485 results about "Word embedding" patented technology

The invention discloses a retrieval generation method and system based on multi-agent collaboration, a terminal and a medium, and relates to the field of artificial intelligence. Performing semantic analysis on the input word embedding converted by the natural language query instruction through a query analysis agent, and determining a semantic intention vector; performing reinforcement learning and meta learning on the semantic intention vector through a strategy construction agent, and determining a retrieval strategy; performing semantic enhancement on the semantic intention vector according to knowledge graph node embedding to obtain a semantic enhancement vector; determining a data channel according to the semantic enhancement vector, a retrieval strategy and a real-time system load, and calling the data channel for retrieval to obtain candidate documents; and generating a target answer according to each candidate document based on an adaptive reflection feedback mechanism in combination with an auto-encoder and a generative adversarial network. The problems that the prior art depends on a fixed retrieval strategy, has limitation when facing complex query, multi-round interaction and cross-modal data fusion, is easily interfered by noise and is not accurate enough in semantic matching are effectively solved.

The invention provides a financial fraud detection method based on a large language model. The method comprises the steps of obtaining a to-be-recognized text; performing word segmentation on the to-be-recognized text through the target word segmentation tool and the financial fraud dictionary, and determining a fraud sensitive word list; calculating the weight of each sensitive word in the sensitive word list according to a target algorithm to obtain a sensitive word weight feature vector; inputting the sensitive word weight feature vector and a to-be-recognized text into a large language model, and determining a context semantic vector of the sensitive word in combination with a word embedding technology; determining the similarity between the context semantic vector of the sensitive word and a preset financial fraud type semantic vector; and determining a financial fraud type according to the similarity. Through the implementation of the method, the generalization ability of the pre-training model is utilized to capture text deep semantics, priori knowledge is injected in combination with a sensitive word weight mechanism, the model is guided to focus high-risk vocabularies, the defect of a traditional method in semantic comprehension is overcome, the financial fraud recognition rate is increased, and the omission ratio is reduced.

The invention relates to the technical field of power marketing, in particular to an AI large model-based power marketing data analysis method, which comprises the following steps of: acquiring structured data and unstructured data in a power marketing system, and generating unified coded data after space-time alignment and pre-training word embedding model processing; inputting the unified coding data into a pre-trained power field large model, and extracting static, dynamic and semantic features through multi-modal fusion, feature decoupling and semantic anchoring; constructing entity link feature pairs in combination with a power knowledge graph, and enhancing fusion feature expression through a graph attention mechanism; and finally, inputting a dynamic weight gating network, outputting an abnormal user identification tag, a demand response strategy and a customer loss early warning probability, and executing strategy optimization under specific conditions. The method can be widely applied to risk identification, strategy making and user behavior prediction tasks in power marketing.

The invention relates to a retrieval method and device based on static word embedding, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of performing word segmentation on an original text to obtain a first word segmentation result, training a static word embedding model by utilizing the first word segmentation result to obtain word vectors, and generating a synonym word library; the method comprises the following steps: establishing a full-text inverted index by utilizing an original text, and expanding query words by utilizing a synonym library in a retrieval stage; encoding the original text into a semantic vector by using a semantic generation model, and constructing a vector index based on the semantic vector; based on a to-be-queried text in a user query request, performing retrieval by using the full-text inverted index to obtain a first candidate document, and performing retrieval by using the vector index to obtain a second candidate document; performing fusion processing on the first candidate document and the second candidate document to obtain a target candidate document; and inputting the target candidate document into the text generation model to obtain a retrieval result. By adopting the method, the accuracy of text retrieval can be improved.

