11 results about "Semantic analyzer" patented technology

A system for dynamic latent space adaptation using spatiotemporal kernel context for multiscale rendering with hierarchical and Lorentzian autoencoders. The Spatiotemporal Kernel Estimator (SKE) analyzes media through motion field, temporal recurrence, frequency band, and scene semantics analyzers to generate adaptive kernel parameters encoding content-specific importance distributions. The system dynamically adapts latent manifold geometry by modifying metric tensor properties according to kernel context, enabling content-aware compression that allocates representational capacity based on visual significance. A multiscale cache implements kernel-adaptive retention policies prioritizing important regions. An adaptive renderer provides intelligent level-of-detail selection based on zoom level and kernel-estimated importance, optimizing processing allocation. The self-optimizing architecture continuously refines kernel context and geometric adaptation based on user interaction and performance feedback, achieving superior compression ratios and perceptual quality. Applications include bandwidth-efficient video streaming, virtual reality, scientific visualization, and cognitive video analytics requiring intelligent context-aware visual processing.

The invention discloses a streaming knowledge injection and adversarial self-optimization large language model training method and system, and the method comprises the steps: collecting the newest knowledge of an authoritative information source in real time, converting the newest knowledge into a structured constraint rule through a semantic analyzer, and dynamically updating a knowledge base; based on the updated knowledge base, adopting a PPO algorithm to optimize a generator, actively constructing a high-risk adversarial sample, and forcing the model to expose security vulnerabilities; after user input and adversarial samples are input into the model, total loss is calculated through constraint detection, gradient updating is blocked if the total loss exceeds a threshold value, and otherwise, a multi-level safety verification stage is started; and finally, fusing a security verification result and the adversarial loss, updating a total loss function and cooperatively adjusting model parameters to form a continuous self-optimization training cycle. According to the method, compliance is guaranteed through streaming knowledge injection, vulnerabilities are actively mined in combination with adversarial training, verification precision is improved by means of multi-level detection and dynamic threshold adjustment, and safety and reliability of a large language model in a complex scene are enhanced.

A system for dynamic latent space adaptation using spatiotemporal kernel context for multiscale rendering with hierarchical and Lorentzian autoencoders. The Spatiotemporal Kernel Estimator (SKE) analyzes media through motion field, temporal recurrence, frequency band, and scene semantics analyzers to generate adaptive kernel parameters encoding content-specific importance distributions. The system dynamically adapts latent manifold geometry by modifying metric tensor properties according to kernel context, enabling content-aware compression that allocates representational capacity based on visual significance. A multiscale cache implements kernel-adaptive retention policies prioritizing important regions. An adaptive renderer provides intelligent level-of-detail selection based on zoom level and kernel-estimated importance, optimizing processing allocation. The self-optimizing architecture continuously refines kernel context and geometric adaptation based on user interaction and performance feedback, achieving superior compression ratios and perceptual quality. Applications include bandwidth-efficient video streaming, virtual reality, scientific visualization, and cognitive video analytics requiring intelligent context-aware visual processing.

This invention discloses a method for extracting yeast transcription factor-gene relationships from biological text. The method includes: retrieving literature related to yeast transcription factor-gene relationship extraction; constructing an input representation; constructing a relationship extraction model; inputting the input representation into the trained relationship extraction model to obtain a pre-selected output representation containing transcription factor-gene binding or regulatory information; inputting the output representation into the trained relationship extraction model to obtain a first probability; inputting the output representation into a syntactic analyzer and a semantic analyzer respectively, and then into a graph neural network to obtain a second probability; setting a loss function for secondary training; inputting real-time collected literature into the model each time; and finally obtaining a joint probability based on the first and second probabilities. If the joint probability is greater than a preset value, the output representation is considered to contain transcription factor-gene binding or regulatory information, and the corresponding output representation is output. The advantage of this invention is that the relationship extraction is more accurate.

The invention provides a multi-modal data fusion power distribution network fault diagnosis method and system, and the method comprises the steps: carrying out the feature extraction of a real-time current and voltage waveform measured by an SCADA system based on an edge feature extractor, and generating an electrical feature vector; performing semantic analysis on the marketing repair text based on a text semantic analyzer, extracting a semantic feature vector, and obtaining fault probabilities of different types of marketing systems based on the semantic feature vector; performing health assessment on the standing book and the historical data of the to-be-monitored equipment based on an equipment health assessor, and obtaining a health score of the to-be-monitored equipment; according to the electrical feature vectors, the fault probabilities of different types of marketing systems, the health score of the to-be-monitored equipment and a dynamic weight distribution algorithm, the confidence and category of the current power distribution network fault are determined, and the accuracy of fault diagnosis is effectively improved.

