60 results about "Information coverage" patented technology

The invention provides an RAG adaptive evidence refining method based on submodule optimization, and belongs to the technical field of artificial intelligence. Collecting data to construct a knowledge base, and generating a candidate document set according to the knowledge base inquired and retrieved by a user; designing an instruction and calling a DeepSeek model to evaluate the answer contribution degree of the candidate document, calculating the relevancy between the query and the candidate document and the information coverage degree score of the optimal document to the candidate document, and constructing a sub-module objective function; calculating a marginal gain threshold value, iteratively calculating optimal documents with marginal gains greater than the threshold value, and generating an optimal document set; and splitting the optimal document set according to the paragraph level, generating a candidate paragraph set, calculating the redundancy score of the optimal paragraph set, constructing a sub-module objective function added with redundancy penalty by multiplexing the method, and performing iterative calculation to generate a final optimal evidence paragraph subset. The method effectively solves the problems of over-high redundancy, over-low information coverage and the like caused by a traditional processing method, so that the final model reasoning performance is improved.

The invention discloses an equipment information management method and an equipment information management device, and relates to the technical field of server hardware management, and the method comprises the steps that VPD information is read in a manner of enumerating PCIe equipment through a BIOS, the VPD information is an inherent attribute of the equipment and is extracted by the BIOS at the beginning, comprehensive coverage of basic information of the equipment is guaranteed, and the equipment information is stored in the BIOS; the BMC reads FRU information of equipment by using an I2C bus and an SMBus, covers function information of the equipment, reads data through the BMC, makes up for the defects of VPD information in function description and state monitoring, combines the VPD information obtained by the BIOS and the FRU information obtained by the BMC to form a target information set, the target information set comprises equipment manufacturing information and function information, and the equipment manufacturing information and the function information are stored in the BMC. Omnibearing information coverage of the PCIe equipment is realized, and the technical problem that the coverage range of information acquired by the PCIe equipment in the prior art is not comprehensive is solved.

The invention discloses a data query method, system and equipment based on distributed retrieval enhancement generation, and belongs to the field of data processing. The method comprises the following steps: a first node receives a query request initiated by a user through a gateway; the first node sends a query request to at least one second node through the gateway; the first node and the second node retrieve own knowledge bases respectively to obtain candidate documents matched with the query request, and generate retrieval results based on the candidate documents through own gateways; the second node sends a retrieval result generated by the second node to the first node through the gateway; the first node reorders the retrieval result generated by the first node and the retrieval result sent by the at least one second node through the gateway, and determines a target retrieval result based on the reordering result; and the first node generates a query result corresponding to the query request based on the target retrieval result through the large model. According to the embodiment of the invention, the information coverage rate and the scene adaptability can be improved, so that the accuracy of the query result is improved.

The invention relates to a document analysis and retrieval enhancement generation method fused with a multi-modal embedding model, and belongs to the technical field of AI large model application. The method comprises the following steps of: analyzing a file and extracting contents of texts, tables and pictures; performing joint semantic embedding on the extracted texts, tables and pictures by using a multi-modal embedding model to generate uniform vector representation, and constructing a vector database supporting multi-modal data; and constructing a multi-modal retrieval chain based on a vector retriever. According to the method, total element analysis and joint modeling of complex documents are achieved, the problem that multi-modal information is difficult to process in a unified mode through a traditional vectorization technology is solved to a certain degree, and the information coverage capacity and the answer accuracy of a retrieval enhancement generation system in a real scene are remarkably improved.

The invention discloses a retrieval enhancement generation method and system fusing multi-modal and hybrid retrieval, and relates to the technical field of multi-modal fusion and hybrid retrieval, and the method comprises multi-modal information fusion framework construction, row hybrid retrieval execution, adaptive weight distribution and context construction, and structured intelligent content compression and generation. According to the retrieval enhancement generation method and system fusing multi-modal and mixed retrieval, the limitation that a traditional RAG system depends on a single text vector library is broken through by integrating multi-source information, multi-modal information sources such as contract original texts, expert suggestions, dialogue history and external similar cases are integrated through a unified framework, the comprehensiveness of information coverage is ensured, and the retrieval enhancement generation efficiency is improved. According to the method, more abundant factual basis is provided for result generation, and for double guarantee of semantics and keywords, a parallel mixed retrieval method is adopted to be combined with vector retrieval and full-text retrieval, so that the defect of a single retrieval mode is avoided, and the accuracy and correlation of retrieval results are improved.

