24 results about "Multi domain" patented technology

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

This invention relates to a self-driven, self-generated electric energy recovery system for multi-domain aircraft, encompassing the fields of aircraft, new energy transportation, and general electromechanical technology. The invention employs an integrated upper and lower stacked coaxial architecture, with the upper layer housing the drive motor unit, the lower layer the power generation unit, and a shared stator in the middle. The stator disk is equipped with multiple sets of perforated holes to achieve weight reduction and convective heat dissipation. The stator extends outward to form an extended protective section, the width of which can be flexibly adjusted as needed. Correspondingly positioned strip-shaped limiting tracks are provided on the outer shell, forming a roller-free anti-deviation limiting structure. Conventional insulation or high-temperature resistant reinforced coils can be selected according to operating conditions to adapt to various complex application scenarios. This invention features a unified core principle, a flexible and adaptable shape, and can be applied across multiple fields including aircraft, road vehicles, rail transportation, and rotating electromechanical equipment. Compared to existing split-type, roller-limited, single-energy-harvesting systems, this invention offers higher integration, superior heat dissipation, lighter weight, self-driven collaborative power generation, stronger reliability, and a wider range of applications.

This invention provides a multi-domain image enhancement method and apparatus based on memory and decision fusion, relating to the field of intelligent image enhancement technology in scenarios with large distribution differences across multiple domains. The method includes: upon receiving an image to be enhanced from an unknown domain, extracting at least one image pair from a memory unit, the image pair including an original image from a known domain and an enhancement image label corresponding to the original image; invoking a domain alignment unit in a decision unit to align the image to be enhanced with each original image, obtaining aligned images corresponding to each original image; optimizing the image enhancement unit in the decision unit based on the aligned images and enhancement image labels corresponding to each original image; and performing image enhancement on the image to be enhanced based on the optimized image enhancement unit, obtaining the enhanced image corresponding to the image to be enhanced. This invention can significantly improve the image enhancement effect and cross-domain generalization ability of multi-domain images.

This invention provides a client-side domain-aware, maskless privacy identification and rewriting system and method. The system includes an offline training module deployed on a server, used to generate domain privacy prototypes representing privacy semantics in different domains based on labeled privacy fragments from multiple domains. An initial rewriting model is trained using these domain privacy prototypes to obtain a privacy rewriting model. An online inference module deployed on the client side receives user input text, automatically identifies privacy fragments in the user input text using the domain privacy prototypes, and calls the privacy rewriting model to rewrite the identified privacy fragments. This invention solves the technical problem that existing privacy rewriting technologies rely on explicit user annotation of privacy fragments and lack domain adaptability, making it difficult to balance automated privacy protection and semantic utility preservation in actual client deployment scenarios.

This invention relates to a method and system for dynamic quality analysis of standardized residency training bases. It constructs a hierarchical model for the training base, including at least a target layer, a criterion layer, and a protocol layer. Based on the judgments of multi-domain experts, a judgment matrix for the criterion layer indicators is constructed, and the weights of these indicators are calculated and consistency checks are performed. Structured operational data of the target base in the protocol layer is automatically collected from at least one external medical information system. The comprehensive quality score of the target base is dynamically calculated and stored periodically or continuously based on the weights and structured operational data. Based on the stored comprehensive quality scores obtained from multiple consecutive calculations, a quality change trend chart of the target base is generated. Based on the comprehensive quality score, its weak points are identified, and the comprehensive quality score, quality change trend chart, and weak points are output. Through the effective utilization of daily operational data of the residency training base, objective and continuous observation of the base's quality is achieved.

The application provides a word element aggregation large model architecture optimization method and system based on multi-field knowledge mixing, relates to the technical field of artificial intelligence, and obtains an initial word element sequence set and a multi-source field label set, then performs hierarchical word element grouping to obtain a pure domain word element segment set and a boundary bit sequence set thereof, then performs context segment aggregation to obtain a cross-domain segment association chain set and a semantic dependency distribution description between segments, dynamically regulates a word element aggregation radius according to the cross-domain segment association chain and the semantic dependency distribution, generates a radius self-adaptive domain fusion word element aggregation cluster set and a distribution density map in the cluster, generates a large model architecture parameter adjustment instruction set based on the domain fusion word element aggregation cluster set and the distribution density map to adjust an attention receptive domain range and a hidden layer expansion coefficient. The application solves the feature representation confusion problem caused by multi-field mixed input, realizes the analysis of the word element sequence field structure and the adaptive adjustment of the model architecture parameters.

