145 results about "Semantics encoding" patented technology

The invention relates to the technical field of artificial intelligence, in particular to a self-organizing knowledge base construction method, system and device based on a multi-agent system and a storage medium, and the method comprises the steps: S1, monitoring a dialogue information flow, calculating the self-reply confidence coefficient of an agent, and if the self-reply confidence coefficient exceeds a preset dynamic threshold value, judging that a potential knowledge value exists and triggering an extraction request; s2, in response to the request, performing semantic coding on a dialogue fragment, extracting knowledge elements, and generating a structured knowledge unit by utilizing graph neural network modeling and relation calibration; s3, through a pre-training semantic coding model, mapping the multi-dimensional semantic coding model into a multi-dimensional semantic vector; s4, carrying out hierarchical clustering comparison and combination, and if a threshold value is exceeded, establishing a new classification and adjusting a knowledge base index; and S5, knowledge fingerprints are generated and compared, and fusion updating is carried out based on the reputation scoring model when semantics conflict or redundancy occurs. According to the method, unstructured dialogue high-precision knowledge extraction is realized, so that the knowledge base structure is dynamically self-organized according to semantics, multi-source knowledge is coordinated, and the knowledge management automation level and quality are improved.

The invention relates to the technical field of artificial intelligence dialogue systems, in particular to an intention correction and optimization method and system for multi-round dialogue context memory. According to the method, joint semantic coding is carried out on user input and historical dialogues, key semantic elements are extracted to construct an intention evolution relation graph, context consistency verification is carried out on an initial intention recognition result, and an intention correction candidate set is generated when conflicts are detected; and dynamically adjusting the context coding weight of the historical dialogue based on the corrected intention recognition result. According to the method, the intention recognition accuracy and context coherence in multiple rounds of conversations are effectively improved, and the semantic migration risk is reduced.

The invention discloses a data analysis method and system based on a large model. The method comprises the steps that natural language query of a user is received; performing semantic coding and intention analysis on the natural language query based on a pre-trained large language model; based on metadata retrieval of vector similarity, obtaining metadata related to natural language query semantics from a vector database; generating corresponding structured query parameters through the large language model in combination with the obtained metadata and the natural language query; according to the structured query parameter, generating an executable structured query language statement; and executing the structured query language statement to obtain a data query result. According to the method, the problems of insufficient semantic comprehension ability, low metadata retrieval precision and poor query generation controllability in the prior art are solved.

The invention discloses a venous thrombosis risk assessment method based on a large language model, and relates to the technical field of medical artificial intelligence, and the method comprises the steps: collecting thoracic surgery diagnosis and treatment data of a patient, carrying out the space-time alignment, generating a standard diagnosis and treatment data flow, and carrying out the homomorphic encryption of the standard diagnosis and treatment data flow, and forming an encrypted patient data package; inputting the encrypted patient data packet into a multi-task large language model, performing feature extraction and semantic coding by a feature coding layer, performing time sequence modeling and risk probability calculation by a risk quantification layer, and outputting a venous thromboembolism risk level of a patient; and performing feature decoupling and potential space mapping on the encrypted patient data packet to obtain thrombus semantic potential features. Through the multi-task large language model, the dual machine learning algorithm and the homomorphic encryption, the accuracy of venous thrombosis risk early warning is improved, and the safety of the risk assessment process is enhanced.

The invention discloses an automatic causal structure generation method based on semantic representation and logical reasoning of a large language model. The method comprises the following steps: acquiring an input text; performing semantic coding and clustering on the obtained input text by utilizing a large language model, and establishing a candidate causal variable set; causal relationship detection is carried out on the established candidate causal variable set based on anti-fact intervention and do-calculation; performing causal direction judgment, and generating a directed acyclic causal graph meeting logic consistency; and on the basis of the generated directed acyclic causal graph, natural language interpretation is generated by using a large language model, and logic consistency closed-loop verification is carried out. According to the method, automatic generation from the natural language to the causal structure is realized, the causal variable set is automatically extracted and constructed from the unstructured natural language text, the defects that variables need to be manually defined and modeling depends on field experts in the existing causal modeling process are avoided, and the labor cost and professional threshold of causal structure construction are remarkably reduced.