The invention discloses a dynamic risk assessment method based on multi-source heterogeneous data and application. The method comprises the following steps: extracting risk features from the multi-source heterogeneous data; constructing a multi-dimensional risk assessment model based on the labeled risk features; and based on the multi-dimensional risk assessment model, the risk is scored. According to the method, multi-source heterogeneous data can be effectively fused, risk features are extracted through a unified data processing flow, a multi-dimensional risk assessment model is constructed, risk scoring based on the multi-dimensional risk features is realized, the problems of coarse assessment granularity and insufficient information coverage caused by dependence on a single data source in the prior art are solved, and the risk assessment efficiency is improved. And the comprehensiveness and the accuracy of a risk assessment result are remarkably improved.

The invention provides a mobile terminal skin parameter inversion method based on spectral information, and the method comprises the following steps: constructing an optical model according to the characteristics of a skin layered structure and a spectral response mechanism of three skin parameters: melanin content, hemoglobin concentration and dermis thickness, and building a skin parameter and reflection spectrum database based on the model; based on the database, establishing a skin multi-parameter inversion MLP neural network structure adaptive to the collection characteristics of a mobile phone end sensor, and screening an optimal four-waveband channel through multi-waveband combination traversal; the mobile phone spectrum sensor is subjected to wave band optimization, data acquisition and skin parameter inversion. According to the method, through efficient decoupling of a nonlinear coupling relation among multiple skin parameters by waveband optimization and an MLP model, and targeted design of four optimal visible light narrow-band channels, synchronous quantitative detection of melanin, hemoglobin and dermis thickness is realized, core physiological parameters from epidermis to dermis superficial layer are covered, information coverage is comprehensive, and specificity is high.

The invention discloses a standard text semantic analysis and key clause extraction method and system, and the method comprises the steps: carrying out the multi-level semantic analysis of a standard text, recognizing a to-be-evaluated clause and a corresponding type in the standard text, extracting a key parameter entity of the to-be-evaluated clause, and constructing a clause-level semantic association structure; based on the semantic association structure, a standard term parameter knowledge graph is constructed, and the knowledge graph supports semantic query of technical parameters; and performing weight calculation and screening on the to-be-evaluated terms based on the standard term parameter mapping knowledge domain, and generating a structured standard requirement abstract list based on the screened key terms and associated parameter information in the corresponding mapping knowledge domain. In combination with multi-level semantic analysis and parameterized knowledge graph construction, automatic and structured extraction of standard text core requirements is realized, and the technical problems of low efficiency, incomplete key information coverage and lack of semantic association caused by dependence on manpower or shallow text processing are solved.

According to the multi-source retrieval and agent collaborative decision-making medical question-answering system and method, a three-layer retrieval framework is constructed, a local medical knowledge base vector retriever, a PubMed intelligent retriever and an academic deep crawling device are integrated, dynamic fusion of multi-source heterogeneous medical information is achieved, and the medical question-answering efficiency is improved. Through the synergistic effect of the three, the information coverage range is remarkably expanded, more comprehensive and multi-dimensional evidence support is provided for various medical problems, the retrieval efficiency and accuracy of a professional database are effectively improved, and the use threshold of non-professional users is reduced. Meanwhile, the technical scheme of the invention supports dual-scene task processing, can meet diversified medical information requirements, provides reliable reference for diagnosis and treatment decisions of medical staff, also provides structured and high-quality information support for subject investigation of medical researchers, and powerfully promotes intelligence and precision of medical information processing.