This invention discloses a multi-domain adaptive design rule method based on EMC data assets. It includes: acquiring the target application scenario identifier, functional type, IC combination, and design objectives; retrieving matching typical basic products and their derivative product cases to obtain their feature sets and mapping relationships; constructing an initial design concept by combining IC-level and system-level data assets; retrieving the standard essential characteristics of the target application scenario; constructing an EMC domain adaptation model, establishing an adaptive mapping based on verified mapping relationships, and quantitatively correcting and generating design rule parameters; outputting design rules containing constraint objects, types, target values, source metadata, and protection layer priorities; and structurally storing and assigning multi-dimensional tags. This invention establishes cross-application scenario mappings through typical basic products and derivative products, achieving quantitative adaptive generation of design rules, supporting cross-domain reuse, test prediction, rule self-evolution, and automatic generation of design documents, shifting EMC design from experience-based to data-driven.

This invention presents a unified framework for addressing complex, multi-domain challenges through adaptive design optimization, equation discovery, and hypothesis generation. Central to the framework is the Mathematical Sphere Framework (MSF), which employs advanced machine learning and hybrid computing to interrelate equation families across fields such as CFD, FEM, CAD, and PLM. MSF enables cross-domain optimization with transparent, physics-based representations, fostering continuous insight generation and system evolution. A neural-assisted system refines equations from experimental and real-world data, advancing theoretical models and design exploration. Hybrid computing combines classical preprocessing with quantum optimization to enhance computational efficiency. A digital twin provides real-time simulation and predictive analysis, while lifecycle adaptability dynamically optimizes parameters across product stages. By improving efficiency, scalability, and interoperability, this invention transforms engineering workflows, supports innovation, and fosters cross-industry applications, including aerospace, energy, and manufacturing, driving progress in an ever-evolving landscape.

This application relates to a multi-domain intelligent interactive system based on bottle QR codes. The system includes: an encoded data acquisition module that collects encoded data; a storage, query, and judgment module that retrieves data stored locally on the communication device and generates corresponding judgment result requests; a data visualization module that receives the encoded data and judgment result requests, mines tourism geographic information datasets using data extraction algorithms, and then transforms this information into visualized content using visualization algorithms; and a user recommendation module that collects social feature parameters input by users, comprehensively analyzes these parameters and the visualized content, and generates a personalized list of recommended users. This system achieves deep integration of bottle QR codes with multiple domain functions, greatly improving the convenience of information acquisition and social interaction, and meeting diverse needs in tourism and social interaction.

This invention provides an AI-based method for classifying and identifying incompatible hazardous chemicals, relating to the field of intelligent classification and discrimination technology. The method specifically includes the following steps: Step 1, acquiring multi-domain data on the entire lifecycle of hazardous chemicals, integrating and fusing the multi-domain data to construct a multi-source heterogeneous database of hazardous chemicals; Step 2, defining text and numerical variables based on the naming rules and attributes of chemicals, and extracting hazardous chemical judgment conditions and incompatible relationships from the multi-source heterogeneous database using natural language processing and knowledge extraction techniques, establishing a hazardous chemical rule base, and constructing an initial knowledge graph of hazardous chemicals containing chemical nodes, attribute edges, and relationships based on graph neural network technology. By integrating hazardous chemical information through multi-source data fusion and knowledge graph technology, intelligently analyzing and mining causal relationships, and combining graph neural network reasoning, accurate classification and incompatible judgment are achieved.

A kind of multi-domain simulation tool automation software architecture and method based on SaaS deployment, it is related to simulation design software architecture technical field, the software architecture includes: front-end architecture, back-end architecture and communication interface;The front-end architecture is used to provide interactive interface to user, user sends request to back-end architecture through the interactive interface;The back-end architecture is used to carry out demand analysis after obtaining front-end request, creates simulation calculation task according to analysis result, and processes the simulation calculation task;The communication interface is used to mutual transmission front and back end information, uses REST form, and communicates with JSON data format by HTTP request.This application supports simulation model zero code integration, non-professional simulation personnel do not need to learn professional simulation software operation, can quickly build simulation scene only by front-end business parameter configuration, greatly expands the application coverage of simulation technology in enterprise.

A data processing system implements systems and methods for performing multi-domain queries that combine explicit graph walks and fuzzy semantic reasoning of language models. These systems and methods leverage the fuzzy semantic reasoning capabilities of a language model during query execution to generate query results that are based on inferences that are not expressly defined in the graph. These techniques utilize grounding and prompt construction techniques to ensure that the language model generates an output that is consistent with a global schema that represents the types of data available from the graph data sources to significantly reduce the risk that the language model will hallucinate and generate erroneous results.

