105 results about "Entity linking" patented technology

The invention belongs to the technical field of natural language processing and geographic information systems, and particularly relates to a region address standardization method based on knowledge graph enhanced retrieval. Cleaning and preprocessing the original address text input by the user; utilizing a fine-tuned large language model to identify a geographic entity and performing standardized expansion on variant expression to generate a query candidate set; entity linking and context retrieval are carried out based on the knowledge graph, and attributes, hierarchy and spatial topology information of associated entities are obtained; in combination with the original input and the map context, an enhanced retrieval query text is generated through large language model reconstruction; vectorized semantic retrieval is carried out through the fine-tuned embedding model, and a preliminary candidate address set is obtained; carrying out multi-dimensional refined sorting by adopting a resorting model; and generating a structured standard address by using a large language model, and outputting the structured standard address after multi-level verification. According to the method, the problems of ambiguity resolution, alias recognition and context understanding in address processing are solved, and the accuracy and robustness of address standardization are improved.

The invention discloses a lightweight knowledge graph rapid construction method and system based on an NLP technology. The method comprises the steps of preprocessing an unstructured text and segmenting the unstructured text into semantic segments; unsupervised clustering is combined with the contour coefficient to determine the optimal clustering number, and a semantic association fragment cluster is obtained; performing word segmentation, part-of-speech tagging, NER and entity linking on the fragment cluster, extracting an entity and initial relationship, and fusing semantic similarity, TF-IDF word frequency collaboration degree and co-occurrence frequency to calculate a relationship edge weight; constructing a lightweight knowledge graph; in the question and answer stage, questions are disassembled through a process engine, related sub-graphs are retrieved, and answers with reasoning links are generated. The system correspondingly comprises a text preprocessing module, a clustering module, an entity relation processing module, a graph construction module, a question and answer reasoning module and a storage module. According to the method, the construction cost is reduced, the interpretability and the module coupling degree are improved, multiple scenes such as government and enterprise public opinions and medical assistance are adapted, and the problems of weak generalization, poor real-time performance and'black box 'in the traditional technology are solved.

The invention discloses an equipment fault diagnosis method and device based on a dynamic knowledge graph and a large model fine tuning technology. The method comprises the following steps: firstly, identifying a core entity from multi-source heterogeneous equipment fault data through a named entity identification model for fine tuning of domain data and a relation extraction model for special fine tuning of a fault diagnosis domain corpus, mining deep semantic association, and injecting the deep semantic association into a graph database after cleaning to form an initial knowledge graph; receiving user natural language fault description, realizing term and standard entity linking through editing distance fuzzy matching and Sension-BERT semantic vector similarity calculation, and combining bidirectional retrieval and attention mechanism fusion to obtain an enhanced context; and finally, generating a structured diagnosis report containing thinking chain reasoning based on an enhanced context by utilizing a specialized fine-tuning fault diagnosis large language model. According to the method, the limitation of a traditional diagnosis method is effectively solved, high-precision and interpretable equipment fault diagnosis is realized, and the diagnosis efficiency and reliability are improved.

Techniques for handling natural language query processing are described. In some examples, semantic meanings of words are determined during the natural language query processing. These semantic meanings are generated from metadata associated with the query and are to be used by an entity linker to help the linker link candidates to columns.

Systems and methods for entity linking using a graph neural network are disclosed. In one aspect, a method for entity linking can include extracting a first attribute set of an unknown entity from an information source and retrieving second attribute sets of known entities from a database, wherein each of the second attribute sets corresponds to one of the known entities. The method can further include generating an unknown entity graph based on the first attribute set, generating known entity graphs based on the second attribute sets, generating an unknown entity graph embedding by applying the unknown entity graph to a graph neural network, and generating known entity graph embeddings by applying the known entity graphs to the graph neural network. The method can further include assigning the information source to one of the known entities based on the unknown entity graph embedding and the known entity graph embeddings.