The invention discloses a recommendation method based on semantic enhancement and a heterogeneous hypergraph network. The recommendation method comprises the following steps that semantic information in an explicit feedback text is coded and serves as an auxiliary signal of a recommendation task; classifying the articles into predefined categories by using LLM, constructing article-category association, and mining a potential co-occurrence relationship of the articles; constructing a heterogeneous hypergraph network; spreading and aggregating hypergraph information; performing semantic alignment and model training; and performing recommendation calculation based on the final representation of the user and the representation of the article, and outputting a recommendation result. According to the method, through technical paths of semantic coding, hypergraph modeling, information spreading and alignment supervision, comment semantics of LLM coding are aligned to the recommendation space through GAE, and the problem of degradation of LLM representation in the recommendation space is effectively solved.

The invention discloses a track user association method based on semantic perception and space-time coding, relates to the technical field of location service and user behavior analysis, and aims to solve the problems of excessive dependence of POI identifiers, limited space-time representation capability, insufficient cross-city generalization capability and the like of the existing TUL method in practical application. By introducing a pre-trained large language model to carry out POI category semantic coding, multi-frequency sine space-time coding and a double-flow transfer learning mechanism, the method can effectively improve the prediction precision and the model generalization ability.

The invention discloses an aspect-level sentiment analysis optimization method based on multivariate external knowledge fusion, which relates to the technical field of sentiment analysis optimization, and comprises the following steps: constructing a multivariate external knowledge source comprising a Chinese sentiment dictionary, a domain knowledge graph and a user comment prior mode library; the sentiment module is used for providing vocabulary-level sentiment polarity, entity attribute relations and high-frequency evaluation semantic modes; semantic coding is performed on the input text and the specified aspect words to generate context semantic representation, and global semantic features and local position features are extracted in combination with aspect word position information; based on a semantic coding result, converting the multivariate external knowledge sources into structured knowledge representations, and dynamically adjusting contribution weights of various types of knowledge through a gating fusion mechanism to generate fused knowledge representations; performing dependency syntactic analysis on the input text, constructing an original syntactic structure, and calculating the correlation strength of each grammatical component and aspect words in combination with context semantic representation; and pruning the original syntactic structure according to the correlation intensity.

The invention relates to the technical field of natural language processing, and provides a multi-granularity text compression method and system for large-model multi-round dialogues, and the method comprises the steps: carrying out the semantic coding of historical question and answer pairs and a current dialogue question through a pre-training model, obtaining a semantic vector, generating a historical dialogue semantic representation through a self-attention mechanism, and carrying out the semantic representation of the historical dialogue. A semantic vector of a current dialogue problem is combined, a correlation enhancement vector is obtained through a target attention mechanism, weighted fusion is carried out on the correlation enhancement vector and historical dialogue semantic representation, a fusion vector is obtained, a feature vector is generated through a maximum pooling and average pooling combination strategy, clustering is carried out, positive and negative sample pairs are constructed, and semantic similarity loss calculation is carried out. Iteratively training the classification model; and based on the trained classification model, screening question and answer pairs related to the current dialogue question, and generating a structured abstract. Redundant content is remarkably reduced while key semantic information is reserved, and efficient context representation adaptive to large model reasoning is formed.

A method for training a speech synthesis model includes obtaining training data; obtaining an initial speech synthesis model; training a semantic encoding network and a semantic decoding network in the speech synthesis model respectively based on a style sample speech, a timbre sample speech, an input sample text, and an output sample speech in training samples of the training data, to obtain a trained speech synthesis model.

