global semantics is realized, so that the expression ability of the model to a cross-sentence relationship, long-distance reasoning and a complex relationship is enhanced. In addition, the invention provides a hybrid adaptive loss function to improve the robustness of the model to difficult-to-classify samples and low-frequency relationships; and a teacher-student type knowledge distillation mechanism is introduced, and the student model learning is guided by using pseudo labels and evidence distribution, so that the overall relationship inference performance is improved. The method has high relation modeling ability, reasoning ability and generalization ability, and can be widely applied to tasks such as knowledge graph construction and information extraction.

The invention discloses a document-level relation extraction method and system based on knowledge enhancement and evidence guidance, and belongs to the technical field of natural language processing and information extraction. According to the invention, three core modules are mainly used for cooperative work: a document graph enhancement module is used for constructing a hierarchical heterogeneous graph and fusing co-reference analysis to enhance semantic representation; the knowledge enhancement module introduces an external knowledge base and adopts a confidence coefficient filtering mechanism to reduce knowledge noise; the evidence guidance reasoning module realizes multi-hop reasoning through axial attention and evidence supervision, and solves the technical problems of decentralized modeling of reasoning capability, large knowledge integration noise, insufficient evidence guidance and limited long-range dependence capture capability in the existing method. Experiments show that the method can effectively capture inter-sentence dependence, suppress knowledge noise and improve multi-hop reasoning stability, and can be widely applied to scenes such as knowledge graph construction, intelligent question and answer and information retrieval.

The invention provides a document-level relation extraction method for fusing subgraphs and displaying a constructed reasoning path, and belongs to the technical field of natural language processing. The method comprises the steps that an input text sequence is converted into a word vector sequence through an encoder; constructing a heterogeneous graph comprising entity nodes, mention nodes and sentence nodes through a document graph construction layer; explicitly defining three reasoning paths in sentences, between sentences and integration; extracting a sub-graph from the periphery of the target entity based on the document graph and the reasoning path, and reasoning by applying an R-GCN network; global encoder feature information and local subgraph feature information are sent to a fusion relation classification layer for relation distribution probability prediction; and optimizing a weighted adaptive loss function, and dynamically adjusting loss contribution degrees of different types of samples. According to the method, the reasoning ability for complex relations, the long-distance dependence capture effect and the accuracy of long-tail relation recognition are improved, and meanwhile, the accuracy, efficiency and generalization ability of document-level relation extraction are remarkably improved.

The invention discloses a long text abstract generation method based on hierarchical graph comparison themes, which comprises the following steps of: 1, preprocessing an original document, dividing sentence sequences, and obtaining global context-aware sentence and document representation through a hierarchical encoder network; 2, deducing document-level and sentence-level topic distribution by using a neural topic model; and 3, constructing a supervision graph based on a standard abstract to perform graph comparison learning so as to close the topic representation of a document and a key sentence and push redundant information. According to the method, the deep semantic structure of the long document can be effectively captured, so that the theme consistency and the information coverage degree of the abstract can be improved, and the redundancy is reduced.

The application discloses an event causal relationship identification method and device based on an iterative graph neural network, equipment and a medium, relates to the technical field of artificial intelligence and machine learning, and comprises: performing sentence encoding and event extraction on input text to obtain sentence embedding and event mention results. Then, a multi-granularity context perception mechanism is used to generate sentence-level embedding and document-level embedding of the events. Then, an initial event causal graph structure is constructed and encoded to obtain graph embedding. Finally, through an iterative graph optimization mechanism, the sentence-level embedding, the document-level embedding and the graph embedding are combined to dynamically update the event causal graph structure, so that the accurate identification of the event causal relationship is realized. Through the multi-granularity context perception mechanism and the iterative graph optimization mechanism, the local and global context information is effectively integrated, the accuracy and robustness of the document-level event causal relationship identification are improved, especially when long texts and cross-sentence causal relationships are processed, the application can better adapt to complex document structures.

A method includes discarding a current knowledge corpus; selecting a new knowledge corpus; performing operations as follows using an Artificial Intelligence (AI) retriever engine: dividing the new knowledge corpus into a plurality of sub-documents; encoding a query for the plurality of sub-documents using a query encoding model; encoding each of the plurality of sub-documents using a document encoding model; and determining at least one matching sub-document of the plurality of sub-documents that is a match for containing an answer to the query based on the encoded query and each of the plurality of encoded sub-documents; performing operations as follows using an AI reader engine: generating an inference about the answer to the query based on a concatenation of each of the at least one matching sub-document with the query, each of the at least one matching sub-document having an associated reader loss function result for the inference; identifying one of the at least one matching sub-document having a lowest reader loss function result; and associating the identified one of the at least one matching sub-document with a truth label for the query.