The invention provides a prediction method and device for operation parameters of a virtual power plant, a storage medium and a computer program product. Comprising the steps of performing data preprocessing on acquired influence factor data; using a time sequence decomposition algorithm to decompose the influence factor data into three-dimensional sequences including a trend component, a season component and a residual component, and segmenting each sequence into sequence blocks and mapping the sequence blocks into time sequence feature vectors; performing K-means clustering on word embedding used for pre-training the large language model, and selecting K clustering centers as semantic anchor points to be spliced with the time sequence feature vectors; finely adjusting the position embedding parameters of the large language model, the weight of the feedforward neural network in residual connection and the parameters of the normalization layer; and according to the semantic enhancement time sequence feature vector, predicting to obtain a trend component, a season component and a residual component of each dimension, and carrying out splicing and reverse normalization processing on the components to obtain a prediction result. According to the prediction method, the accuracy and real-time performance of virtual power plant load and electricity price prediction are improved.

A method, computer program product, and computing system for generating a plurality of chunks for a plurality of text portions of a document, wherein the document includes the plurality of text portions and a plurality of images. Each chunk is indexed using a word embedding. Each of the plurality of images is indexed based upon, at least in part, a position of a respective image relative to a corresponding chunk. An image placeholder is generated for each of the plurality of images. A plurality of image-enhanced embeddings is generated by inserting the image placeholder for each of the plurality of images into a respective word embedding for the corresponding chunk. The plurality of image-enhanced embeddings are provided for processing a query using a generative artificial intelligence (AI) model.

The invention provides a full-text retrieval method and system fusing various types of documents, and relates to the technical field of information retrieval, and the method comprises the following steps: obtaining document representation through document content extraction and structure recognition, generating a cross-modal semantic vector by using word embedding and nonlinear transformation, constructing a hierarchical index and a cross-document association graph, and obtaining a full-text retrieval result; the basic correlation score is calculated after the query request is received, and the comprehensive score of the candidate content segments is calculated based on the association graph to determine the optimal retrieval result, so that unified representation and retrieval of heterogeneous documents are realized, the cross-document retrieval precision and relevance are improved, and the processing capability of a retrieval system on complex queries is enhanced.

The invention relates to an enhanced question answering method based on agricultural large model training and RAG, and belongs to the field of agricultural large data, the method comprises the following steps: collecting agricultural data, and constructing an agricultural large model training data set; the method comprises the following steps of: constructing input embedding of a Transform and a Transform encoder to train a large model based on a Transform framework; after a user inputs a question text, the large model sequentially executes dynamic word embedding disambiguation, regional term replacement, entity perception position coding and generation of enhanced representation of an input sequence, then, a Transformer encoder extracts deep semantic features, a classification task directly outputs prediction labels, and a generation task generates answers word by word through a decoder; and performing RAG enhanced questioning and answering in combination with a retrieval enhanced generation mechanism. According to the dynamic word embedding mechanism, the ambiguity problem of agricultural terms in different contexts and regions is effectively solved, and the accuracy of agricultural semantic understanding is remarkably improved.

The invention provides a global and local feature fusion SMPL-X action-to-text generation method, and belongs to the field of artificial intelligence. The method comprises the following steps: preprocessing and coding an input SMPL-X action sequence into a double-flow action feature; through a cross-modal mapping module, the double-flow action features are mapped to a pre-trained large language model through independent projection branches, and global conditions and local action prefix embedding are obtained; through a text generation module, a decoder of a pre-trained large language model is used as a trunk network, text cue word embedding is extracted based on a text instruction given by a user, local action prefix embedding and text cue word embedding are spliced and then input into the decoder, and global conditions are injected into each layer of the decoder through a cross attention mechanism. And generating a description text in an autoregression mode. According to the method, the description text which is consistent with action semantics and has sufficient details can be stably and accurately generated, and the generation stability and the cross-scene applicability are improved when disturbance exists in the action sequence.

The invention discloses a knowledge data de-duplication method and device, a storage medium and computer equipment, relates to the technical field of data processing, is suitable for businesses such as financial science and technology and smart medical treatment, and mainly aims to solve the problem of poor de-duplication effect when large-scale knowledge data is processed in existing knowledge data de-duplication. Comprising the steps of obtaining large-scale knowledge data of related businesses; performing clustering processing on the large-scale knowledge data by adopting a MinHash LSH model to obtain a repeated text clustering result; the repeated text clustering result comprises a plurality of groups of similar data sets; performing word embedding calculation on each group of similar data sets by adopting a bge-m3 model to obtain word embedding corresponding to each group of similar data sets; and respectively carrying out semantic similarity de-duplication processing on the word embedding in each group of similar data sets to obtain a de-duplication result of the knowledge data.