The invention discloses a network-on-chip debugging method and device, computer equipment and a storage medium, and relates to the technical field of integrated circuit design verification, and the method comprises the steps: generating network-on-chip data files in a preset file format, and generating a network-on-chip database containing network debugging information according to the files; acquiring target information corresponding to the network-on-chip debugging request from a network-on-chip database; and performing on-chip network debugging according to the target information, and writing a debugging result into an on-chip network database. The problem that NoC debugging is low in debugging efficiency and depends on verification of experiences of engineers can be solved. According to the method, the target information corresponding to the debugging request is provided for the user through the network-on-chip database, the dependence on the experience of a verification engineer is eliminated, the network-on-chip debugging difficulty is reduced, and the debugging efficiency is improved. All possible communication scenes and boundary conditions in the debugging process are covered through the target information, and the debugging coverage rate is increased. And continuously improving the database according to the debugging result to form knowledge resources.

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 a kind of multi-document abstract generation methods based on key event, first by internet collection under the same theme multiple media articles, and in document set basis, according to artificial rule, standard abstract is generated to construct sample dataset;Then the dataset is preprocessed, and the input data of training model is generated;Then construct the sequence-to-sequence multi-document abstract generation model based on key event fusion;Finally, based on the model after training, the output model is constructed, and the document set to be summarized is automatically summarized using the output model.The application uses event extraction technology to extract key events containing dynamic and static information from multiple documents to mine the relationship of multiple documents, which can guide the abstract generation process at multiple levels, thereby improving the information coverage and factual consistency of the abstract result, highlighting the event information in the original text, and enhancing the logicality of the abstract.

The utility model discloses an anti-static protection device for oil loading and unloading of an oil tank truck, which belongs to the technical field of oil loading and unloading protection devices and comprises a support rod, a detection component is fixedly connected to the side surface of the support rod and comprises a touch screen, and one side of the touch screen is electrically connected with an electrostatic sensor. And the electrostatic sensor is fixedly connected to the side face of the supporting rod, the other side of the touch screen is electrically connected with a broadcasting sound box, and the broadcasting sound box is fixedly connected to the side face of the supporting rod. According to the anti-static protection device for oil loading and unloading of the oil tank truck, static electricity is sensed through the static electricity sensor and transmitted to the touch screen to be displayed, a detection numerical value is broadcasted through the broadcasting loudspeaker, and therefore under the condition that the sight of an operator cannot focus on the screen all the time, the newest static electricity detection numerical value can be obtained at the first time through an auditory channel; omnibearing and dead-corner-free electrostatic monitoring information coverage is really achieved, and powerful support is provided for taking anti-static measures in time and guaranteeing operation safety.

A single-cell level spatial omics multi-modal fusion method, comprising: extracting differential expression genes and nuclear spatial morphological features from spatial transcriptome data, single-cell sequencing data and histological images, and realizing field adaptation between different platforms by using conditional variational autoencoder. Based on a probability inference model, spatial transcriptome expression, single-cell omics and morphological features are fused to jointly infer the type and gene expression level of each cell. A spatial cell network is constructed by a graph attention mechanism to realize the spatial diffusion and recognition of cell types in the whole slice range. Combined with a multi-omics enhancement module, the unmeasured gene and protein expression are completed based on expression similarity, and the prediction consistency is improved through spatial correction. The method realizes high-resolution reconstruction of single-cell multi-omics information in three-dimensional space, improves the information coverage and spatial resolution of spatial omics data, and provides an efficient and low-cost solution for spatial biology and precision medicine research.

The application relates to a chip verification method, device, equipment, storage medium and program product. The method comprises the following steps: acquiring a simulation log file and an instruction information file of a chip; matching first instruction information and second instruction information; and performing coverage verification on the instruction information file according to a matching result. The first instruction information actually executed by the chip in a simulation verification process is recorded in the simulation log file, the second instruction information required to be executed by the chip in a starting stage is contained in the instruction information file, and the coverage verification is used for verifying a coverage degree of the instruction information file by the simulation log file. In the method, the dynamic execution track of the chip in the simulation verification process can be matched with the static instruction required in the starting stage by acquiring the simulation log file and the instruction information file, the instructions actually executed in simulation and not actually executed in simulation in the instruction information file can be quickly identified, and the chip verification efficiency and accuracy are improved.

