17results about How to "Improve retrieval speed" patented technology

The invention relates to the technical field of three-dimensional data processing, and discloses a three-dimensional grid data processing and multiplexing method and system, and the method comprises the steps: S1, obtaining and analyzing grid data, obtaining three-dimensional grid original data, and carrying out the format recognition, geometric analysis and topology analysis of the three-dimensional grid original data; s2, carrying out reuse-oriented standardization processing; s3, constructing multiplexing description information; s4, grid resource storage and index management; s5, determining a subset based on the request; s6, on-demand conversion and result generation; and S7, unified scheduling control. By constructing multiplexing description information containing a unique identifier, a hierarchy mapping relationship, a block mapping relationship, version information and a service adaptation mark, unified management and efficient multiplexing of grid data under different hierarchy and service scenes are realized, and unified data semantic support is provided for hierarchical retrieval, on-demand extraction and multi-service collaboration; therefore, the reusability of the three-dimensional grid data and the system cooperation efficiency are remarkably improved.

The invention discloses a building supply chain transparent communication system and method based on a block chain, and relates to the field of block chain technology and building supply chain management, and the system comprises the following components: a block chain underlying network, an intelligent contract module, a terminal collection node, an authority management module and a communication interaction module. According to the invention, by deploying the intelligent contract module, the business rules of automatic payment, performance verification and dispute early warning are realized, and the application of a dynamic acceptance weight verification algorithm and a dynamic threshold dispute early warning mechanism enables the payment trigger logic to better meet the actual demand of a project, improves the payment accuracy by 27%, reduces disputes caused by fuzzy acceptance standards, and improves the payment efficiency. Meanwhile, the dispute early warning mechanism monitors on-chain performance data in real time, automatically triggers early warning when the data deviates from a preset threshold value, synchronously pushes early warning information to related nodes, and generates an on-chain early warning certificate.

This application provides a data storage method, apparatus, and business system across an HBase cluster. The method includes: upon receiving a data upload message, obtaining the location information, first time information, and first identification information of the data to be stored; encoding the location information, first time information, and first identification information to obtain an index number, which uniquely identifies the data to be stored; reversing the index number to obtain the row key value of the data to be stored; and storing the data to be stored in a first target partition on a target server according to the row key value. The first target partition is the storage partition corresponding to the byte of the first identification information in the row key value, and the byte order of the row key value and the index number is reversed. This method solves the problem in existing technologies where storing files sequentially in the same area causes hotspot phenomena and reduces data storage speed.

The invention provides a method and device for establishing a space grid system and a remote sensing image processing system. The method belongs to the technical field of remote sensing images, and carries out unified structured management on remote sensing image data by constructing a three-dimensional holographic management framework composed of a time grid, a space grid and a spectrum grid. The space grids adopt global subdivision grids and are endowed with unique codes, the time grids establish indexes according to years, seasons and months, and the spectrum grids are divided according to professional spectrum segments. The corresponding device comprises a data receiving module, a grid coding module, a storage module, a scheduling module and a processing module. The system supports micro-service cluster deployment, visual low-loss compression and open API integration. According to the method, pixel-level scheduling, dynamic processing and real-time service release of the remote sensing data are realized, the bottlenecks that the traditional file-level management efficiency is low, the preprocessing is tedious, and the instant service demand of massive high-resolution images is difficult to meet are overcome, and the data management, processing and service efficiency is remarkably improved.

This invention provides a database optimization method and terminal for storing game items. The method involves obtaining general attributes of item data and extended attributes for each type of item data; constructing a general table based on the general attributes and an identifier field; and constructing an extended table based on the extended attributes and the identifier field. An extensible key-value pair field is added to the extended table to store variable attributes. This invention avoids the redundancy problem of distributing general attributes of different types of item data across multiple tables by storing them in a general table, thus improving retrieval speed. Furthermore, the addition of a key-value pair field to the extended table to store variable attributes allows for direct addition of new attributes using the key-value pair field without modifying the extended table structure, further reducing the impact on the original table structure, lowering the difficulty of extending attributes, and increasing the flexibility of modifying item data.