The invention discloses a similarity calculation method based on semantic editing fusion. The method comprises the following steps: calculating global semantic similarity of a source text and a target text by utilizing a semantic coding model; obtaining a first keyword list of the source text and a second keyword list of the target text through word segmentation processing, taking each first keyword in the first keyword list as a target word, and respectively forming a plurality of to-be-compared word pairs with a corresponding word in the second keyword list and a plurality of adjacent second keywords; based on the to-be-compared word pairs, calculating the semantic similarity between each group of target words and the corresponding words by utilizing a semantic coding model, and determining the editing distance between the source text and the target text; and determining the local semantic similarity of the source text and the target text according to the editing distance, and combining the global semantic similarity to obtain the final text matching similarity. According to the method, the problem that text matching only depends on character-level surface matching and neglects semantic association between word pairs is solved, and the text matching precision is improved.

The invention provides a robust Deep JSCC semantic communication system based on mutual information maximization, and relates to the field of semantic communication and joint source channel coding, and the system carries out semantic coding on to-be-transmitted source data according to a preset compression ratio and power constraint at a transmitting end to generate a channel input sequence; at a receiving end, performing semantic decoding on a noisy receiving signal transmitted by the additive white Gaussian noise channel to obtain reconstructed data and execute a downstream semantic task, and calculating a mutual information lower bound between channel input and the receiving signal by using a mutual information estimation module; a composite loss function fusing task loss, reconstruction loss and mutual information loss is constructed, and parameters of a mutual information estimator, a semantic encoder, a semantic decoder and a task execution module are alternately optimized through a joint training mechanism, so that the semantic encoder learns robust semantic representation with a self-adaptive compression ratio. The system does not need to depend on a discrete codebook, and the anti-noise capability of semantic features is improved through explicit maximization of mutual information.

The invention discloses an associated field reasoning method based on a semantic model, and the method comprises the following steps: S1, constructing a multi-dimensional semantic enhancement data set, and fusing the data set with field basic information, context features, domain knowledge and an association relationship; s2, a semantic model is constructed, the semantic model comprises a semantic fusion module, a causal semantic coding module and a symbol and semantic fusion reasoning module, according to the associated field reasoning method based on the semantic model, associated field matching can be more accurate in medical, financial, e-commerce and other scenes, and the reasoning efficiency is improved. For example, medical accurate recommendation check items, financial accurate association risk control early warning fields and e-commerce reasonable matching association commodities, deviation caused by semantic ambiguity or association non-fitting business logic is avoided; reasoning response is rapid, and real-time business requirements can be met.

The application relates to the field of artificial intelligence and digital medicine, and proposes an intention recognition model training method and device based on artificial intelligence, an electronic device and a storage medium. The intention recognition model training method based on artificial intelligence comprises the following steps: performing word embedding on a plurality of natural sentences collected in advance to obtain a sentence vector of each natural sentence; identifying an intention vector and a slot vector of the sentence vector by using a preset semantic coding model; labeling the intention vector and the slot vector to construct a training data set; constructing an initial intention recognition model, training the initial intention recognition model by using the training data set, and obtaining an intention recognition model. The method can jointly train the intention recognition model by using the intention vector and the slot vector of the natural sentence, so that the accuracy of the intention recognition model can be improved.

The invention discloses a semantic fusion detection method and device for short circuit, open circuit and open circuit defects, and the method comprises the steps: extracting a candidate region from a to-be-detected image, carrying out the multi-scale cutting of the candidate region, generating a plurality of input images of different scales, and carrying out the semantic coding of the input images, and generating a semantic vector; obtaining a cue word set containing short circuit, open circuit and open circuit defect form description, matching the semantic vector with the cue word set, and calculating a semantic score of each defect type; selecting a short circuit expert, an open circuit expert and an open circuit expert to execute corresponding reasoning tasks according to the semantic vector and the context information, and outputting diagnosis results and suggestions; and calculating measurement features related to connectivity for the candidate region, performing fusion arbitration on the measurement features, semantic scores and expert output, and outputting defect types, confidence scores and re-judgment layering. According to the scheme, the precision and reliability of short circuit, open circuit and open circuit defect detection are improved.

