18 results about "Semantic system" patented technology

The invention discloses a vision-language-action robot control method and system based on task stage semantic enhancement, and the method comprises the steps: firstly constructing a task stage semantic system, and disassembling a task into stages with definite semantics; constructing a task stage observation module based on a pre-trained vision-language model, deducing a task state and abnormality in real time, and generating stage semantic features; modulation parameters are generated through a lightweight network, and dynamic modulation is carried out on the vision-language features to strengthen the vision feedback weight; and finally, fusing the modulated characteristics and the body state characteristics, and outputting a control instruction through an action generation network to form closed-loop control. According to the method, the multi-modal information weight can be dynamically balanced, task abnormity can be identified in real time, error correction is triggered, the false completion phenomenon is remarkably reduced, the task execution reliability, robustness and generalization ability of the robot in a complex dynamic environment are improved, and the method is suitable for various robot systems needing high-credibility operation.

The invention discloses a remote bank fuzzy question-answer accurate matching method based on deep learning. The method comprises the following steps: firstly, constructing a business semantic system (clear business class-scene-specific intention), a term and variant mapping library and a sentence pattern template containing a replaceable entity slot position; extending the original similar question pairs to generate derivative sentences and candidate similar sentences, filtering, clustering to generate cross-sentence pattern similar pairs, and selecting negative samples to control the literal overlapping rate; then constructing a twin BERT network, mixing original and enhanced samples into a training set, and training in two stages by using improved triple loss; finally, sentence vectors are generated through the trained twin network, semantic similarity is judged according to cosine similarity, question pairs which are judged to be similar through a model are subjected to secondary verification through a service rule base, and score reduction is carried out if the question pairs do not pass verification.

The invention relates to the technical field of MCU instruction driving, and discloses an AI-based MCU instruction driving generation method and system.The AI-based MCU instruction driving generation method comprises the steps that a concept ontology network is obtained according to a unified semantic representation framework, and a mapping relation between multi-language expression and a core concept is established; according to the concept ontology network, a holographic semantic system architecture is obtained; determining a conceptual language joint learning mechanism according to the holographic semantic system architecture; zero sample cross-language migration capability is realized according to a conceptual language joint learning mechanism; and obtaining an ontology-based knowledge organization system in combination with the zero sample cross-language migration capability to form structured MCU-driven domain knowledge representation. Through the unified semantic representation theory and the concept ontology network, the understanding accuracy of the same technical requirement expressed in different languages is improved.

The invention belongs to the technical field of intelligent identification and classification of multi-modal data, and discloses an intelligent data identification and classification method based on multi-modal feature fusion, and the method achieves the precise and unified representation of heterogeneous modals through a semantic anchoring double-space architecture and a two-stage alignment mechanism. The shared semantic space depends on an LLM pre-training semantic system and a triple anchor point constraint conversion direction, the modal specific space reserves unique attributes of a modal, and dual-space dynamic weight adjustment can balance semantic consistency and specificity; two-stage alignment is combined with semantic anchor point rough matching and adaptive temperature coefficient fine optimization, and time sequence offset is corrected by matching with time-space synchronization verification, so that semantic dislocation and error accumulation are avoided; a multi-dimensional credibility evaluation system and a dynamic complementation strategy are constructed, according to different credibility grading, matching transfer learning, hybrid generation or semantic guidance and other complementation modes, the semantic validity of complementation features is guaranteed in combination with a quality feedback mechanism, and semantic deviation after complementation is reduced.

A system and method guide an Artificial Intelligence engine to determine the semantic and grammatical correctness of a user-expanded sentence in real-time. The sentence validation process involves receiving input from the user, the input includes sentence fragment that the user wishes to expand and user-expanded sentence that the user constructs on the fragment provided. The inputs are broken down into tokens. The word-level tokenization algorithm is used, which identifies tokens by splitting the text into spaces, punctuation marks, and other delimiters. Further, a token comparison algorithm is used to assess the relationship between the sentence fragment and the user-expanded sentence to analyze order and placement. Once the token comparison is complete, a prompt is generated using prompt generator to evaluate grammatical and semantic evaluation of the user-expanded sentence. Real-time feedback is provided to the user based on grammatical and semantic evaluation.

A cost effective big data retention management system can be implemented. The system may include a retention tagging system, which allows metastore tables managed by a query engine to be tagged with a retention policy. The system may include a partition semantics system, which can understand the retention tags and produce retention events. The system may include an event consumption system, which can process the retention events and execute a set of operations for the retention events. The system may include a data restoration system, which allows archived data to be restored as needed. The system may include a listener, which can enforce one or more rules relating to retention-related table properties.