The invention relates to the field of intelligent report generation, in particular to an intelligent insight report generation method and system based on an AI large model, and the method comprises the steps: inputting an insight demand, and generating an insight data package comprising insight contents, associated data and industry labels; extracting a basic keyword set of the insight data packet to form a mixed feature code; after mixed feature coding preprocessing, weight distribution is carried out; and after weight distribution of the AI large model, injecting a high-weight feature vector into a semantic understanding core layer, injecting a low-weight feature vector into a logical reasoning layer, and outputting analysis data to form an intelligent insight report. According to the method, word embedding parameters are optimized according to field semantic characteristics, a parameter verification mechanism is introduced, field text characteristics are adapted by adjusting vector dimensions, context windows and low-frequency vocabulary filtering threshold values, window parameter validity is verified through cosine similarity, word frequency threshold value reasonability is verified through standard deviation, and field text characteristic matching is achieved. And it is ensured that the feature vectors can accurately capture domain-specific semantics.

The embodiment of the invention provides an address data matching method and related equipment, and belongs to the technical field of geographic information services. The method comprises the following steps: constructing an address annotation corpus according to input address information data and a preset address database; the method comprises the following steps: generating a geographic information embedding vector according to a preset geographic information knowledge graph, performing address element analysis in combination with an address annotation corpus to obtain an address element sequence so as to construct a dictionary tree, and performing similarity screening through a spatial hierarchical matching algorithm to obtain a similar address set; generating an address embedding vector matrix through a preset word embedding vector model, and performing feature extraction through a preset semantic feature extraction model to obtain semantic-level similar features; according to input address information data, multi-dimensional character similarity matching is carried out to obtain character-level similar features, then weighted fusion is carried out in combination with semantic-level similar features, and target matching address data is determined according to a weighted fusion result. According to the embodiment of the invention, the address data matching accuracy and efficiency can be improved.

The invention provides a large model output content illusion detection method and system oriented to a zero knowledge scene, and belongs to the field of large model security application, and the method comprises the following steps: S1, extracting internal state features and output probability features from a large language model LLMs, the internal state features including full context average embedding ACE and end word embedding FTE; the output probability features comprise word probability TP entropy En; s2, inputting the features ACE, FTE, TP and En into a classification detection module for training; a trained classification detection module is obtained; and S3, when the LLMs receives an input prompt and starts to generate a text, extracting features ACE, FTE, TP and En in real time, inputting the features into the trained classification detection module, and judging whether the content of the text is illusion or not. According to the method, external knowledge sources are not needed, and illusion detection is efficiently and accurately carried out by combining internal and external features.

The invention discloses an optical metasurface structure parameter design model training and application method and a related device, and relates to the technical field of nanometer photonics nanometer structure design, the method comprises the following steps: obtaining a data set, the data set comprises multiple groups of sample data of each optical metasurface in a plurality of optical metasurfaces, the sample data comprises a sample structure name, a sample structure parameter and a sample spectral response, word embedding and position code embedding are respectively carried out on the sample structure name and the sample structure parameter to obtain an embedded sample structure name and an embedded sample structure parameter, a feature extractor is further utilized to extract and obtain a fusion feature, and the fusion feature is extracted to obtain a fusion result; and taking the fusion features as input, generating an embedded sample structure parameter predicted value by using a structure parameter generator, subsequently calculating total loss, and performing iterative updating based on the total loss until training is completed. The method can be suitable for design of structure parameters of optical metasurface of different structure types and different spectral targets, and generalization ability is improved.

