334 results about "Semantic similarity" patented technology

This invention relates to the field of industrial equipment fault diagnosis, and provides a self-supervised multi-sensor fault diagnosis method and system based on semantic embedding. The method preprocesses raw normal operating condition monitoring data, generates a fault window through single-channel fault injection using temporal semantic embedding, and generates a supervision signal for the fault window. Features are extracted from the training sample window composed of the pure normal operating condition window and the fault window using long-term Transformer branches and short-term CNN branches. The extracted long-term and short-term features are concatenated and mapped to a shared fault feature representation. A fault attribute matrix is ​​constructed, a semantic prototype library is generated, and semantic similarity is calculated. Based on the shared fault feature representation and semantic similarity, a multi-task prediction head is constructed. Based on the multi-task prediction head and the supervision signal, a multi-task loss function is designed to train an end-to-end fault diagnosis model. Post-processing optimization is combined to correct easily misjudged fault results. Online fault diagnosis results are output.

The invention discloses a medical image treatment method and system for multi-source heterogeneous data, and relates to the field of medical image treatment, and the method comprises the steps: firstly, respectively extracting an image embedding vector and a text embedding vector from original medical image data and a description text thereof through a deep learning model; then, the vectors of the two different modes are fused, and a unified semantic embedding vector is formed; based on the unified vector, through semantic similarity calculation with a standard term library, candidate mapping can be automatically generated, and dependence on rigid artificial rules is eliminated. More importantly, a closed-loop mechanism of manual auditing-feedback learning is introduced in the scheme, high-confidence mapping is adopted automatically, low-confidence mapping is audited by experts, and an auditing result is absorbed into a mapping knowledge base, so that the system has continuous learning and self-evolution capabilities, and a semantic gap of cross-mechanism data can be eliminated more intelligently and more accurately.

The application discloses a kind of multi-dimensional data intelligent matching method and system, including the following steps: collection multi-source heterogeneous data set, and execute pre-processing;Perform feature extraction, and calculate the correlation strength between features, generate feature weight matrix;Calculate structural similarity, semantic similarity and statistical similarity, and execute weighted fusion according to feature weight matrix;Nonlinear resonance optimization is executed, and stable semantic resonance spectrum is obtained;Establish reinforcement learning matching strategy network, and train matching strategy network;Match strategy network is used after training in stable semantic resonance spectrum Dynamic matching decision is executed, and matching result set is generated.The application fuses semantic resonance spectrum and reinforcement learning method, realizes the intelligent matching of multi-dimensional heterogeneous data, with the advantages of high precision, strong self-adaptation and stability.

This application discloses a method and system for consolidating evidence of network fraud through multimodal semantic consistency verification, comprising: collecting and modally identifying multimodal electronic evidence and generating original data fingerprints; extracting semantic information from each modal of evidence across five dimensions: time, amount, identity, behavior description, and geographical location; constructing a cross-modal evidence semantic association graph with a heterogeneous graph structure including evidence nodes, element nodes, and relational edges, and establishing attribution edges and cross-modal association edges based on element categories using differentiated semantic similarity calculation methods; generating initial feature vectors for each node, inputting the semantic association graph into a graph attention network model for consistency inference, performing consistency, contradiction, and uncertainty classification on cross-modal association edges, and obtaining a global consistency score; and performing differentiated on-chain consolidation operations based on the score to generate consolidation certificates. This invention can automatically discover semantic contradictions between different modalities of evidence, improving the quality of electronic evidence consolidation.

Embodiments of the present application disclose an intention recognition method and device, electronic equipment and medium, relating to the technical field of artificial intelligence. A specific embodiment of the method comprises: constructing an intention knowledge graph reflecting the association relationship between historical intention labels based on at least one historical intention label in historical dialogue text; performing semantic topic division on at least one historical intention label in the intention knowledge graph to obtain at least one historical intention label set; performing semantic similarity matching on target dialogue text based on the at least one historical intention label set to obtain at least one candidate intention label associated with the target dialogue text; and performing intention recognition on the target dialogue text based on the at least one candidate intention label to determine the target intention of the target dialogue text. The intention label system is dynamically expanded, the intention semantic association is utilized to improve the interpretability of the recognition result, and the complex customer intention recognition requirement is adapted.