The invention discloses an event causal reasoning method for intelligence analysis and electronic equipment. The method comprises the following steps: acquiring a plurality of evidence fragments related to a query text from a plurality of sources, and taking the plurality of evidence fragments as an evidence fragment set; extracting entities from the evidence fragments, and linking the entities into a knowledge graph; event elements are extracted from the evidence fragments, event nodes and causal edges are generated according to the event elements, and one or more candidate causal edges are generated based on the event pairs; retrieving from the evidence fragment set to obtain one or more binding evidences of each candidate causal edge; and calling a large language model for each candidate causal edge to generate a causal assertion, performing consistency verification on the binding evidence of the causal assertion and the same candidate causal edge, combining into a traceable causal evidence chain, and performing event causal reasoning analysis. According to the method, the event-level representation is constructed, and the event-event candidate causal relationship is generated, so that a reasoning object is improved from entity co-occurrence to a reasonable and maintainable event causal structure.

The embodiment of the invention relates to the technical field of knowledge maps, and particularly discloses a knowledge map construction method and system for enterprise dynamic risks. According to the embodiment of the invention, risk original information is obtained; performing entity, attribute, relation and time information extraction on the risk original information; performing unified processing of entity linking, attribute merging and relation conflict resolution on the multiple pieces of structured information; constructing a basic knowledge graph; and based on the basic knowledge graph, constructing a dynamic risk model, performing future relationship prediction and risk trend prediction, and generating and displaying dynamic prediction information. According to the method, time span credibility evaluation and timeliness conflict interval positioning are carried out on multi-source attributes, the problems of attribute description conflicts and timeliness dislocation in multi-source data are effectively solved, meanwhile, permeation expression of risk signals on the attribute level is enhanced, and a consistent data basis is provided for accurate enterprise risk portraits.

The invention provides an electric power marketing service event risk hierarchical identification method and system based on knowledge graph semantic modeling, belongs to the technical field of electric power marketing service event risk identification, and solves the problems of data islanding, semantic understanding deviation, poor scene adaptation and high manual dependence in the prior art. According to the technical scheme, the method comprises the steps that multi-source data including 95598 appeals, electric vehicle charging / live broadcast electricity utilization and the like are collected, a knowledge graph with entity links is constructed, and risk identification is achieved through hierarchical semantic modeling, three-level risk identification and dynamic optimization. According to the method, the data association rate and the risk identification precision can be improved, the enterprise operation cost is reduced, and the digital transformation of enterprise power marketing is promoted.

The application discloses a BERT model optimization method combined with knowledge graph entity correlation. The method firstly obtains an entity set through entity linking according to a knowledge graph and a text data set, and calculates the correlation between entities according to the shortest path length between the entities on the knowledge graph in the entity set, then inputs the text data set into the BERT model as a training sample, trains the BERT model by using the loss function in the BERT model, and optimizes the attention distribution of the corresponding entity part in the text data under the multi-head attention of the BERT model according to the correlation between the entities during the training process, obtains the optimized BERT model, and applies the BERT model to processing a downstream task. The application guides the training of the BERT model through the knowledge graph entity correlation, and effectively improves the effect of the BERT model in processing the downstream task.

The application relates to a kind of port entry and exit special goods risk identification methods based on knowledge graph construction and entity linking, comprising: according to the risk identification demand of port entry and exit special goods in advance, constructing hierarchical risk factor knowledge graph;Obtain the declaration material of risk to be identified, carry out entity extraction from the declaration material, retrieve historical similar cases according to the extracted entity, build enhanced context containing declaration material and its historical similar cases, and carry out overall vectorization processing;Adopt three-granularity matching strategy: first, through Top-K screening, to match special goods subclass with coarse granularity;Then through threshold filtering, to match risk factor with medium granularity, finally, fine-grained matching risk feature node, to obtain the risk identification result of port entry and exit special goods.Compared with the prior art, the present application has the advantages of more precise risk feature expression, explicit risk knowledge audit, and realizes accurate matching from declaration material to knowledge graph node.