Embodiments of the present application disclose a picture retrieval method and device, a storage medium and an electronic device, and relate to the technical field of picture processing. According to a feature point code table, sample picture feature points in each sample picture are quantitatively coded, and a same picture index is constructed according to sample quantization values of the sample picture feature points. After target picture feature points of a target picture are extracted, the target picture feature points can also be quantitatively coded according to the feature point code table to obtain target quantization values, and then a sample picture corresponding to the target quantization values is searched in the same picture index, that is, a retrieval result corresponding to the target picture is obtained. Since the calculation amount of the quantitative coding is much smaller than that of directly comparing the feature points in the sample picture and the target picture, the calculation amount when retrieving similar pictures can be reduced, and the retrieval speed when retrieving similar pictures can be effectively improved.

This invention provides an intelligent storage chip integrating an edge computing data processing system, relating to the field of electronic digital data processing. It includes a data acquisition module, an edge computing module, a data storage module, and a chip management module. The data acquisition module receives externally input information data; the edge computing module performs computational analysis on the data; the data storage module stores the processed data; and the chip management module controls and manages the chip's computational and storage functions. By adding an edge computing module, this chip utilizes computing power to process edge tasks while simultaneously managing data storage, effectively improving the retrieval efficiency and access speed of stored data.

This invention discloses a method for identifying the operating conditions of a pumping unit based on orthogonal depth hashing and graph reasoning, comprising: (1) acquiring, labeling, classifying, and training an orthogonal depth hashing network; (2) using the trained orthogonal depth hashing network to extract features from each dynamometer in the sample set, generating and saving bit-independent binary hash codes; (3) constructing an inverted index library based on the hash codes of different sample graphs, storing the storage node information of different sample graphs in the corresponding hash buckets; (4) acquiring the load and displacement data collected by the dynamometer of the pumping well to be detected, and drawing the dynamometer; (5) performing pumping condition fusion reasoning judgment based on a graph convolutional network with dual sorting. The binary hash codes generated by this invention are mutually uncorrelated in each dimension, maximizing the information entropy of the encoding. This ensures the uniform distribution of massive samples in Hamming space, significantly improving the recall rate and retrieval speed in the coarse screening stage.

The embodiment of the invention discloses a vector retrieval device and method based on an RRAM, a chip and equipment, and the vector retrieval device comprises an instruction analysis module which is used for analyzing an input query instruction to obtain a query vector and a threshold value; the plurality of digital-to-analog conversion modules are used for converting the query vector output by the instruction analysis module into an analog signal; the RRAM module is used for carrying out multiply-accumulate operation on the stored vector to be matched and the received query vector in the form of the analog signal to obtain an accumulation result; and the comparison module is used for converting a threshold value in a digital signal form into a threshold value signal in an analog signal form, comparing an accumulation result output by each storage array of the RRAM module with the threshold value signal, and taking a to-be-matched vector in the storage array corresponding to the accumulation result greater than the threshold value signal as a retrieval result. By adopting the technical scheme, the vector retrieval speed is increased, and the circuit area is reduced.

The application discloses a kind of big model knowledge retrieval methods based on knowledge graph enhancement, comprising: S1, collection multi-source data, constructs original knowledge triple set, generates knowledge graph;S2, input natural language query data, execute semantic analysis and intention identification, form query semantic vector and structured query template;S3, through improved combination formula graph convolution network CompGCN, target entity node and relation vector are combined convolution operation, generate candidate knowledge list and divide configuration confidence;S4, utilize semantic embedding mechanism and DistilBERT acceleration calculation query vector and knowledge vector similarity, obtain first retrieval result set and confidence score;S5, fusion two kinds of retrieval results, obtain knowledge result set;S6, according to the fusion knowledge result set generation answer text, and output reasoning path, entity reference and relationship link.The application improves the accuracy of complex query processing, response speed and knowledge reasoning transparency.

The invention discloses a depth data measuring head, a depth camera and a depth data acquisition method. A projection device of the depth data measuring head can project masked structured light to a shooting area, and an image sensor can shoot a masked area of the shooting area to obtain an interference light two-dimensional image frame under irradiation of the masked structured light. And obtaining a processed two-dimensional image frame based on the two-dimensional image frame obtained by conventional shooting and the interference light two-dimensional image frame, and generating depth data of the measured object in the shooting area. The interference light information is obtained by projecting the masked structured light and shooting the masked part, and the influence of the interference light is eliminated in the image finally used for depth calculation according to the interference light information, so that the anti-interference capability of the structured light depth camera is improved, and normal acquisition of depth data is ensured.