The invention relates to the technical field of data processing, in particular to an import and export dangerous chemical intelligent analysis method based on machine learning, and the method comprises the steps: constructing a dangerous chemical knowledge graph, and carrying out the path detection of the dangerous chemical knowledge graph, and obtaining at least one effective path and at least one invalid path; for any invalid path, obtaining a value index of any invalid path according to the information coverage degree between the intermediate entities of any invalid path, the similarity characteristics between the adjacent intermediate entities of any invalid path and the reliability degree of each connection segment in any invalid path; and screening at least one high-value path from all the invalid paths by using the value index of each invalid path, and performing feature extraction and model training on all the high-value paths and all the effective paths by using a PRA path ranking algorithm to obtain a dangerous chemical analysis model. The method is used for comprehensively and accurately deducing similar substances or substitutes of dangerous chemicals.

The application discloses a data acquisition method and device, electronic equipment and storage medium. The method comprises: acquiring audio to be processed collected through a microphone array, the audio to be processed comprising multiple sentences of multi-channel audio; for each sentence of multi-channel audio, acquiring a phase difference feature between at least one group of specified channel audio as target data, the specified channel audio comprising at least two position-associated channel audios. For each sentence of multi-channel audio, the method performs phase difference calculation by incorporating the audio of at least two position-associated microphone channels, which can generate higher-dimensional phase information coverage on one hand, so that the DOA feature obtained after phase difference is more intensive, thereby helping to improve the accuracy of the DOA encoding result; on the other hand, since the position-associated microphone channel audio is used for phase difference calculation, data redundancy can be avoided, thereby helping to improve the efficiency of DOA encoding.

The invention relates to the technical field of road tunnel construction traffic control, and provides an intelligent control method of a tunnel intelligent traffic dispersion system in order to solve the problems of low traffic efficiency, large potential safety hazard, incomplete information coverage and the like of an existing inclined shaft and main tunnel intersection. Meanwhile, the driving direction of the vehicle is judged through a direction recognizer; based on the vehicle position and the driving direction, the number of vehicles waiting in line and the vehicle arrival rate are obtained through statistics, calculation is carried out according to an intelligent scheduling algorithm, and the optimal green light prolonging time is obtained; and controlling a traffic light group in the tunnel intelligent traffic dispersion system according to the optimal green light prolonging time, adjusting traffic light timing schemes in the main tunnel direction and the inclined shaft direction, and performing multi-modal information release based on the traffic light timing schemes. According to the invention, the dynamic optimization scheduling of the traffic flow of the tunnel intersection is realized, a non-blind area safety early warning system is constructed, and the traffic efficiency and safety level of the tunnel intersection are remarkably improved.

The invention discloses a secret communication method based on energy-information integrated transmission, and relates to the field of power line communication. The method mainly comprises an anti-eavesdropping circuit design, a pseudo-random information coverage-based secret communication method, and realization of secret communication based on energy-information integrated transmission of the AC micro-grid. The method provided by the invention aims at the secret communication of the alternating-current micro-grid, and the design provides a new encryption strategy for the user communication of the alternating-current micro-grid due to the fact that the research on the energy-information integrated communication of the alternating-current micro-grid at present rarely relates to the privacy protection of the user information. According to the method, additional communication lines are not needed, information sending, transmission and receiving can be achieved through an existing controller and an alternating-current power line, cost is reduced, and secret communication of alternating-current micro-grid users is achieved through a pseudo-random information coverage method.

The invention discloses a tobacco leaf classification method based on complementarity of TGA data and NIR data, and the method comprises the following steps: (1) collecting the TGA data and NIR data of a sample, carrying out the preprocessing of the data, eliminating the noise and interference of the data, extracting effective features, and constructing a data matrix; (2) quantizing complementarity of TGA data and NIR data based on an orthogonal projection principle; (3) analyzing similarity and difference of different types of samples, and completing classification of the tobacco leaf samples; and (4) constructing a discrimination model, and verifying the classification enhancement effect of the orthogonal projection features. According to the classification method provided by the invention, the limitation that the TGA and NIR single-mode technology information coverage is incomplete and the complementarity lacks a quantification means in the prior art is solved.