The present application relates to a kind of biological image online retouching method and system supporting natural language interaction, belong to image auxiliary processing technical field.Wherein, the method includes: based on recursive data structure analysis matrix, the multi-level channel tensor is structured storage as key-value pair dictionary, and output lightweight browsing compatible data string;Multi-modal visual rendering model is constructed, and image attention weight is aggregated layer by layer, and text-image semantic correlation diagram is obtained;Based on vector database semantic matching library, the cosine similarity of high-dimensional embedding vector and candidate vector in vector database semantic matching library is calculated, and user retouching matching scheme is generated;Establish adaptive online feedback optimization mechanism, use user feedback as reward signal to adaptively adjust semantic similarity matching threshold and vector space distribution, assist in correcting semantic retrieval parameter, and convert browser end contour rendering image into interactive real-time visual output image.

A general-purpose grasping method and apparatus for retrieval enhancement in the open world can effectively balance semantic and geometric similarity, providing more reliable knowledge support for grasping point migration, thereby achieving high-precision and highly generalized zero-shot robot grasping point prediction. The method includes: (1) constructing a grasping memory bank, which integrates data from three channels; (2) determining whether the target task is a visible task, if so, executing step (3), otherwise executing step (4); (3) retrieving source domain images of the same category as the object to be grasped by category name; (4) further retrieving source domain images that are semantically and geometrically similar to the object to be grasped from the source domain image set retrieved in step (3); (5) based on the source domain images retrieved in step (4), performing the migration of grasping points from the source domain image to the target image using a diffusion model, and completing the grasping point prediction of the target image.

Systems and methods are described for comparing execution of two or more artificial intelligence (AI) AI agents. A platform can provide a user interface (UI) that allows for selection or creation of multiple AI agents. The AI agents can utilize different agent objects, such as different prompts, datasets, or models. The AI agents can be displayed on a single UI screen, where execution of the AI agents is simultaneously simulated. The same inputs can be provided to the multiple AI agents, and the corresponding outputs can display on screen. The platform can also vectorize and compare the semantic similarity of the outputs, presenting an indicating of the semantic similarity on the same UI screen.

The application discloses a kind of based on ecological environmental protection supervision big language model address resolution and geographic coding acquisition method and system.For the fuzzy, non-standard and positioning difficult problems of environmental supervision letter and visit complaint address description, the method first uses big language model to parse the original text into structured address, and cooperates multi-level fault tolerance mechanism to ensure the reliability of parsing result.In the matching stage, the system adopts local priority strategy, and preferentially searches the environmental special position database containing key polluting units and other information;If the initial matching is not hit, the self-defined part-of-speech tag segmentation technology is used to accurately extract the core environmental entities such as pollution outlet and water source protection area for secondary matching.To ensure accurate positioning, the system verifies by semantic similarity and geographical distance, to prevent errors caused by the same name in different places.When local matching fails, the system automatically accesses the network query, and uses big language model to simplify the redundant description and retry, forming a closed-loop error correction mechanism.The application significantly improves the accuracy and efficiency of pollution source positioning, protects data privacy, and provides strong technical support for environmental law enforcement personnel to achieve accurate positioning, rapid verification and scientific division of responsibility area.

The application discloses a kind of homework single topic batch re-batch method, storage medium and equipment, from the same topic of the image of the answer to the extraction of the same topic to be re-batched, and each answer image is attached Metadata label including student identity label;Each answer image is carried out semantic similarity calculation, clustering and grouping by visual semantic analysis model;According to the preset layout rule, each answer image in grouping is spliced into a composite image, and the corresponding student identity code information is embedded;For single composite image, execute once OCR identification processing to extract the answer text corresponding to all answer images, construct composite prompt word, guide large language model to return structured score array;Each score result is dynamically mapped to corresponding student individual;Update the original score record of each student's electronic homework.The application can significantly improve processing efficiency, enhance scoring consistency and explainability, consider engineering practicability and education compliance, have good expansibility and compatibility.

This application relates to a method, apparatus, computer device, and storage medium for detecting hidden encryption attacks. The method includes: performing topic analysis on input text content to determine the topic semantics of the text content; performing intent parsing on the text content to extract multiple intent information contained in the text content and obtain the intent semantics corresponding to each intent information; performing anomaly analysis based on multiple intent semantics to determine abnormal intent information; determining the semantic similarity between the intent semantics of the abnormal intent information and the topic semantics, and identifying abnormal intent information with a semantic similarity lower than a preset similarity threshold as irrelevant intent information; and performing instruction nature identification on the irrelevant intent information to determine the hidden encryption attack detection result of the text content. Using this application can improve the ability to identify malicious content with high concealment and semantic complexity.