This application discloses a security protection method, device, storage medium, and electronic device. The method includes: analyzing the source code of the target program by combining the business semantics of the target program, the kernel semantics of system calls, and the scenario semantics of program execution; generating a system call whitelist; forming a three-dimensional semantic constraint rule set based on the linkage relationship between business semantics, kernel semantics, and scenario semantics; binding the protection strategy to the running process of the target program to intercept system calls initiated by the process that are not on the system call whitelist; collecting the runtime business semantics, kernel semantics, and scenario semantics of system calls within the system call whitelist to generate a system call sequence; determining deviation items based on the system call sequence and the three-dimensional semantic constraint rule set; dynamically quantifying and evaluating the deviation items to obtain a risk level; and implementing corresponding permission control or blocking operations on the process according to the risk level to achieve runtime security protection for the target program.

The application discloses a museum immersive exhibition optimization method based on cross-domain digital twinning. The method carries out three-dimensional modeling on entity exhibits, generates exhibit digital twinning bodies in combination with geometric shape, material and photoacoustic calibration data; collects geographical, architectural, historical and social and cultural data associated with the exhibits, establishes a unified space-time coordinate and semantic system, and forms a cross-domain context data set; inputs the structured information of the exhibit digital twinning bodies and the context data set into a multi-source fusion engine, constructs a cross-domain semantic space-time knowledge graph and generates a cross-domain coupling graph, and realizes bidirectional constraint of the entity environment and the virtual scene; under a given exhibition theme and immersion performance index, a context rendering parameter set is generated to guide virtual reality rendering; context parameters are dynamically adjusted according to visitor behavior and physiological feedback, and adaptive optimization of the exhibition is realized. The application improves the intelligent interaction and immersion experience of virtual and real exhibitions.

The invention discloses an organization-level semantic cognitive map construction system which is applied to enterprise talent and knowledge management. The system constructs a comprehensive enterprise skill knowledge map by integrating internal multi-source data, and draws a semantic map of internal knowledge and talent distribution of an organization. And through time sequence analysis, the semantic trajectory change of the employees is tracked, and a team potential map and future trend prediction are generated. The system can identify and organize potential knowledge blank and talent risks, such as talent faults or knowledge islands, and early warns and prompts managers to take measures in advance. By introducing a semantic sovereignty concept, an enterprise can autonomously define a semantic system thereof, and it is ensured that the map conforms to the own context. The invention provides a visual white-box analysis tool to help a decision-making layer to understand and organize the current situation and trend of intellectual capital, so that the talent strategic layout is optimized, and the method has high authorization possibility and industrial application prospects.

A method for UAV non-orthogonal communication and edge scheduling based on semantic symbol matching comprises the following steps: acquiring and partitioning the dataset, constructing an air-to-ground semantic system, matching semantic symbol sequences, correcting semantic errors and recognizing images, constructing a feature fusion strategy network, determining unloading, training the feature fusion strategy network, and testing the feature fusion strategy network. This invention employs a Transformer sequence prediction model, using semantic information to analyze the spatial correlation of the context, and repairs damaged semantic symbols caused by multi-vehicle non-orthogonal collisions or fading, thus improving the success rate. The UAV calculates the direction vector and spatial similarity of the observed signals, and uses the Hungarian algorithm to complete global matching, enabling concurrent vehicle unloading and improving the processing scale of the vehicle-to-everything (V2X) network. A multi-branch feature fusion strategy network is adopted, combining the policy network and the value network to construct a near-end policy optimization framework, achieving adaptive perception of real-time state changes of vehicles and UAVs, and improving system resource utilization.

The invention discloses a text prompt type heart function intelligent evaluation method based on a visual language large model. The method comprises the following steps that 1, an echocardiography mark data set is constructed; 2, dividing four chambers of the heart; step 3, hierarchical feature learning; 4, performing multi-scale feature fusion; and 5, predicting the cardiac function (ejection fraction). According to the method, the heart can be modeled into a multi-scale semantic system, and features from macroscopic chamber dynamics to microstructure motion are captured through a hierarchical feature learning mechanism. Deep integration of imaging data and clinical knowledge is realized by aligning hierarchical visual features with professional medical descriptions in a shared semantic space. The method has excellent performance in the aspect of ejection fraction prediction.

The invention discloses a whole-process judgment auxiliary legal semantic system supporting legal applicable rule matching. The system is constructed based on a DIKWP semantic model and comprises a case element input module, a law knowledge base module, a semantic retrieval and matching module, a law reasoning and screening module, a judgment rule integration module, an intention layer calibration module, an interpretation output module and the like. According to the system, candidate law articles are preliminarily matched from a legal knowledge base through semantic retrieval, then applicability verification and conflict detection are conducted through a legal knowledge graph and an inference engine, and competing and contradiction between laws and regulations are automatically recognized and resolved; and finally, the matching result is calibrated in combination with the judgment purpose and the legal value in the intention layer, and the legality and rationality of the recommendation scheme are ensured. According to the method, law application suggestions with white box inference chain explanation can be output, the problems of inaccurate law matching, lack of deep semantic understanding and opaque decision making process in the prior art are effectively solved, and the efficiency and accuracy of law retrieval and application by a judge are remarkably improved.