The present invention facilitates communication between sign language users and machines by translating sign language and text using AI models, deep learning computer vision, and word embeddings. Users interact via sign language, captured and processed through deep learning and NLP modules. The system converts sign language videos into text, constructs coherent sentences, and generates contextually appropriate responses using a Retrieve and Generate (RAG) model. Responses are translated back into sign language videos, spelling out words not found in the dictionary. If requested, a human agent can respond. Key features include high-accuracy recognition, context-aware response generation, dynamic vocabulary updates, and optional human interaction. The method ensures efficient processing with LLM, embedding techniques, and deep learning, optimizing translation accuracy and user experience. The system adapts to multiple languages and dialects by training on specific sign languages, making it applicable globally.

An information learning apparatus includes a memory and a processor configured to perform generating, for each of processing units constituting an input sequence included in training data, a third embedded vector based on a first embedded vector for the processing unit and a second embedded vector corresponding to an unknown word; executing a process based on a learning target parameter, with the third embedded vector generated for each of the processing units as an input; and learning, for a processing result by the executing, the parameter based on an error of an output corresponding to the input sequence in the training data.

The invention belongs to the technical field of anomaly detection, and particularly relates to an abnormal sample detection system based on a diffusion model. The system comprises a word embedding vector training module which applies a set number of abnormal samples, initializes the abnormal samples into a word embedding vector, and inputs the word embedding vector into a diffusion model for guiding the diffusion model to learn; the abnormal sample generation module is used for generating an abnormal sample by combining the word embedding vector trained by the word embedding vector training module with the abnormal mask, and inputting the abnormal sample into the abnormal reconstruction module; the anomaly reconstruction module is used for carrying out image reconstruction on the input anomaly sample, carrying out loss calculation on the reconstructed image and the corresponding normal sample to train a model, splicing the reconstructed image and the anomaly sample and then inputting the spliced image and anomaly sample into the anomaly judgment module; and the abnormity judgment module is used for generating a model, speculating an abnormal mask, and comparing the abnormal mask with an actual abnormal mask to calculate loss.

The application discloses an application method of a Transformer architecture based on a biological regulation mechanism, S1, original data is divided into basic units token, and then each discrete token is mapped into a continuous high-dimensional vector through a word embedding layer; S2, a prototype activation unit identifies an initial semantic core from the input embedded token, and scores an activation value based on the embedded vector of the token; S3, a multi-scale prototype aggregation module performs multi-scale pattern aggregation on the prototype candidate, and forms a prototype vector across tokens; S4, a regulation path diverter determines a main / auxiliary path distribution strategy of each token according to the context semantic weight and the prototype responsiveness of the token; S5, the main path performs a standard QKV calculation process; S6, a biological regulation loop module performs feedback regulation on the auxiliary path token by using the aggregated prototype vector in the S3 step; S7, two token expression streams generated by the main path and the auxiliary path are integrated, so that the final output of the Transformer layer is generated.

The invention provides a quantization method and a quantization device, which can perform singular value decomposition on a first matrix formed by word embedding vectors to obtain the maximum characteristic value of each block matrix in M block matrixes, and can determine the quantity of quantization bits of each block matrix according to the maximum characteristic value of each block matrix. The method comprises the following steps of: dividing each block matrix into a plurality of block matrixes, quantifying each block matrix to obtain each quantized block matrix, combining each quantized block matrix to obtain a quantized first matrix, and representing the amount of semantic information carried in each block matrix by a maximum characteristic value of each block matrix to a certain extent, according to the embodiment of the invention, the quantized bit number can be allocated to each block matrix according to the amount of semantic information carried by each block matrix, so that different bit numbers can be allocated to the block matrixes carrying different semantic information, differential bit number allocation can be realized, semantic information loss can be reduced, and user experience can be improved.

The present invention relates generally to computer systems based on specific computational models, in particular to language reasoning servers, methods for language reasoning, systems, media, and products for visual language big model reasoning. A language reasoning server comprises computing resources, a large language model, a cache pool and a task scheduling component. The task scheduling component, in response to detecting that the large language model needs to perform a first portion of language reasoning associated with the first word embedding and the language element, checks the cache mapping dictionary based on the first request identifier, selectively loads the first word embedding from the cache pool based on a result of the checking, and selectively loads the second word embedding from the cache pool based on the second request identifier. And selectively instructing the large language model to invoke the computing resource to perform a first portion of the language reasoning task. Wherein execution of the first portion by the large language model is at least partially contemporaneous with execution of the visual coding calculation by the visual server. Some embodiments improve system throughput and reduce system latency.