This application discloses an intelligent question-answering method, apparatus, computer device, storage medium, and product, belonging to the field of artificial intelligence technology. It includes: displaying an intelligent question-answering interface; obtaining input question information based on input operations on the intelligent question-answering interface; performing a statement matching query in a structured database based on the question information to obtain a first query result; performing a semantic similarity query in an unstructured knowledge base based on the question information to obtain a second query result, wherein the unstructured knowledge base includes multiple unstructured knowledge items, and the second query result includes the retrieved candidate unstructured knowledge items; determining candidate answer information based on the first and second query results; and performing logical reasoning and natural language generation on the candidate answer information to obtain a first target answer information, wherein the first target answer information is used to describe the candidate answer information using natural language, and the first target answer information contains the logical reasoning process of the candidate answer information.

This application discloses a podcast generation method and related apparatus, relating to the field of artificial intelligence technology. Based on podcast configuration information, a central coordinator agent and N anchor agents are automatically created. By invoking the central coordinator agent, the N anchor agents are scheduled to speak alternately in each dialogue round according to their dynamic speaking weights, simulating the dialogue mode of a real human podcast. The dynamic speaking weights comprehensively consider time decay factors, content relevance factors, and repetition penalty factors, ensuring that anchor agents who have not spoken for a long time, and anchor agents whose potential speaking content has high semantic similarity to the current dialogue context, receive more speaking opportunities. The repetition penalty factor suppresses repetitive speaking behavior, reducing the phenomenon of viewpoint loops and content repetition in multi-round dialogues, achieving precise scheduling of multi-round dialogues, and ensuring the structured, coherent, and interactive nature of the dialogue among the N anchor agents.

The application discloses a kind of redundancy information removal methods of public opinion information, it is related to language processing technical field, the application includes for public opinion information Construction dataset;Data set constructs positive and negative examples, carries out sample enhancement;Construct pre-training model BERT, text is encoded based on pre-training model BERT, whether the similar output text is passed through average pooling layer;Positive example sample and negative example sample are input into pre-training model BERT, and pre-training model BERT is unsupervised training;The text to be identified is input into the pre-training model BERT of training, and the semantic similarity of text sentence vector is identified, if synonymous, this information is discarded, if different in meaning, the information is written into database, and redundancy information is removed;Comprehensively, the method is analyzed to the semantic similarity of the obtained public opinion data, removes the same news data, and is conducive to the efficient execution of the whole public opinion monitoring work.

The application discloses a kind of achievement transformation report intelligent generation method and system, comprising: constructing achievement transformation corpus;Based on corpus, use five-element prompt word framework to extract entity and relationship, check, implicit completion and weight update are carried out in combination with achievement transformation value evaluation model, and optimized knowledge graph is constructed;Knowledge cluster is divided to graph, and dual-mode index network of paragraph vector index and structured knowledge index is constructed;After receiving user input, core entity set is obtained by semantic coding and entity recognition;With the set as center, structured knowledge enhancement retrieval is executed, and evidence fragment is screened by fusing semantic similarity and knowledge correlation degree score;Finally, sort constructs input context, calls natural language generation model to generate structured report, simultaneously extracts reasoning path and is associated with evidence source and outputs.The application solves the problems of poor precision in manual writing, poor adaptation of general model and insufficient semantic mining, and improves the scientificity, pertinence and traceability of the report.

The application provides a semantic retrieval sorting method based on archive level weight, and related equipment. The method comprises: obtaining a natural language query sentence input by a user, and obtaining a plurality of to-be-retrieved archives, each to-be-retrieved archive comprising a plurality of level texts, the plurality of level texts comprising an archivist level text, a catalog level text, a volume level text, a file level text, and a page level text; determining a level semantic similarity between the natural language query sentence and the archivist level text, the catalog level text, the volume level text, the file level text, and the page level text of each to-be-retrieved archive; performing weighted summation processing on the level semantic similarity of each to-be-retrieved archive according to a plurality of level weights corresponding to the to-be-retrieved archive, to obtain a comprehensive semantic similarity of the to-be-retrieved archive, thereby sorting the to-be-retrieved archives to determine a semantic retrieval sorting result of the natural language query sentence. The application can improve the accuracy of retrieval result sorting.