The invention relates to the technical field of natural language processing and information retrieval, and discloses a question and answer pair automatic generation method and device, computer equipment and a storage medium. The method comprises the following steps: firstly, extracting an initial question and answer pair from an input text through a large language model (LLM); slicing and vectorizing the original text, and constructing an original text vector database; vectorizing answers in the initial question and answer pair, matching the vectorized answers with an original text vector database, and positioning one or more original text fragments for each answer to form a traceable triple; performing quality verification on the triple by adopting a comprehensive verification mechanism based on integrity and accuracy evaluation and key information coverage calculation; and finally, de-duplication is carried out in combination with editing distance calculation and semantic similarity calculation, and high-quality question and answer pairs are output. According to the method, the problems of unstable generation quality, lack of answer basis and the like are solved, and automatic construction of the high-credibility knowledge base is realized.

The invention discloses a self-organizing earthquake early warning method based on an emergency information issuing system, and particularly relates to the technical field of earthquake early warning. Acquiring an initial feature vector of a seismic event and broadcasting the initial feature vector to edge nodes, and constructing a local propagation weight and a relay queue by each node based on received information to realize multi-hop self-organizing propagation path construction; the local receiving density and the relay delay are dynamically monitored in the propagation process, a signal attenuation area is identified, and a cooperative relay request is triggered; the central node schedules available equipment to establish a temporary relay group and automatically deploys a hotspot communication protocol, so that the information coverage range and the transmission reliability are improved; and finally, constructing a complete propagation topological graph, and calculating response timestamps of the nodes to evaluate system performance. The method has the advantages of high fault tolerance, fast deployment capability and good adaptive expansibility, and is suitable for emergency early warning system construction in complex disaster scenes.

The invention provides an importance perception KV coverage method based on Transform model reasoning, and the method comprises the steps: analyzing the importance characteristics of tokens in a Transform model based on a storage and calculation integrated architecture of a ReRAM, and proposing a thought that the KV information of more important tokens is preferentially stored and the low-importance KV is covered when the storage resources are insufficient; the KV coverage method for importance perception of the ReRAM architecture is designed and realized, the resource quantity occupied by the KV can be effectively reduced under the condition that the model reasoning precision is guaranteed, and smooth reasoning is guaranteed.

The invention relates to the field of semantic bridging, in particular to a semantic bridging-based question and answer method for fusing knowledge generation and knowledge retrieval, which comprises the following steps: constructing an enterprise manual question and answer system based on an RAG architecture, the enterprise manual question and answer system comprises an enterprise manual vector database, a knowledge retrieval model of an SBERT architecture and a large language model, and a knowledge fusion model based on a cross attention mechanism and a multi-layer perceptron architecture. In a training stage of the enterprise manual question-answering system, a comprehensive evaluation function based on matching accuracy, enterprise term semantic similarity and enterprise key information coverage is calculated by fusing knowledge vectors and sample standard knowledge, and a combined loss function is constructed; and inputting a fused knowledge vector generated by the knowledge fusion model into the large language model to generate a system answer text. According to the method, semantic-level bridging fusion of the retrieval knowledge and the generation knowledge is achieved, and the coverage degree and adaptability of the RAG question and answer system to the enterprise manual knowledge are improved.

The application relates to the technical field of education data processing, and discloses an education knowledge graph construction method and system based on artificial intelligence, which comprises a multi-scale representation analysis unit, a structure parameter configuration unit, a multi-modal graph element assembly unit and an intelligent link optimization unit. The system first extracts the multi-layer convolution response distribution of a normalized teaching image by using a convolutional neural network, calculates a semantic coverage decay index reflecting the change of information coverage range with scale, then maps the index to node bearing potential, and calculates the number quota of visual anchor points and concept entities for the image according to the node bearing potential, then positions the visual area and extracts associated text concepts according to the number quota to establish a weighted image substructure, and finally performs relationship reasoning by using a graph completion model to finalize the final graph. The application can adaptively adjust the composition scale according to the knowledge density of the image content, solves the problem of imbalance of multi-modal graph structure density, and is suitable for a smart education scene.