The invention discloses a printing field knowledge graph automatic construction method based on a large language model, which comprises the following steps: collecting a printing field multi-source heterogeneous document, processing the document to form a structured text block set, and taking the structured text block set as the input of the large language model; the model generates a primary abstract for each document, then performs cross-document semantic association analysis, constructs a document-level access sequence and a potential association network, and generates a secondary abstract; based on a printing field ontology mode, constructing a knowledge extraction mechanism fused with a structured prompt template, and guiding the model to perform structured knowledge extraction to form a candidate triple; and performing entity alignment, conflict resolution and multi-source knowledge fusion on the triple to form a structured knowledge graph. According to the method, end-to-end and automatic construction from the unstructured document to the structured knowledge graph is realized, the integrity, reasonability and construction efficiency of knowledge organization are remarkably improved, and core technical support is provided for printing intelligent diagnosis, process optimization and knowledge reuse.

The invention provides an entity linking method and system based on semantic enhancement of a large language model, and the method comprises the steps: collecting related data according to an entity linking research subject, constructing a tool retrieval adaptation strategy for carrying out the fine granularity of a document-level task, and obtaining an initialized search entry set; screening a candidate set according to the initialized search entries, and adopting different optimization strategies for different structured entries to obtain a high-quality candidate entity set; and according to the target mention and the high-quality candidate entity set, constructing a task conversion prompt template, converting an entity link task into a multi-choice problem, inputting the multi-choice problem into a large language model, and outputting a link result in accordance with a set output format. The method has the advantages that the screening efficiency of the candidate set is remarkably improved, and the entity linking accuracy is effectively improved.

The invention discloses a document-level event argument extraction method and system based on a semantic fusion graph, and the method comprises the steps: building a semantic mention graph by taking entity mentions in a document as nodes of the graph and semantic relationships as edges, fusing a graph structure into a large language model for embedding representation through a multi-layer encoder, and dynamically updating node and edge information through an attention mechanism. The embedded representation of the enhanced context semantics is obtained; constructing an event graph based on all event structures in the document, modeling a multi-event association relationship, and fusing the multi-event association relationship into a large language model decoder to strengthen the understanding of the model on the association between events; the method comprises the following steps: constructing cue sentences according to event types, synchronously inputting multi-event cue sentences in a training stage, designing a weighted loss function, guiding a model to learn multi-event information interaction, inputting the multi-event cue sentences in a reasoning stage, and only taking a target event argument result as an evaluation standard. According to the method, document noise can be suppressed, the problems of scattered distribution and long-distance dependence of events and arguments are solved, and the accuracy and robustness of argument extraction are improved.

The application relates to a document retrieval method and system based on power semantic enhancement and electronic equipment, and belongs to the technical field of natural language processing. The application solves the problem of low utilization of complex knowledge and insufficient power professional semantic understanding, which leads to low retrieval accuracy. The application comprises the following steps: receiving user query content, obtaining a query embedding vector by using a modal joint embedding model; obtaining a structured query vector of the user query content by using a large language model and a text embedding model based on a power knowledge graph; obtaining a plurality of candidate documents, a document-level similarity score and a page-level similarity score of the candidate documents according to the query embedding vector and the structured query vector, and then obtaining a comprehensive similarity score of each candidate document by using a double-path prediction model; obtaining a total score according to the document-level similarity score, the page-level similarity score and the comprehensive similarity score of each candidate document, and selecting a plurality of candidate documents with the highest total score as retrieval results. The application improves the retrieval precision.

The invention discloses a document anomaly detection method and system based on a Transform-LSTM (Long Short Term Memory) hybrid model, and relates to the technical field of computer security and natural language processing. The method comprises the following steps: performing format analysis and text cleaning on an original document to generate a standardized text sequence; performing dependency syntactic analysis based on the sequence to generate a syntactic path set, and matching the syntactic path set with a predefined global syntactic rule set to generate a rule matching vector; splicing the vector, a word embedding vector and a dependency path embedding vector obtained by syntactic path coding into a composite semantic vector; a Transform encoder is input to capture global semantic dependency, and document-level semantic representation is obtained through pooling; and inputting an LSTM model according to a semantic unit sequence to carry out time sequence modeling, and completing document detection according to an output abnormal probability. According to the method, deep fusion of the grammar structure and the semantic context is realized, malicious documents with disordered structures and abnormal semantics can be effectively identified, and the detection precision and the adaptive capacity of unknown attacks and complex documents are improved.