This disclosure provides systems, methods, and devices for automatic product classification using deep learning and generative artificial intelligence (AI) models for a harmonized system (HS) product classification. A method includes generating embeddings based on an input dataset. The method includes applying a deep learning model to the embeddings to produce a prediction set including classifications corresponding to the embeddings. The method includes converting the classifications to a first set of similarity metrics. The method includes determining a second set of similarity metrics based on the embeddings using a semantic similarity model. The method includes generating a third set of similarity metrics based on an output of the semantic similarity model. The method includes outputting a ranked set of predictions corresponding to the input dataset based on the first set of similarity metrics, the second set of similarity metrics, and the third set of similarity metrics.

The application relates to a question and answer matching method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: obtaining a user question and a plurality of candidate questions corresponding to the user question; screening target candidate questions from the candidate questions; the target candidate question is a candidate question having at least one same character with the user question; for any same character corresponding to any target candidate question, determining a user word in which the same character is located in the user question, and determining a candidate word in which the same character is located in the target candidate question; in the case that the user word and the candidate word are different, adjusting the semantic similarity according to the difference between the semantic represented by the user word and the semantic represented by the candidate word; and screening a target question matched with the user question from the candidate questions according to the new semantic similarity corresponding to each target candidate question. The method can improve the accuracy of question and answer matching.

This disclosure provides a model training method, apparatus, and device for question response, which can be applied to the field of data processing technology. The method includes: performing a text generation task on an initial sample question based on prompt options to obtain an inference chain corresponding to the prompt options; scoring the inference chain to obtain a scoring result, where the scoring result represents the semantic correlation between the sample question and the sample answer in the inference chain, and the semantic similarity between the sample answer and preset sensitive words; training a scoring model based on the scoring result and the sample scoring result obtained by processing the inference chain using the scoring model to obtain a target scoring model; training a policy model based on the sample question to obtain a target policy model, where the policy model is used to process the sample question and output sample response results, and the target scoring model is used to score the sample response results to obtain sample scoring results; and training the policy model based on the sample scoring results to obtain the target policy model.

The application provides a data center computing power time sequence prediction method and device, and relates to the technical field of resource scheduling. The method comprises the following steps: based on the semantic similarity between a target variable sequence of a target data center at a current time and each historical power sequence in a vector database, retrieving semantic representations of multiple prediction power sequences; performing time sequence dependence and variable association modeling on the target variable sequence and a covariant sequence to extract time sequence dynamic feature representations; and performing weighted fusion of the semantic representations of the multiple prediction power sequences and the time sequence dynamic feature representations through dynamic gating fusion to obtain a computing power prediction value of the target data center at a future specified step. The data center computing power time sequence prediction method and device provided by the application can guarantee high precision while significantly improving the prediction robustness and generalization ability of the model in the face of unknown modes and scene mutations through deep fusion of semantic-level historical experience retrieval and deep time sequence context modeling.

The application provides a floating floor slab impact sound insulation performance optimization method, device, equipment and storage medium, keyword extraction is performed on target building space information input by a user through a natural language processing module, semantic similarity matching is performed in a pre-constructed noise source library, a target noise source spectrum corresponding to the target building space and a sound insulation standard limit value are automatically obtained, then a floating floor slab sound insulation performance target is determined based on spectrum analysis, subsequently, floating plate material parameters and elastic cushion layer parameters are substituted into a floating floor slab sound insulation performance prediction model established based on a finite element method, impact sound insulation performance prediction values under candidate material combinations are quantitatively calculated, finally, a multi-objective optimization model is constructed and solved in combination with user design constraint conditions, and an optimal design scheme of the floating floor slab matched with actual requirements of the target building space is output, so that precise matching and global optimal design of the acoustic requirements of the building space and the construction scheme of the floating floor slab are realized.

This invention relates to a scientific literature clustering method and system based on unsupervised keyword extraction. First, it effectively extracts keywords from scientific literature by comprehensively considering factors such as the occurrence of words in document abstracts and titles, the semantic similarity between words and the documents themselves, and the characteristics of domain keywords. Then, based on the characteristics of Chinese and English, this invention uses different embedding methods to cluster the extracted Chinese and English keywords, thereby achieving effective clustering of Chinese and English scientific literature. This invention considers the importance of words from multiple perspectives, comprehensively considering their occurrence in document abstracts and titles, and uses a method that automatically adjusts the preset keyword length based on domain characteristics to calculate keyword scores, thus incorporating more features of the words. This invention improves upon existing unsupervised keyword extraction algorithms.