The application relates to the technical field of artificial intelligence, and discloses a multi-language user matching recommendation method and system based on multi-dimensional label fusion, which comprises the following steps: extracting language feature symbols and interactive behavior features in multi-dimensional label data to construct multi-dimensional label vectors of users in a digital service platform; performing semantic alignment processing on the multi-dimensional label vectors by using a cross-language fusion layer in a deep learning model to obtain aligned semantic vectors; calculating information coverage corresponding to the aligned semantic vectors by using a coverage function in the deep learning model; determining core label vectors in the multi-dimensional label vectors to generate enhanced semantic representation vectors corresponding to the users; and calculating semantic correlation degrees between the enhanced semantic representation vectors to match out recommendation content corresponding to the users from the digital service platform. The application can improve the accuracy of multi-language user matching recommendation based on multi-dimensional label fusion.

The invention relates to a software information updating method and device, a storage medium and electronic equipment, and relates to the technical field of software testing, and the software information updating method comprises the steps that reference information of target software is read from a preset software information document, and the reference information at least comprises software identification information, a reference downloading address and reference file verification information; according to the reference downloading address, initiating an automatic network request to download target software, and determining actual file verification information of the target software; and performing consistency comparison on the actual file verification information and the reference file verification information, and if a comparison result is inconsistent, covering and updating the reference file verification information by using the actual file verification information. By means of the scheme, it is ensured that the reference file verification information can reflect the latest state of the software file in time, the problems that the downloaded file is not matched with the display information and downloading fails when verification is conducted according to original information due to the fact that software file modification is not synchronized to a popularization party are solved, and the downloading experience and credibility of a user are improved.

A subway platform screen door display system and method based on CAN communication are disclosed. The platform screen door management host and the display units in the display system exchange message commands via a CAN bus. Each display unit is located at a platform screen door, and the number of display units is the same as the number of platform screen doors. The platform screen door management host sends display message commands carrying display information and update message commands carrying update information to the display units. The display units send a second message command carrying their status information to the platform screen door management host. Each message command includes an identifier and a data field. The identifier includes the device type, message type, and display unit serial number, while the data field stores the specific message content. This invention offers broad information coverage, enables centralized management of the display system, improves display and update efficiency, enhances information synchronization efficiency, and allows for diversified display content.

The invention relates to the technical field of artificial intelligence, and discloses a dialogue content adjustment method and device based on emotional state recognition, and the method comprises the steps: collecting user interaction data in a user interaction process, analyzing the emotional state information of a user according to the user interaction data, and carrying out the adjustment of the dialogue content according to text data and voice data; analyzing the information coverage degree of the interaction target in the user interaction process and the current dialogue context information, calculating the dialogue load corresponding to the user at present according to the current emotional state and the current dialogue context information, and determining the dialogue state of the user according to the dialogue load, the information coverage degree of the interaction target and the emotional development trend. Questioning control parameters for the current dialogue round are generated, and questioning questions are generated according to the questioning control parameters. It can be seen that the emotion state of the user can be sensed in real time, the dialogue strategy can be adjusted in a self-adaptive mode, the adaptability of the dialogue process and the emotion of the user is improved, the dialogue pressure and the conflict emotion of the user are reduced, and the dialogue information obtaining efficiency is improved.

The invention provides an intelligent question and answer method, system and device and a computer readable storage medium, and the method comprises the steps: obtaining a corresponding preliminary search result from at least one information source for a user search request, carrying out the knowledge gap analysis of the preliminary search result, and carrying out the multi-round iterative search if the knowledge gap analysis does not meet a preset search termination condition. In the multi-round iterative search process, planning information used for guiding the next round of search is dynamically generated based on the result of knowledge gap analysis, so that the search process has adaptive ability, missing information can be continuously supplemented, and compared with single retrieval, the information coverage and exploration depth of answers are remarkably improved. Moreover, the planning information of the iterative search is dynamically generated, so that the search target of the subsequent iteration is more focused, the resource waste on irrelevant information is avoided, and the overall search efficiency is improved. And an answer is output after the information quality of iteration termination is ensured to reach a standard through a preset termination condition, so that the output accuracy is improved.