This application provides an adaptive semantic joint source-channel coding method, system, electronic device, and storage medium. The method includes: Step S1: acquiring raw input data and real-time channel state information, wherein the real-time channel state information includes at least the signal-to-noise ratio (SNR); Step S2: using a lightweight semantic coding module to extract multi-scale semantic features from the raw input data, and incorporating the SNR embedding vector in a feature modulation manner during the coding process to obtain channel-adaptive semantic features; Step S3: using an SNR embedding and channel-adaptive attention module to adjust the attention weights of the semantic features to obtain attention features; Step S4: using a dynamic codebook generation and vector quantization module to convert the attention features into a discrete codebook index sequence; Step S5: using a joint source-channel coding module to modulate and map the codebook index sequence into complex channel symbols and transmit them.

The application discloses an attention mechanism-based adversarial text defense method and system, which comprises the following steps: inputting to-be-recognized text into a natural language processing model with an encoder-decoder as a basic structure; calculating the importance score of each word in the text by using a word scoring function; taking the reciprocal of the importance score to form a reconstruction score vector; calculating the weight of each hidden layer vector according to an attention formula to obtain an attention weight vector; balancing the reconstruction score vector and the attention weight vector by using the way of multiplying the reconstruction score vector by a hyperparameter, multiplying corresponding elements in the reconstruction score vector and the attention weight vector one by one to obtain a final reconstruction attention vector; and multiplying the reconstruction attention vector by a hidden layer feature vector to obtain a reconstructed semantic code, and obtaining output after decoding. The application has good generalization performance, does not need to retrain a model when coping with new adversarial attacks, and has a certain effect on character-level adversarial attacks and word-level adversarial attacks.

The invention discloses a short entity recall model construction method and device, a short entity recall model comprising an encoder and a decoder is constructed based on a pre-training language model, the encoder is responsible for inputting semantic codes of a context, the decoder generates entity names in an autoregression mode, and the model parameter scale is related to the size of a vocabulary; after obtaining training data containing an input context and a corresponding target entity name, inputting the training data into the model, and training the model in an autoregression mode to learn a mapping relationship between the input context and the target entity name; configuring a limited decoding strategy to ensure that an entity name generated during model reasoning belongs to a preset candidate entity set, and obtaining a trained model; and inputting the context to be retrieved into the model, generating a corresponding entity name through the model, and completing short entity recall.

The application discloses an intelligent terminal semantic understanding optimization method based on deep learning, and relates to the technical field of intelligent terminal semantic understanding. When a current short instruction is received, the method reads multi-source interface events in a preset time window and constructs a short-time interface state transition graph. Based on the short-time interface state transition graph, a local causal chain is extracted and an interface attribution clue sequence is formed. Double-granularity semantic coding and slot constraint decoding are performed on the current short instruction text, and cross alignment is performed with the interface attribution clue sequence to generate a task attribution candidate result. Path backtracking and consistency checking are then performed to determine a target attribution result. Finally, a call entry is determined according to the target attribution result and written into a short-range semantic memory cache. The application is suitable for multi-application switching, page jumping and short instruction continuous input scenes.

The invention belongs to the field of artificial intelligence and recommendation systems, and discloses a two-channel recommendation method and system based on graph structure perception and representation alignment, and the method comprises the steps: obtaining the text data of a user article and the interaction data of the user article; constructing a user article interaction graph based on the interaction data, and extracting collaborative embedding from the interaction graph; performing multi-layer perception semantic coding on the text attribute data, and injecting structure signals of local neighbors into a semantic space to obtain semantic embedding; compressing semantic embedding and collaborative embedding into an indexable discrete potential space, and generating index representation; performing weighted fusion on the collaborative embedding, the semantic embedding and the index representation to generate a unified fusion representation; and based on the unified fusion representation, taking a learnable prompt vector as a task condition signal, guiding the model to complete a plurality of recommendation tasks, and outputting a recommendation result. According to the method, embedding degradation and semantic drift caused by data sparsity are effectively relieved, and the robustness and generalization ability of the model in a cold start scene are remarkably improved.