The application relates to the technical field of knowledge graphs, in particular to a cross-domain data fusion and dynamic knowledge graph construction system for industry chain analysis, which comprises a data processing module, a graph construction module, an analysis module, a robustness determination module, an update adjustment module and a trigger adjustment module. The data processing module comprises a collection unit for collecting multi-source heterogeneous data of multi-field data sources. The graph construction module comprises a construction unit for constructing a knowledge graph according to the cross-domain aligned data. The analysis module is used for analyzing an industry chain according to the incremental knowledge graph to obtain an analysis result. The robustness determination module is used for determining whether the robustness of the multi-source heterogeneous data meets the requirements according to a multi-source data fusion representation value. The update adjustment module is used for determining an attenuation coefficient of multi-source heterogeneous data update credibility according to an update lag rate of the knowledge graph. The trigger adjustment module is used for determining an extension trigger threshold of the knowledge graph according to the fitting degree of the analysis result and an actual scene. The application improves the stability of industry chain analysis.

A data processing system implements systems and methods for performing multi-domain queries that combine explicit graph walks and fuzzy semantic reasoning of language models. These systems and methods leverage the fuzzy semantic reasoning capabilities of a language model during query execution to generate query results that are based on inferences that are not expressly defined in the graph. These techniques utilize grounding and prompt construction techniques to ensure that the language model generates an output that is consistent with a global schema that represents the types of data available from the graph data sources to significantly reduce the risk that the language model will hallucinate and generate erroneous results.

The embodiment of the specification discloses a kind of CAX-based intelligent collaboration method, equipment and medium, it is related to intelligent collaboration technical field, method includes: receiving the natural language design requirement input by user, to carry out multimodal retrieval in the preset CAX tool library, determine candidate tool set, and pre-verification is carried out to the candidate tool set in lightweight sandbox, and screening executable tool;According to the natural language design requirement, identify potential multi-domain constraint conflict in the preset unified industrial semantic network, to carry out automatic conflict resolution between multiple professional agents based on predefined negotiation rules by A2A agreement, and generate executable task sequence;Through the multiple professional agents, according to the executable task sequence, the corresponding CAX operation is executed by calling the executable tool, and intelligent collaboration is realized.

This application provides a method, apparatus, electronic device, and storage medium for constructing a multi-domain knowledge graph, relating to the field of knowledge graph construction technology. The method includes: constructing a text dataset and performing domain filtering based on a pre-selected large language model to obtain domain text datasets for target domains (single-domain, related-domain, and full-domain); further performing sentence segmentation, word segmentation, and part-of-speech tagging; extracting nouns from the segmented word dataset to obtain target entities; performing word segmentation and dependency syntax relation extraction on each sentence in the target domain text dataset, and reconstructing a dictionary list of target dependency syntax relations to extract relationships between different domains; and constructing a multi-domain knowledge graph based on entities and relations. The multi-domain knowledge graph constructed in this application is more comprehensive, accurate, and efficient, requiring no large amounts of data for model training, thus saving significant time, manpower, and computational resources.

The application belongs to the technical field of knowledge engineering, and discloses an AI knowledge base construction method and system for cross-domain collaboration, which comprises the following steps: based on a symbolic dynamic knowledge graph, an adversarial neural network is used to decouple the domain-independent intention and the domain feature vector in a user query, a generalized intention vector and a personalized weight parameter are generated; the semantic similarity between the generalized intention vector and the dynamic knowledge graph is calculated, a multi-hop retrieval is used to recall multi-domain high-relevance knowledge fragments by using HNSW, and a cross-domain knowledge set is obtained; a variational Bayesian uncertainty estimation method is used to scan the knowledge fragments in the cross-domain knowledge set and shield illegal content, and a personalized cross-domain answer is generated in combination with the personalized weight parameter. Through the generation of the generalized intention vector and the personalized weight parameter, the application realizes user intention decoupling and personalized weighting, adapts the query to multiple domains, and improves the relevance of knowledge retrieval.