The invention discloses a whole-process judgment auxiliary legal semantic system supporting fact semantic extraction and dispute focusing. The system is constructed based on a DIKWP mesh cognitive model and comprises a data acquisition and preprocessing module, an information extraction and semantic association module, an intelligent analysis and dispute focus recognition module, a judgment intention association and priority ranking module, a white-box reasoning chain recording and visualization module and a Bug detection and consistency verification module. The system realizes layer-by-layer abstraction and feedback optimization from original data to judgment intentions by performing multi-level semantic analysis on case materials, and can accurately identify case dispute focuses. The method is characterized in that a consciousness relativistic mechanism is introduced to realize multi-role semantic alignment, a white-box interpretable inference chain tracing function is realized, and semantic contradictions can be automatically detected to realize logic self-correction. According to the method, the depth of factual understanding of complex cases and the accuracy of dispute focus extraction are effectively improved, and the method can be widely applied to the judgment auxiliary whole process of civil and criminal cases and the like.

A method for operating a semantic robotic system in communication with a plurality of devices, the semantic robotic system having at least one processor and at least one transceiver includes storing in the memory an intrinsic semantic associated with an object type, the intrinsic semantic further indicating a purpose of the object type. The method further causes the at least one processor to detect a first object of a first object type based on one or more first data inputs, infer a first semantic in rapport with the first object of the first object type based on one or more second data inputs, generate semantic augmentation data based on a determination that the first inferred semantic is non-affirmative by having a high entropy with respect to the intrinsic semantic, and cause the at least one transceiver to transmit the semantic augmentation data to a remote device.

The invention relates to a heterogeneous resource-oriented virtual power plant standardized information model and an automatic registration method, and belongs to the technical field of energy internet and virtual power plants. The method comprises the steps that a unified information description framework is determined, a modular architecture is constructed, a standardized information model is established, and meanwhile semantic labels are given to data elements to form a machine-readable semantic identification system; the resource end generates and reports a modular information packet according to the model, and the platform performs automatic analysis, matching and instantiation on the information packet based on a semantic system to form a standard resource object containing static attributes and a dynamic service interface; and performing multiple verification of data integrity, logic consistency and service interface accessibility on the resource object, and after the verification is passed, completing resource storage, index construction, global unique identifier allocation and schedulable state activation. Plug and play and efficient automatic registration of heterogeneous resources are realized, and the access efficiency, the management precision and the system compatibility of the virtual power plant are improved.

The invention relates to the technical field of judicial information, and discloses a whole-process referee auxiliary legal semantic system supporting referee suggestion generation and interpretation tracking. According to the system, on the basis of a DIKWP mesh semantic model, automatic generation and whole-process semantic reasoning from case facts and legal bases to judgment suggestions are achieved through judgment element summarization, conclusion reasoning generation, logic consistency verification, document generation, interpretation tracking and visualization and a multi-mode output module. The system can generate a judgment conclusion with a legal basis, and records and visualizes an inference chain by using a white box technology. Meanwhile, contradictions or omissions in reasoning are automatically detected and corrected through a logic consistency checking mechanism. For different roles such as judges and parties, the system provides a customized output form, and the transparency and understandability of judgment results are enhanced. According to the invention, the intelligence, interpretability and credibility of the judgment auxiliary system are effectively improved, and the system is suitable for court judgment, dispute mediation, legal service and other scenes.

The invention belongs to the technical field of medical artificial intelligence, and particularly discloses a radiology report generation method based on focus attention and semantic knowledge fusion. According to the invention, based on the pre-trained medical image encoder, multi-focus label prediction is carried out on the radiation light image; performing coarse and fine granularity retrieval on the similar images and the corresponding similar reports; mapping the identified medical entities to a unified medical semantic system; and generating a radiology report corresponding to the radiation light image according to the screened candidate reports based on a report integration generation strategy of chain reasoning. According to the mode, the high-precision candidate reports are screened by adopting a multi-stage retrieval enhancement mechanism formed by label matching, coarse and fine granularity retrieval and knowledge screening, retrieval errors are remarkably reduced, a report integration generation strategy based on chain reasoning controls a large language model to generate texts only in a reference range, medical illusion is avoided, and the accuracy of retrieval is improved. Therefore, the accuracy and interpretability of generating the radiology report can be effectively improved.

A robotic semantic system includes one or more smart robotic devices, which may be configured as a stack of modules including a mobility module and one or more sensor modules. A plurality of robotic modules is communicatively coupled to one another, and use variable semantic coherent inferences to allow the devices to perform semantic augmentation.