The invention relates to the technical field of building engineering information, and discloses an engineering project digital archive management method based on BIM. The method comprises the following steps: initializing a BIM platform environment, importing a BIM model and related archive data, and carrying out standardization processing; constructing a project component association graph and an archive data flow graph by using a graph technology; and capturing state change through event monitoring, and synthesizing a project operation state diagram and an archive data dynamic flow diagram in real time. The core innovation lies in that a row calls a semantic analyzer and an authority checker to scan and check a double-dynamic graph, and outputs an abnormal identifier set; positioning an abnormal node and performing dependency chain backtracking along an archive flow graph, and accurately extracting a problem root; a test case is generated by combining static analysis and a novel test strategy, and defect details are monitored and recorded during operation. And integrating the problem root and the defect record to generate a comprehensive management report. According to the invention, the real-time performance of engineering file management and the accuracy of problem tracing are improved.

The invention discloses a non-inductive parking fee payment management method and platform, and relates to the technical field of intelligent parking. The method comprises the following steps: acquiring vehicle videos shot by cameras inside and outside a field, and forming a visual sequence after frame extraction; analyzing the sequence by using a visual semantic analyzer to generate a semantic long description text; constructing an abnormal behavior query problem set according to a charging rule and a historical abnormal record; calling a multi-modal question and answer model, answering questions based on the semantic text, and outputting fine-grained answers and confidence; integrating question and answer results, and constructing a multi-node relation heterogeneous graph; and carrying out anomaly analysis on the graph, if an anomaly is found, sending out an early warning, and managing the cost according to the early warning. The technical problems that in an existing non-inductive payment parking scene, due to the fact that visual data understanding is shallow and multi-source information is isolated, abnormal behavior detection is inaccurate, and tracing is difficult are solved, and the technical effect that through deep semantic understanding and multi-relation graph analysis, the parking lot vehicle behavior recognition precision and parking fee management safety are improved is achieved.

The invention discloses a parameterized base rapid modeling method based on a Smart3D platform, which comprises the following steps: step S1, packaging a modeling function interface based on a user-defined interface provided by Smart3D, and abstracting a complex modeling function into a simple calling method; s2, creating a simple and efficient semantic analyzer, performing semantic analysis on the script file, and performing modeling control on keywords and parameters; s3, a designer writes a script file in a text format to serve as a specific modeling input definition, script recognition and analysis are conducted through a program, batch modeling of bases is conducted in a three-dimensional space, and a generated base model is exported to be in a needed format. According to the method, the modeling efficiency is greatly improved, the operation is simple and free of technical barriers, code development is not involved, and a designer can quickly operate without a coding basis.

The application discloses a PE compiling method of an AI chip, comprising the following steps: obtaining source code, preprocessing the obtained source code to obtain preprocessed source code, performing lexical analysis and syntax analysis on the preprocessed source code in sequence to obtain a syntax tree corresponding to the source code, processing the syntax tree by using a semantic analyzer to generate a symbol table, generating an intermediate representation corresponding to the preprocessed source code according to the syntax tree and the symbol table and using a code conversion module inside a low-level virtual machine (LLVM) compiler, converting each of all operations and all data structures in the intermediate representation corresponding to the preprocessed source code into a corresponding PE instruction according to a PE instruction set specification of the AI chip, and performing parallel optimization on each PE instruction by multiple PEs to obtain an execution result of the optimized PE instruction. The application can solve the technical problem that a traditional compiler cannot fully exert performance because PE optimization is not performed for the AI chip.

The application discloses a kind of meteorological water conservancy disaster intelligent grading early warning linkage calling routing method and system.Through the construction meteorological water conservancy double factor early warning matrix, establish cross-department data fusion channel, develop early warning coupling model, realize multidimensional early warning fusion.Formulate water conservancy special response strategy library, including river water level grading route and mountain torrent risk area customization strategy.Develop intelligent water conservancy enhancement module, such as flood control engineering topology analysis engine and water rain work condition semantic analyzer.With the first dynamic coupling algorithm to build three-element risk rating model, based on double dimension water conservancy emergency response topology network and water conservancy engineering BIM data and meteorological early warning space overlay analysis module and other characteristic innovation features.Can improve the accuracy and comprehensiveness of disaster warning, ensure the timeliness and effectiveness of emergency response, improve the intelligent level of system and disaster response capability.

The application discloses a kind of legal document generation method and device, electronic equipment, storage medium, it is related to legal document generation technical field or other related fields, wherein, the method comprises: receiving first language case information and legal document generation demand;First language case information is analyzed by the semantic analyzer corresponding to first language, and the case elements of target case are obtained, and first language case information is translated by translator, and second language case information is obtained;Case elements are based on inquiring target legal knowledge base, and the query result is obtained;Target document template is determined based on legal document generation demand, and target legal document is generated based on target document template, second language case information, case elements and query result.The present application solves the technical problems that related art cannot accurately understand and convert professional legal terms and context in the process of cross-language generation of legal documents, leading to inaccurate or even non-compliant legal documents.