The present application provides a communication method, comprising: receiving a first message from a second entity, the format of the first message being a first format; determining a semantic similarity between the first message and each first-format message indicated in a first mapping relationship, the first mapping relationship comprising a mapping relationship between at least one first-format message and at least one second-format message; on the basis of the semantic similarity between the first message and each first-format message indicated in the first mapping relationship, selecting a target first-format message; and, if the semantic similarity between the first message and the target first-format message is greater than or equal to a first threshold value, sending to a third entity a second-format message in the first mapping relationship corresponding to the target first-format message. The third entity can understand the first message. Thus, communication between the second entity (for example, an intelligent network element) and the third entity (for example, an access network device, a terminal device or a conventional network element) is implemented.

An electronic device and an application search method thereof are provided. The method includes the following steps. A plurality of semantic feature vectors of each of a plurality of application is recorded. The semantic feature vectors of each of the applications at least include a first semantic feature vector associated with one of application categories and a second semantic feature vector associated with one of expanded related words. A query string is obtained via the input device. A query semantic feature vector of the query string is generated by using a natural language model. A plurality of semantic similarities between the query semantic feature vector and the semantic feature vectors of each application are determined. A target application corresponding to the query string is determined from the applications according to the semantic similarities corresponding to the semantic feature vectors of each application.

Systems, methods, and computer program code for motion segmentation using a motion from structure approach; and related image processing tasks. Structure-from-motion systems are also described. In some implementations feature representations for pixels of an image frame are used to extrapolate from high reliability regions to lower reliability regions, based on semantic similarity as represented by the feature representations. This can link different visible parts of the same object, e.g. for generating a segmentation map for moving objects.

The application provides a knowledge graph question and answer reasoning method based on potential energy increment triggering, which comprises the following steps: S1, obtaining a user question, constructing a neighborhood subgraph according to the user question, generating an initial candidate operation set based on the neighborhood subgraph, and screening a candidate operation set in the initial candidate operation set according to the executability judgment and the semantic similarity ranking; S2, inputting the user question and the candidate operation set into a large language model, and outputting the selected operation by the large language model; S3, calling a retrieval execution unit, executing the selected operation output by the large language model, and outputting the corresponding entity; S4, obtaining the subject entity quality and the virtual entity quality according to the entity, extracting the user question, and calculating the semantic distance; S5, calculating the potential energy of the current step, and calculating the potential energy increment, and performing a rollback operation when the potential energy increment exceeds a preset threshold value, until the final answer is output.

The application discloses a kind of based on machine learning's intellectual property precision retrieval analysis method, it is related to artificial intelligence and information retrieval technical field, including: generating semantic vector and dimension spectrum vector to document library literature and query literature respectively;Recall candidate literature based on semantic vector and calculate semantic similarity;Based on dimension spectrum vector, calculate dimension spectrum distance, and determine the intensity of deviation of dimension semantics in combination with semantic similarity;Based on the statistical distribution of deviation intensity, calculate inhibition coefficient, calculate ranking score in combination with semantic similarity and deviation intensity and output result.The application can improve the precision and technical matching of retrieval result.

The application discloses a hyperspectral image reconstruction method based on a prior image and an auto-encoder model, presets an auto-encoder model, trains and calculates a loss, obtains a trained auto-encoder model, acquires an RGB image and a two-dimensional compressed measurement image of a target scene and pre-processes, receives the pre-processed two-dimensional compressed measurement image through a preset hardware encoding model and processes, obtains an initial three-dimensional data cube of the target scene, takes the pre-processed RGB image as a prior image, designs a regularization term based on semantic similarity of the prior image, constructs a software decoding model according to the regularization term, takes the trained auto-encoder model as a constraint term, obtains the software decoding model with the constraint term, and the software decoding model with the constraint term receives the initial three-dimensional data cube of the target scene and adopts an optimization iterative algorithm to optimize and solve, and obtains a reconstructed three-dimensional data cube of the target scene. The method helps to reduce noise and artifacts and has good reconstruction effect.

This invention provides a method, apparatus, device, medium, and program product for generating adversarial examples based on reinforcement learning, applicable to the fields of artificial intelligence and information security technology, and involving the application of large models in fintech scenarios. The method includes: performing multiple rounds of iterative mutation on a target statement based on a reinforcement learning algorithm to generate adversarial examples until at least one of the semantic similarity or fluency of the adversarial examples no longer improves; wherein, in each iteration, a first adversarial example is generated based on a target action selected from the action space of the reinforcement learning algorithm, and the first adversarial example is evaluated from at least one dimension of semantic similarity or fluency to obtain a reward value; then, after the iteration terminates, preferred adversarial examples that meet a first condition are selected from the generated adversarial examples.