The invention discloses a document-level saliency detection method and a related device, and the method comprises the steps: obtaining a document-level entity, and extracting an entity text, a theme text and a context text from the document-level entity; the entity text, the subject text and the context text are input into an entity saliency detection model based on a multi-source attention mechanism to obtain the saliency type of the document-level entity, the method and the related device can accurately detect the saliency type of the document-level entity, and the method and the related device are particularly suitable for texts in the field of electricity markets, for example, the saliency type of the document-level entity can be accurately detected. The policy file and the transaction record can improve the accuracy of market decision and knowledge graph construction.

The invention discloses a document-level relation extraction method based on double-attention fusion. The method comprises the following steps: A, determining word embedding of each word; b, obtaining an initial hidden state representation of word embedding by using a BERT encoder; c, obtaining an attention score of the initial hidden state; d, converting the initial hidden state into an attention weight; e, obtaining a final hidden state representation and a mention representation of word embedding; f, obtaining an entity embedding expression through the coarse-grained and fine-grained entity embedding expressions; g, entity-level interaction representation between the entity pairs is obtained through double-attention fusion; h, calculating axial attention representations of the horizontal axis and the longitudinal axis of the two entities; i, calculating entity pair-level interactive representation of the entity pair on a transverse axis and a longitudinal axis; j, obtaining a final representation of the entity by connecting the entity-level interaction representation and the entity pair-level interaction representation; and K, calculating the probability of each relationship of the entity pairs by using a feedforward neural network, solving the relationship type with the maximum probability corresponding to each entity pair in the document as the relationship type to which the entity pair belongs, and finally completing document-level relationship extraction. According to the invention, accurate relation prediction between entity pairs can be realized.

The invention provides an index database management method and device, a retrieval method and device, a medium and a retrieval system, and relates to the technical field of computers, in particular to the technical field of information retrieval. The implementation scheme comprises the steps of receiving a first library building request; performing document-level popularity determination on the first document data; lexical item level popularity determination is carried out on the at least one first lexical item; and for each first lexical item in the at least one first lexical item, performing hierarchical storage on the first lexical item and first inverted index information of the first document data according to a lexical item level popularity determination result of the first lexical item and a document level popularity determination result of the first document data.

The application provides an element extraction method for a document-level low-resource scene based on self-labeling and a prompt, adopts a self-labeling method to construct label core words for named entity recognition, adopts a prompt method for relation extraction, and solves the problems of difficulty in constructing label core words for element extraction in a document-level low-resource scene, slow decoding speed and poor performance; and the method is suitable for the fields of justice, tourism, e-commerce, electric games, social interaction and finance.

This invention discloses a surgical editing method for complex documents based on markup language skeleton preservation. When parsing a complex document into an intermediate data model, the original markup language fragments of each content block are saved as skeleton anchors, and the global resource file is preserved as a whole document-level skeleton; unrecognizable files are stored in a pass-through repository. Upon exporting after editing, only the smallest markup node carrying the text content is replaced for the content block holding the skeleton anchor; format attribute nodes remain unchanged at the byte level; the document-level skeleton and pass-through repository are written back as is. This method solves the problem of inevitable loss of format information in existing complete conversion routes, achieving zero number of format difference nodes in tests with 10 complex documents.

The application provides a network document relationship extraction method and system, the network document relationship extraction method comprises the following steps: replacing all mentions in each entity with the first occurrence of the mention; training a pre-training model using document-level contrast pre-training and mask language model pre-training respectively; using the trained pre-training model to encode the context of a document sample to obtain a vector representation of each token, converting the named entity recognition information and part-of-speech tagging information of each token into a vector representation respectively and splicing the vector representation of each token to obtain a word embedding representation and a document embedding representation; performing representation enhancement on the entity to obtain a final embedding representation; using a binary cross-entropy loss function to calculate the possibility of the existence of a relationship between entities, and selecting the relationship with the highest possibility as the result of relationship classification. The application can effectively extract the relationship of network documents and provide effective support for network text processing.

The application provides an end-side adaptive document structure understanding method, which comprises the following steps: uniformly rendering and normalizing a document to be parsed, outputting a page-level pixel raster and basic metadata; performing lightweight layout analysis and region classification, obtaining the boundary box, reading order and region type label of each region in the page; routing each document region to a corresponding special parsing channel for parallel parsing, and each parsing channel outputs a structured intermediate result and a confidence; performing consistency checking and completion reasoning on the intermediate results output by each channel, and generating traceable proof chains for low-confidence segments; fusing the results of each channel after checking into a unified document-level structured output; for new document types or continuous low-confidence modes, starting the adaptation process of the parameter-efficient fine-tuning technology to generate a channel-level incremental weight package and update the model parameters of the parsing channel. The application has the beneficial effects that it can realize parallel and accurate parsing of different elements such as tables, formulas and texts.