This invention discloses a method and apparatus for knowledge fusion and push based on reinforcement learning for equipment fault diagnosis, comprising: selecting single-domain fault diagnosis and multi-domain fault diagnosis domains, constructing original corpora for each fault diagnosis domain, encoding the original corpora to obtain knowledge data and text block segmentation results; performing graph clustering on the encoding vectors of each entity node obtained based on reinforcement learning, and grouping entity nodes with similar semantics into a community; generating community reports for the divided graph communities based on a large model, constructing community nodes to build a hierarchical knowledge graph based on the generated community reports; encoding user input based on a preset model, and performing encoding matching between user query encoding and entity encoding using a cosine similarity-based method, using the matched entity nodes to perform knowledge query in the constructed hierarchical knowledge graph, and inputting the queried knowledge into the large model to generate a fault diagnosis solution.

The present application relates to a kind of long context question and answer generation method based on logic template and anchor point dispersion, comprising the following steps: data input and preprocessing, structured segmentation and overlap buffer, content type discrimination and routing, logic template driven construction, anchor point position planning, long reference text assembly and anchor point dispersion delivery, approximate interference construction and injection, question and answer generation, quality check and screening output.Compared with prior art, the present application has the advantages of constructing logic skeleton in short window and forming long-range dependence through programmed layout, effectively suppressing keyword matching shortcut, accurately filtering pseudo-long samples and flexibly adapting multi-domain and multi-length settings.

Methods and systems are presented for configuring and training a machine learning model using transfer learning techniques that can transfer knowledge among multiple domains that do not share an identical feature set. Instead of using any feature set associated with a domain, a feature arrangement that combines all of the feature sets associated with the multiple domains in a particular organization for configuring and training the machine learning model. The feature arrangement includes a domain independent section and multiple domain-specific sections corresponding to the multiple domains. The domain independent section includes common features that are common across the multiple domains. Each of the domain-specific sections includes a feature set associated with the corresponding domain. The machine learning model that is configured in this manner can be trained to learn knowledge across the multiple domains and subsequently perform tasks for the multiple domains.

The application discloses a kind of lightweight cross-domain recommendation method and system based on user alignment Agent drive.It first obtains the historical behavior data of user in multiple fields, fuses text, image and other multi-modal content, generates fine-grained interest prototype through cross-domain semantic encoder, and constructs personalized Agent to simulate user intent.Then, in the multi-field collaborative environment, the behavior strategy of Agent is optimized using reinforcement learning and hybrid reward mechanism, and the general preference and domain-specific preference are modeled through hierarchical strategy network, and knowledge fusion is realized through gating mechanism.Subsequently, combined with preference distillation technology, transferable representation is extracted from Agent behavior, and lightweight cross-domain knowledge graph is constructed.Finally, by using behavior trajectory compression and cross-domain preference mapping, efficient and low-consumption personalized recommendation is realized.This method effectively alleviates the problems of cross-domain data sparsity and model complexity, improves recommendation accuracy and system response efficiency, and is suitable for real-time recommendation services in multiple scenarios.

This invention discloses a multi-domain collaborative industrial storage end-to-end redundancy correction system. The invention relates to the field of telecommunications technology and includes an industrial data preliminary processing end, an industrial data redundancy feature acquisition end, an industrial data redundancy feature classification end, and an industrial data redundancy filtering and elimination end. The preliminary processing end periodically receives industrial sensor data packets, extracts five-tuple information for dirty data cleaning, and outputs standard data blocks stored in a cache pool. The redundancy feature acquisition end extracts data from the cache pool, identifies the device cluster, analyzes the standard data blocks within the same monitoring period, and outputs a set of redundancy feature vectors. The classification end evaluates redundant and non-redundant data packets according to priority. The filtering and elimination end extracts redundant data packets, removes old data packets based on timestamps, and retains new data packets. This achieves end-to-end redundancy correction for industrial storage.

A system and method for administering a game in which live sports game is displayed and / or results retrieved, and a virtual game associated with the live sports game is established along with other game attributes of game play are combined to determine player scores a rankings.

The invention relates to the technical field of data management systems, and particularly discloses a cross-domain trusted data exchange and access management and control system, which comprises an interaction management gateway, a strategy control engine, a trust evaluation engine and a block chain core layer. Through collaborative design of attribute unified mapping, dynamic multi-dimensional trust evaluation, double-chain heterogeneous evidence storage and synchronization, self-adaptive encryption compression transmission, TEE hardware isolation protection and ciphertext strategy fine-grained management and control, pain points such as difficulty in cross-domain attribute adaptation, trust evaluation staticization, interaction non-traceability, transmission efficiency and security imbalance and the like are effectively solved; high credibility, high efficiency and high security of cross-domain data exchange are realized, differentiated management and control requirements and system performance balance are considered, and the method is adaptive to a multi-field cross-domain data sharing scene.