The application provides an entity linking method and system based on dynamic time steps, and relates to the technical field of natural language processing. The method comprises the following steps: obtaining a plurality of entity mentions of a related topic in a document to be analyzed; obtaining candidate entities of each entity mention from a knowledge base; constructing a mapping entity association graph by using a dynamic time step method, and extracting global topic consistency features from the mapping entity association graph; obtaining target entities of each entity mention from the candidate entities according to the global topic consistency features, and associating each entity mention with the corresponding target entity. In this way, by using the dynamic time step method to construct the mapping entity association graph, in each time step, the candidate entities related to the topic are selected as the mapping entities of the mapping entity association graph, so that the correct global topic consistency features can be extracted from the mapping entity association graph, thereby improving the accuracy of entity linking.

This invention relates to the field of natural language processing technology, and more particularly to a method, apparatus, terminal, and storage medium for question answering related to power finance knowledge. The method first performs word segmentation and entity linking on the question text, identifying multiple candidate word sets corresponding to multiple keywords in the power finance knowledge graph. The power finance knowledge graph includes multiple entities and multiple relationships. Then, relationship path matching is performed based on the multiple candidate word sets to obtain multiple candidate relationship paths. Finally, the final relationship path is selected from the multiple candidate relationship paths, and the final answer is determined based on the final relationship path. This invention utilizes a link reasoning model based on the intersection of candidate word sets for power finance knowledge graph question answering tasks, which can reduce the scope of relationship matching, improve the accuracy of relationship path matching, and thus enhance the accuracy of the question answering task.

The application provides an entity linking method and system, an electronic device and a readable storage medium, wherein the method comprises the following steps: determining a representation of a denotation and a context entity; determining an entity vector of a candidate entity, wherein the entity vector of the candidate entity is obtained by using an average method of word vectors to initially represent entity words of the candidate entity; inputting the entity type feature vector, the entity distance feature vector, the entity relationship feature vector and the entity vector of the candidate entity into an entity disambiguation model to obtain a similarity between the denotation and the candidate entity. The application models entities with complex interaction relationships, maps the entities into low-dimensional vectors, carries more semantic information, improves the accuracy of the model, can fully utilize the context of the text where the entity is located and the background information of the text, disambiguates the entities in the unstructured text, and serves subsequent graph expansion and graph merging and the like.

A multimodal entity linking method based on dual encoders and a hybrid expert mechanism is proposed. This invention relates to multimodal entity linking technology at the intersection of natural language processing and computer vision. Addressing the problems of low inference efficiency, insufficient cross-modal interaction, and shallow modal fusion in existing methods, this invention proposes a multimodal entity linking method based on dual encoders and a hybrid expert mechanism. A dual-tower architecture is used to independently encode mentions and entities. Entity embeddings can be pre-computed offline and indexed, and linking is completed during inference through fast vector retrieval. A hybrid expert mechanism is introduced to achieve adaptive feature transformation of samples, and a gating network dynamically selects expert combinations. Bidirectional cross-modal attention is used to establish fine-grained alignment at the word-image block granularity. A channel attention mechanism dynamically balances the contributions of textual and visual modalities. The model is jointly optimized by multiple constraints, including load balancing loss. This invention achieves efficient retrieval while maintaining deep inference capabilities, simplifies inference time complexity, and is suitable for scenarios such as knowledge graph construction and intelligent question answering systems.

The invention provides a network content-oriented hierarchical semantic processing label automatic generation method and system, and relates to the technical field of semantic processing, and the method comprises the following steps: receiving a semantic analysis request of a target user, analyzing to obtain a first work and a second work, respectively carrying out core entity extraction and entity derivative analysis, and establishing a retrieval vector, the method comprises the following steps: collecting a first network content set and a second network content set from a network data source, respectively extracting key semantic concepts and associated emotional polarities, performing entity linking with a pre-constructed domain knowledge graph, constructing a first hierarchical semantic tag set and a second hierarchical semantic tag set, and comparing the first hierarchical semantic tag set and the second hierarchical semantic tag set; and performing label alignment classification and semantic aggregation about the knowledge field and the emotional polarity to generate a hierarchical comparison report. The technical problems that in the prior art, network content hierarchical semantic tags are low in generation efficiency and limited in analysis precision are solved, and the technical effects of efficiently and accurately generating precise semantic tags and improving semantic processing efficiency and analysis precision are achieved.