This invention discloses a method, apparatus, device, and storage medium for detecting SQL injection attacks. The method includes the following steps: based on an abstract syntax tree generated by parsing a received SQL statement, semantically and syntactically encoding the nodes in the abstract syntax tree to obtain semantic feature vectors and syntactic feature vectors for each node; concatenating and fusing the semantic feature vectors and syntactic feature vectors to generate initial features for each node; based on a constructed sparse heterogeneous graph, using a graph convolutional network to propagate the initial features between each node and its neighboring nodes to obtain updated features for each node; performing global average pooling on the updated features of the nodes to obtain a global representation vector of the SQL statement, and inputting the global representation vector into a classifier to output the SQL injection detection result. This application can significantly reduce the computational resource requirements of the model while maintaining high detection accuracy.

The application discloses a semantic communication decoding method based on space-based computing, comprising the following steps: receiving a semantic encoding signal transmitted through a channel, and generating an initial decoding distribution through a non-autoregressive semantic decoder; constructing an empty information observation adapting to the current channel characteristics, inputting the observation into the decoder to estimate an implicit language prior, and correcting the initial decoding distribution to obtain a corrected channel likelihood distribution; combining the semantic prior of a pre-trained language model, quantifying the channel uncertainty and semantic uncertainty of each word position respectively, and generating a refinement priority based on the source ratio of the channel uncertainty in the total uncertainty; and iteratively refining the text sequence according to the priority. The application can explicitly eliminate the internal prior bias of a large model by using an empty observation, suppress the illusion phenomenon under a low signal-to-noise ratio, and realize the on-demand allocation of computing resources by distinguishing the uncertainty sources, thereby improving the accuracy and efficiency of decoding.

Embodiments of the present application provide a text processing method, device and computer storage medium. The method comprises: obtaining a text to be processed; calling a first classification network to process a secondary category code representation, category matching parameters and semantic code representation of the text to be processed, to obtain classification information of a secondary category of the text to be processed; and determining a classification result of the text to be processed based on the classification information of the secondary category and classification information of a primary category of the text to be processed, the classification information of the primary category of the text to be processed being obtained by processing the text to be processed by a second classification network, so that the accuracy of hierarchical classification of the text can be improved.

The application provides a risk event identification method and system based on monitoring video, and relates to the technical field of data analysis.In the application, first, semantic coding is performed on a target monitoring video to form a global video coding vector, and the global video coding vector is associatedly coded based on behavior-related semantics to form a behavior video coding vector;second, semantic coding is performed on target non-visible domain data to form a global non-visible domain coding vector, and the global non-visible domain coding vector is associatedly coded based on behavior-related semantics to form a behavior non-visible domain coding vector;then, the behavior video coding vector is constrainedly coded based on the behavior non-visible domain coding vector to form a behavior constraint coding vector;finally, the behavior constraint coding vector is semantically decoded to form a risk event identification result.Based on the above method, the problem that the reliability of risk event identification is relatively low in the prior art can be improved.

The application discloses a power transaction knowledge extraction method based on semantic coding and feature enhancement, comprising the following steps: cleaning and term standardization on unstructured text in the power transaction field to obtain preprocessed text; using a pre-training language model based on an error correction mask mechanism to code semantic features of the preprocessed text to obtain a semantic coding sequence; according to the semantic coding sequence, sequentially performing bidirectional long-distance semantic dependence capture, dynamic attention weight distribution and local semantic detail fusion to obtain an enhanced feature sequence; based on a layered pointer network, performing entity and relation joint extraction on the enhanced feature sequence to obtain a triple composed of a head entity, a relation and a tail entity; and performing deduplication and completion processing on the triple to complete the knowledge extraction of the power transaction. The application realizes the goal of automatically and high-quality extracting structured knowledge of the power transaction from unstructured text.