The present invention relates to the field of remote sensing image change detection, and discloses a remote sensing image change detection method based on language guidance. Efficient SAM is used as a visual basic model to extract semantic feature maps of remote sensing images in two phases, and a multi-layer adapter is constructed to extract multi-scale feature maps in two phases; spatial attention is constructed to extract visual features; the input text description is processed by a text encoder to obtain word embedding and overall expression embedding; a multi-head cross attention method is used to construct a pre-fusion module for the fusion operation of word embedding and spliced ​​multi-layer feature maps; a dual-channel decoder structure is constructed to calculate the similarity between the obtained image features and the semantic embedding, and output the final change detection mask. The present invention designs a multi-layer visual-language feature cross layer to enable the model to fully learn the relevant information between the image and the corresponding type of text, thereby realizing universal detection of changes in multi-category remote sensing images guided by natural language.

The invention relates to the technical field of artificial intelligence dialogue systems, and particularly discloses an artificial intelligence dialogue generation method based on natural language processing. According to the method, response certainty or diversity is adaptively adjusted according to a dialogue scene through a dynamic temperature sampling strategy, and historical dialogue key features are screened in combination with a gating attention mechanism to realize accurate semantic fusion; word embedding and primary coding are migrated to terminal equipment to be executed by adopting an edge-cloud collaborative architecture, and are transmitted to a cloud end through feature compression and encryption to complete deep decoding; establishing a dual-channel sensitive word detection mechanism of input regular matching and named entity recognition, and blocking privacy leakage through low-temperature sampling and risk word filtering in an output stage; and generating a four-dimensional metadata label driving decision containing the dialogue behavior type, the emotion polarity, the confidence coefficient and the interpretable vector. The method improves the generation quality in the algorithm layer, optimizes the deployment efficiency in the system layer, enhances the security and interpretability in the application layer, and is suitable for intelligent customer service, virtual assistant and other scenes.

This invention provides a method and apparatus for trajectory prediction based on a pre-trained large language model, relating to the field of artificial intelligence. The method includes: acquiring historical trajectory data of a target aircraft and performing data preprocessing to obtain a trajectory sequence; constructing prompt words and inputting the prompt words into a pre-trained large language model to obtain word embedding representations; the prompt words contain a description of the trajectory prediction task and a description of the trajectory data; encoding the trajectory sequence to obtain an encoded trajectory sequence and a trajectory position sequence, and segmenting the trajectory position sequence into words; constructing a spatiotemporal graph of the trajectory sequence based on the encoded trajectory sequence and extracting spatiotemporal dependency feature representations of the trajectory sequence; fusing the word embedding representations and the spatiotemporal dependency feature representations using the pre-trained large language model to generate a trajectory information feature representation; and inputting the trajectory information feature representations into a trained prediction module to obtain predicted values ​​of the longitude, latitude, and altitude of the target aircraft at future times.

A method initializes and accelerates training of neural network based large language model, including by: (i) accessing a corpora for training a neural-network based large language model having word embeddings and word projections in respective word embedding and word projection layers and at least one hidden layer; (ii) counting raw token frequencies associated with content within the corpora; (iii) smoothing the raw token frequencies into a series of vector norms based on log or scaled log functions parameterized by maximum norm information; and (iv) injecting vector norm information into word embeddings and / or word projections based on norm-angle reparameterization to prepare the large language model for training.