This invention relates to the field of data analysis technology, and in particular to a query analysis method, apparatus, device, and storage medium based on a large language model. The method provides comprehensive and effective data support for enterprise decision-making. It achieves semantic fusion of multi-source heterogeneous data through a unified knowledge graph, breaking the limitations of traditional data silos and providing a comprehensive and accurate business knowledge foundation for the large language model, reducing model illusions from the source. It achieves differentiated analysis processing through task type classification, automatically scheduling different capability modules of the large language model according to user needs, balancing the efficiency of basic queries with the accuracy of in-depth analysis. It obtains the full business context through entity links, limiting the reasoning process of the large language model to the enterprise's real business scenarios, significantly improving the business relevance of the analysis results. It achieves in-depth diagnostic analysis through causal attribution, fully leveraging the logical reasoning capabilities of the large language model and overcoming the limitations of descriptive analysis.

This application discloses a document analysis method, apparatus, device, and medium based on a retrieval-enhanced generation framework, applied in the field of document analysis technology. The method includes: extracting target information from parsed information obtained from parsing several unstructured documents to construct an initial knowledge graph; the target information includes table information, text information, relationships between table information and text information, and relationships between table information; the nodes in the initial knowledge graph include table-related nodes determined based on the table information; performing entity linking and node relationship reasoning on the initial knowledge graph to obtain a target knowledge graph; the entity linking operation includes identifying target entities pointing to the same real-world object in different documents to establish cross-document links; retrieving the target subgraph of the query request from the target knowledge graph, using a large language model and generating query results based on the target subgraph to achieve the analysis of unstructured documents. This method can improve the document analysis results and enhance accuracy.

The invention discloses a smart city knowledge graph construction method and device and a medium, and relates to the technical field of knowledge graphs. The method comprises the following steps: acquiring multi-modal data in a smart city; performing feature extraction on the image by using a visual encoder to generate an image feature sequence, and performing feature extraction on the text by using a language encoder to generate a text feature sequence; inputting to a feature fusion layer, and performing bidirectional interaction through a cross-modal attention mechanism to generate joint representation; obtaining a preset natural language instruction template, combining the joint representation and the natural language instruction template into a prompt, and inputting the prompt into the multi-modal large model after instruction fine tuning to generate structured entity relationship knowledge; and analyzing the structured entity relationship knowledge, converting the structured entity relationship knowledge into a knowledge triple, and performing entity linking, conflict detection and resolution with the knowledge graph to obtain a new knowledge graph. According to the method, accurate, efficient and continuous dynamic evolution of the knowledge graph is realized.

This application provides a method for multi-source intelligence semantic analysis and compliance risk early warning, applied to intelligent processing devices. The method includes: transforming multi-source heterogeneous intelligence data into a unified semantic representation vector through a cross-modal semantic mapping model; performing entity linking and relationship reasoning on the unified semantic representation vector based on an enterprise multi-source intelligence knowledge graph library to generate an intelligence semantic analysis graph; matching the intelligence semantic analysis graph with a multi-level compliance rule library through an enterprise compliance rule reasoning engine to mine and assess the severity and associated transmission paths of potential compliance risk points to generate a risk transmission path diagram; and generating tiered compliance risk early warning information based on the risk transmission path diagram. This application improves the effectiveness of multi-source intelligence semantic analysis and the rationality of compliance risk early warning compared to existing solutions, and better meets the actual needs of enterprises for refined management of compliance risks.

The invention provides an intelligent customer service robot question and answer method and system based on a charging knowledge graph, and the method comprises the steps: carrying out the intention recognition of a charging session inquiry, and obtaining an inquiry intention; based on the inquiry intention, performing entity linking in the charging knowledge graph to obtain an entity to which the inquiry intention is linked; based on the linked entities and inquiry intentions, executing multi-hop search in the charging knowledge graph, and obtaining nodes consistent with intention slot constraint and associated rule terms; generating an evidence chain based on the obtained nodes and the associated rule terms; based on the evidence chain and predefined decision logic, driving the large language model to generate a natural language answer; and combining the natural language answer with the rule term document information corresponding to the natural language answer to generate a charging session answer. According to the inquiry intention, the entity linked in the charging knowledge graph is combined to perform multi-hop search and generate the answer, so that cross-theme interference is avoided, and the accuracy of charging session answering is improved.