The invention relates to a document-level event argument extraction method fusing role capsules and prompt gating, which comprises the following steps of: extracting data from document-level event parameters, and constructing a data set; preprocessing the data set to obtain input of an event parameter extraction model; constructing a document-level event parameter extraction model, and performing model training and verification; the document-level event parameter extraction model comprises a coding layer, a trigger word modeling layer, a gating prompt fusion layer, a role perception capsule network layer and a boundary prediction and matching layer; and inputting the test set of the data set into the model for testing, and optimizing the model through an ablation experiment. According to the method, a capsule network structure based on role categories is introduced, semantic representation capable of sensing the structural relationship between roles is generated around event trigger words, a prompt gating mechanism is designed to perform fusion modeling on structured prompts and context semantics, a loss function is optimized by using a Hungary matching algorithm and a span, and the probability of event triggering is lowered. The expression ability of the prompt information is improved, so that the matching precision is improved.

This invention provides a method, apparatus, and device for spoken-to-written language conversion based on graph attention networks. The method includes: semantically encoding a spoken document to obtain a semantic representation of the spoken document; determining the initial representation of each node in the document structure graph of the spoken document based on the semantic representation, wherein the document structure graph includes document nodes, sentence nodes, and word segmentation nodes; performing message propagation on the initial representation of each node in the document structure graph based on an attention mechanism to obtain a structure graph representation of the document structure graph; and performing semantic decoding based on the structure graph representation to obtain the written document corresponding to the spoken document. The method, apparatus, and device provided by this invention, by constructing a document graph structure diagram, can obtain a more concise and readable written document, avoiding the omission of spoken terms crossing sentence boundaries during text conversion, and ensuring the effectiveness of document-level spoken text conversion to written text.

The application relates to the technical field of event extraction, and discloses a hybrid prompt generation method and device for document-level event argument extraction. The method realizes accurate constraint through fine-grained concept guidance prompts, realizes dynamic enhancement through soft instance perception memory, realizes synergistic fusion through a hybrid generator, and constructs an efficient, accurate and robust hybrid prompt generation model. Based on the hybrid prompt generation model, the coverage of document-level event argument extraction and the effectiveness of retrieval enhancement can be improved, and more comprehensive event argument extraction can be realized.

The invention is suitable for the technical field of document processing, and provides a document-level event element extraction method and system based on a pre-training language model.The method comprises the following steps that a document is loaded and preprocessed, and the document is segmented into a plurality of text segments through a sliding window method; a word segmentation device corresponding to the pre-training language model is used for marking the text segments, and the text is converted into a token sequence; inputting the token sequences into a pre-training language model, encoding each fragment, fusing the document-level global context, and determining the final representation of each token sequence; event element boundaries are recognized from the final representation of the token sequence, and element types are determined; and taking all the obtained elements as nodes, constructing an element graph, modeling the element graph through a graph neural network, and determining an event structure relationship. According to the method, the global semantics of the long document can be better modeled, the context information of the document can be deeply understood, and the event element extraction quality is improved.

The invention relates to the technical field of event extraction, and discloses a mixed prompt generation method and device for document-level event argument extraction, and the method achieves the precise constraint through fine-grained concept guide prompt, achieves the dynamic enhancement through soft instance perception memory, achieves the collaborative fusion through a mixed generator, and achieves the extraction of document-level event arguments. An efficient, accurate and robust mixed prompt generation model is constructed, the coverage range of document-level event argument extraction and the effectiveness of retrieval enhancement can be improved based on the mixed prompt generation model, and more comprehensive event argument extraction can be achieved.

The application provides a kind of entity pair guided scientific and technical literature document level relationship extraction method and system, its method includes: the entity identification of input scientific and technical literature document is obtained in scientific and technical literature document entity set;Based on the entity pair pre-screening mechanism of multiple sampling and similarity verification, from all possible entity pairs of entity set, candidate entity pair set is screened out;Then generate enhanced relationship description that integrates corresponding entity type information and relationship semantics between entity pairs;Based on the pre-constructed relationship semantic knowledge base, a double-layer filtering mechanism is used to retrieve the corresponding fine-screening candidate relationship set for each enhanced relationship description;Then use the detailed semantic description of candidate relationship to guide the large language model to make triple fact judgment, and get the output result.The application significantly reduces the computing overhead of long text processing while ensuring high-precision relationship extraction, making scientific and technical literature document level relationship extraction more accurate and efficient.