This invention discloses a method for retrieving similar protein structures based on hash learning: First, a protein structure dataset is acquired, and the pairwise similarity information is calculated. For the query sample, positive and negative samples are sampled. The protein structures are modeled as a graph, and node and edge features are extracted, the feature representations of nodes and edges are updated, quantization loss and similarity loss are defined, and the final loss function is obtained. The model is then trained. Based on the trained model, each collected protein structure is represented as a binary vector, resulting in a binary vector database. During retrieval, the trained model represents the new protein structure to be retrieved as a binary vector. Similar structures are retrieved by directly calculating the Hamming distance between binary vectors or by constructing an inverted index. Alternatively, the top-ranked protein structures returned based on binary vector retrieval can be reordered using other real-valued vectors or more complex algorithms. This invention reduces storage overhead and improves retrieval speed.

The application discloses an intelligent question answering method and device, electronic equipment and storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring question information; searching a first knowledge base according to the question information; the first knowledge base stores a plurality of question-answer pairs constructed according to static information data; if a matched target question is searched, outputting the corresponding answer; otherwise, inputting the question information into a question answering model to obtain an answer output based on a second knowledge base; the second knowledge base stores a plurality of knowledge vectors constructed according to dynamic information data. According to the application, the question-answer pairs are constructed according to the static information data, and the knowledge vectors are constructed according to the dynamic information data, so that the workload of manually entering data can be reduced, and the timeliness of the answering basis can be improved; by searching the first knowledge base first, and outputting the answer based on the second knowledge base by the question answering model in the case of a miss, the answering failure or error can be reduced, and the question answering efficiency can be improved.

The invention discloses a multi-mode RAG method and device, electronic equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: performing modal type identification on original multi-modal data, respectively performing structured feature extraction, and storing an original multi-modal data copy into a hierarchical storage database; respectively carrying out vector coding processing and semantic enhancement processing on the mechanism features, and then storing the mechanism features into a semantic vector library and a knowledge graph library; retrieving in the updated hierarchical storage database, the semantic vector library and the knowledge graph library at the same time based on the retrieval intention of the user question, outputting a candidate document set after deduplication and combination of retrieval results, reordering the candidate document set, and finally generating an answer for the user question. According to the invention, accurate answering of complex questions can be realized.

This invention discloses a method and system for intelligent generation and processing of meeting content based on a multimodal large model, comprising: collecting raw meeting data and performing standardized preprocessing; inputting the multimodal large model, extracting multimodal features and semantically aligning them; performing cross-modal hash encoding to generate binary codes and establish an index library; hash-retrieval of relevant segments to construct a directed graph of meeting content; optimizing the path using Monte Carlo tree search based on the directed graph structure; generating structured meeting content and outputting minutes, summaries, and action items. This invention achieves efficient extraction, accurate retrieval, and intelligent structured generation of meeting content by integrating a multimodal large model, cross-modal hash encoding, and Monte Carlo tree search.

The application provides a perimeter defense method and device, a terminal and a readable storage medium, and belongs to the technical field of communication, wherein the perimeter defense method comprises the following steps: training a target segmentation network based on a public data set of perimeter defense; constructing a category database according to a live picture of a target site and the target segmentation network; extracting a target feature vector of a target video frame picture by using the target segmentation network; searching in the category database according to the target feature vector to determine whether a target object exists; and determining whether the target object invades the target site according to pixel information of the target object and a warning contour line. In the application, the target segmentation network is trained by using the public data set, and the category database is constructed, so that the dependence on a training data set is reduced, a user can freely construct the category database according to a specific scene, the perimeter defense method of the application can be adapted to various scenes, the searching speed is accelerated, and the perimeter invasion determination is more accurate.

The application relates to the field of footprint identification, and discloses a barefoot footprint comparison method and device based on deep learning. The method comprises the following steps: obtaining a to-be-identified barefoot footprint, and determining a toe region and a sole region in the to-be-identified barefoot footprint through a footprint segmentation unit; determining a plurality of intermediate target barefoot footprint traces corresponding to the to-be-identified barefoot footprint in a barefoot footprint trace database according to the toe region through a toe region identification unit; determining a final target barefoot footprint trace corresponding to the to-be-identified barefoot footprint in the plurality of intermediate target barefoot footprint traces according to the sole region through a sole region identification unit, and taking the final target barefoot footprint trace as an identification search result; wherein the toe region identification unit is obtained through training of toe region data sets of a plurality of barefoot footprint traces. The application can improve the speed of footprint identification search in the barefoot footprint trace.