The application discloses a structured and unstructured fact knowledge fusion corpus construction method and device. Structured fact knowledge is selected from a pre-constructed structured knowledge base and is converted into a fact retrieval query. A plurality of unstructured fact knowledge text candidates are retrieved from a pre-constructed unstructured text base according to the fact retrieval query. Entities in the unstructured fact knowledge text candidates are obtained by using entity recognition and entity linking. The fact knowledge text candidates are matched with the structured fact knowledge in the knowledge base. The fact relevance of the fact knowledge text candidates is judged based on the matching result. Text candidates with strong relevance and the structured fact knowledge matched therewith are reserved as a piece of structured and unstructured fact knowledge matched corpus and are saved into a corpus. The above process is repeatedly performed, and finally a high-quality corpus containing a plurality of matched corpora is formed.

The application discloses a kind of character information intelligent query and entity linking method supporting time alignment.The method is: extracting the entity mention and its context needing to be linked from the original text input;Using large language model, according to the entity mention and its context, the instruction input large language model is constructed, and a structured query intent containing the standard name candidate list related to the entity mention, dynamic context and inferred time is output;From the structured query intent, the evidence pair with timestamp is extracted, and the candidate entity list matching the standard name in the standard name candidate list under the effective time point corresponding to the evidence pair in the knowledge base is retrieved;According to the entity mention and its context, structured query intent and candidate entity list, evidence summary input is generated to large language model, and the entity linking result between entity mention and standard name and corresponding explanation text are output.The application solves the problem of inaccurate candidate entity recall.

The disclosed concepts relate to contextualization of generative language models. In some implementations, a linked entity database is populated with entity resource identifiers of entities extracted from a search log by an entity linker. A contextualized prompt data structure is generated based on the linked entity database, e.g., by including linked entity context information in the contextualized prompt data structure. A response to the contextualized prompt data structure is received, where the response is conditioned on the linked entity context information.

The present application relates to a kind of entity knowledge index cross-view introspection ordering distillation method, equipment and medium, the method is through ordering knowledge transfer to improve the robustness of entity linking under visual absence.Based on complete multi-modal data, teacher model is trained, and pure text data is input into student model (inherit teacher architecture and parameter).Specifically, it includes: ordering alignment loss: through the differentiable ordering alignment teacher (complete modality) and the entity ordering distribution of student (absence modality), the semantic distinguishability is retained;Cross-view distillation: utilize teacher entity-mention reverse ordering to enhance multi-modal representation understanding;Self-consistency constraint: force student mention-entity ordering and reverse entity-mention ordering consistent, break through one-way search limitation.Three-loss joint optimization, and student model only needs text input is deployed.Compared with traditional probability distillation, through ordering transmission and cross-view introspection mechanism, the stability in absence scene is significantly improved, and there is no additional parameter.

The invention discloses an ETC digital human intelligent customer service question and answer optimization method and system based on a knowledge graph, and the system comprises a data obtaining module which is used for obtaining data information related to the ETC business field from a multi-source heterogeneous data source, carrying out the preprocessing of the obtained data information, and obtaining the processed text data; the knowledge graph construction module is used for converting the text data into structured entities, relationships and attributes through entity extraction, relationship extraction and attribute extraction, and dynamically constructing a knowledge graph in the ETC service field; the question and answer optimization module is used for understanding user questions based on the constructed knowledge graph, generating knowledge graph query statements according to the user questions, retrieving and reasoning related knowledge and outputting corresponding answer contents; wherein user question understanding comprises intention recognition, entity linking and emotion recognition; and the dynamic updating module is used for monitoring the constructed knowledge graph in real time and optimizing the knowledge graph through an incremental updating algorithm.