A method of channel-aware semantic coding (CASC) by a user equipment (UE) includes determining a quality level of a channel during a time period in which a video frame is transmitted over the channel, determining one or more semantic elements to be included in a semantic tile stream (STS) based on the quality level, encoding the video frame with the one or more elements of the STS, and transmitting the STS to a remote device.

This invention discloses an enterprise intelligent knowledge base management system and method based on a vector model, relating to the fields of intelligent information retrieval and natural language processing. This invention parses unstructured documents in various enterprise formats, extracting semantic structure information such as text heading levels and paragraph boundaries, and segmenting them into a set of text fragments. The text fragments are then cleaned to obtain effective text fragments. These effective text fragments are integrated into semantic text blocks with the same topic. Sentence vectors are generated using a pre-trained semantic encoding model. Semantic segmentation points are located based on the semantic similarity of adjacent sentences to obtain standardized knowledge units. A dual-database collaborative query is constructed. After receiving user queries, structured pre-filtering is performed through a relational database, followed by semantic retrieval through a vector database. The intersection of the results from both databases is taken, and a comprehensive score is calculated based on multi-dimensional weights. After filtering, aggregation, and tracing, the results are output. If no matching results are found, a tiered fallback process is triggered, completing the entire retrieval loop.

This invention discloses an industrial fault diagnosis method based on knowledge graph causal chains and a large language model, comprising: (1) a knowledge graph creation step; (2) a key semantic extraction step; (3) an entity recall step, obtaining semantic vector u and entity vector v for comparison, sorting candidate entities according to cosine similarity scores, and selecting the top N as candidate entities, where N is a positive integer; and (4) a large language model decision step. This invention's industrial fault diagnosis method based on the collaboration of knowledge graph causal chains and a large language model, by collaborating the knowledge graph causal chains and the large language model, allows the large language model to address issues such as colloquialisms in on-site questions and insufficient consistency and interpretability caused by ambiguous terminology. Utilizing the advantages of SBERT semantic encoding and fast vector retrieval, it is suitable for rapid filtering from massive amounts of data, significantly improving recall efficiency and meeting the real-time requirements of large-scale data scenarios.

This invention provides a script generation method, apparatus, electronic device, and readable storage medium. The method includes: acquiring at least one initial script idea element described in natural language; inputting the initial script idea element into a pre-trained semantic encoding model, converting the initial script idea element into a first idea vector in a d-dimensional real vector space, where d is a positive integer; wherein all real vectors output by the semantic encoding model together constitute a script idea latent space for representing the semantic distribution relationship of the initial script idea element; performing spatial operations on at least one first idea vector within the script idea latent space to obtain a second idea vector, the second idea vector being used to represent a script concept; inputting the second idea vector into a cross-modal projector, converting the second idea vector into a conditional vector matching the word embedding dimension of a pre-trained large language model; inputting the conditional vector as a generation condition into the large language model, outputting script idea text corresponding to the script concept, the script idea text including at least one of the following: story synopsis, character biographies, and key scene description information.

The application belongs to the technical field of deep learning and knowledge graph completion, and relates to a few-shot knowledge graph completion model and method, wherein the completion model is divided into three modules: (1) LLM-PCA semantic encoder: responsible for extracting the text semantic features of entities and relations in the knowledge graph triplets, and reducing the dimension of the extracted semantic feature vectors to obtain effective low-dimensional semantic representation of the knowledge graph elements; (2) attention mapping network: responsible for fusing the initial semantic encoding vectors of the triplets and mapping the text semantic representation vectors of the triplets under the same relation to a unified semantic space, combining the structural information with the text semantics; (3) authenticity scoring module: responsible for judging the possibility of the real existence of the to-be-queried triplets. The method can not only effectively handle complex relations and sparse data, but also realize high-quality knowledge graph completion under the condition of a small amount of training samples, and significantly improves the accuracy and generalization ability of the completion.