The invention provides a cultural industry digital monitoring system and method based on machine learning, and relates to the technical field of digital monitoring, and the method comprises the steps: obtaining a digital file, monitoring the access of a user to the digital file, and obtaining a user access sequence; determining a plurality of culture entities of the digital archive, and extracting a plurality of user subjects from the user access sequence; entity linking and relation extraction are carried out on all the culture entities, the digital archives and all the user subjects, and a knowledge graph is constructed; performing parallel feature extraction on the digital archive to obtain an access mode reference; judging the normal offset of the digital file when the user accesses based on the access mode reference and the session data to obtain an abnormal state; and positioning entity nodes associated with the abnormal state according to the knowledge graph, evaluating an associated risk range, and triggering an active protection response according to the associated risk range. According to the method and the device, the compliance of user access and the security of data can be monitored in combination with the behavior characteristics of the user.

The application discloses an enhanced large language model medical question and answer method based on a dual-view knowledge graph exploration, relates to the technical field of retrieval enhancement, and comprises the following steps: a large language model is used to extract medical entities from a user question, and a three-step entity linking algorithm is used to link the medical entities with nodes in a knowledge graph; subsequently, based on the mapped entity set, core entities and potential core entities are identified in the knowledge graph as exploration anchors of the knowledge graph; then, one-hop search and fine-grained pruning are respectively performed on the two types of entities to obtain two reliable triple sets; finally, the triple sets are input into the large language model together with the question to guide the generation of an answer. The method can more efficiently retrieve more comprehensive knowledge in the knowledge graph, and provides simple and high-quality external knowledge for the large language model, thereby improving the accuracy and efficiency of the large language in answering medical questions, and having higher economic benefits.

This invention discloses a rumor detection method and apparatus based on knowledge graphs and large language models, relating to the field of artificial intelligence technology. The method includes: identifying key entities from data to be detected; linking the key entities to a preset first knowledge graph; extracting local subgraphs related to the entities; and constructing a second knowledge graph based on these subgraphs. The preset first knowledge graph includes fact triples related to the key entities. The method also includes searching for fact triples corresponding to the key entities in the second knowledge graph; performing multiple rounds of reasoning loops on the data to be detected using a preset large language model based on the data to be detected to obtain reasoning chains and evidence; and generating a rumor detection result based on the reasoning chains, evidence, and corresponding fact triples. According to embodiments of this invention, by integrating the structured factual knowledge of the knowledge graph with the reasoning capabilities of the large language model, the accuracy and interpretability of rumor detection are improved.

The method comprises the following steps: taking an image and a text as input, combining a low-level high-resolution image and a high-level strong-semantic image to enhance image features by using a feature pyramid bidirectional fusion strategy, extracting phrase-level features of the text through a convolutional neural network, and extracting a multi-modal entity link of the visual and text feature enhancement; the method comprises the following steps: firstly, extracting a visual vector, fusing the visual vector with text character-level features obtained by an encoder to form text input, filtering noise embedded in the vector by utilizing a bottleneck fusion network, improving the fusion efficiency of a model, and fusing the text vector and the visual vector to form a multi-modal fusion vector; and finally, the entity mentions are linked to the candidate entity with the highest score in the knowledge base based on the matching scores. According to the method, noise of text vectors and image vectors of an existing model can be effectively reduced, key information such as text features and visual features is reserved, feature dimensions are aligned, a semantic gap between the text vectors and the image vectors is made up, subsequent multi-modal fusion is facilitated, and the overall performance of entity linking is improved.

The invention provides a carbon field-oriented coding-semantic coupling type entity extraction method, and belongs to the field of natural language processing and carbon data intelligent crossing. Aiming at the problem that long entities with highly similar structures in the carbon field are easy to be mistakenly recognized or missed in recognition, the method comprises the following steps: firstly, constructing a special dictionary and a regular naming template based on a national and industrial emission factor database, and pre-screening candidate entities for an original text; secondly, a Span-based architecture is adopted, character-level features are fused in vector representation to enhance sensitivity to tiny coding differences, and a comparative learning mechanism is introduced to enlarge the semantic distance between similar entities; furthermore, disambiguation and complementation are realized in a window in combination with a coding rule, so that the recall rate is remarkably improved; and finally, knowledge fusion is completed through entity linking, and the universality of an identification result is enhanced. The method can be widely applied to carbon emission report automatic generation, emission factor intelligent matching, carbon management platform knowledge service and other scenes, and provides key technical support for carbon data management.