The invention provides an embedded layer access collaborative acceleration system and method of a neural network model, a processing unit layer, a control layer and a routing layer cooperatively work, and the control layer can predict a candidate item of a next lexical element before a processing unit by querying a lexical element statistical table. A routing layer reads word embedding vectors of candidate lexical elements from a memory bank in advance and stores the word embedding vectors in a prefetching buffer area corresponding to the memory bank, and after a processing unit inferes a next lexical element by using a neural network model, if the next lexical element hits the candidate lexical element, the next lexical element is not hit by the candidate lexical element. If not, the word embedding vector of the hit candidate lexical element is read from the prefetching buffer area, and compared with the mode that the processing unit inferes the next lexical element and then reads the corresponding word embedding vector from the storage bank, lexical element prefetching processing is increased, and the access speed of an embedding layer of the neural network model can be increased.

The invention discloses a scene perception non-intrusive load monitoring method based on a large language model. The method comprises the following steps: firstly, acquiring a total load sequence, segmenting a total electrical load sequence into a plurality of total load segments, and performing vector embedding to obtain a total load segment embedding vector; thirdly, constructing a comprehensive cue word and carrying out vector embedding to obtain a comprehensive cue word embedding vector; then, splicing the total load fragment embedding vector and the comprehensive cue word embedding vector to obtain a total load and comprehensive cue word fusion embedding vector, inputting the total load and comprehensive cue word fusion embedding vector into a trained large language model, and outputting a load estimation vector of each target device; mapping the load estimation vector of each target device through a linear transformation layer to obtain a load estimation value of each target device; and finally, correcting the load estimation value by using the load probability distribution of the target equipment in the current operation state. The power consumption scene information and priori knowledge are fully utilized, and the load monitoring accuracy of the model under the complex power consumption condition is remarkably improved.

The invention discloses a big and small model collaborative natural language processing method, system and device and a medium, which are applied to the field of language processing, and the method comprises the following steps: performing data preprocessing on to-be-processed text data to obtain word embedding vector data; obtaining the semantic complexity of the word embedding vector data and the input text length of the to-be-processed text data, and calculating the task complexity according to the semantic complexity and the input text length; splitting the text processing task into a plurality of sub-tasks according to the task complexity; according to a preset task complexity threshold, dynamically allocating each sub-task to the large model and the small model for processing to obtain a text reasoning result of each sub-task; and fusing the text reasoning results to obtain a text processing result. According to the method, by dynamically distributing the tasks to the large and small models and fusing the reasoning results of the large and small models, the reasoning efficiency and precision are effectively balanced, the accuracy and the real-time performance of the natural language processing tasks are improved, and meanwhile the dynamic adaptability of the system is enhanced.

The invention relates to an assembly type bridge modeling method based on a large model and an RAG technology. The method comprises the steps that a bridge standard component library is established based on an assembly type bridge standard image set; calling a large language model, and generating standardized knowledge statements based on the bridge standard component library; a word embedding method is adopted to convert the normalized knowledge statement into vectors, and the vectors are stored in a vector database; converting an input natural statement query into a vector by adopting a word embedding method, and retrieving in a vector database to obtain a normalized knowledge statement related to the query; performing parameter analysis and optimization on the retrieved knowledge statements based on a large language model by utilizing an RAG technology so as to generate component parameters; and calling a component modeling algorithm based on each component parameter to generate a component three-dimensional model, and assembling the component three-dimensional model to obtain a fabricated bridge model. According to the method, component parameter sorting, knowledge statement intelligent combination, efficient semantic retrieval, parameter optimization, three-dimensional model automatic generation and the like are organically integrated, and the modeling efficiency and quality are improved.

The invention discloses an automatic file analysis method based on a semantic evolution graph. The method comprises the following steps: generating a policy document standardized corpus; obtaining a policy document structured corpus unit set; constructing an initial semantic map snapshot containing a version stamp; establishing a time slice semantic evolution graph sequence; obtaining a multi-time sequence fusion vector; generating a multi-time sequence optimization word embedding vector matrix; performing time continuous regularization on the multi-time-sequence optimization word embedding vector matrix, outputting time continuous word embedding vectors, and mapping the time continuous word embedding vectors to nodes and edges of the time slice semantic evolution graph sequence to form a global semantic evolution graph fusion representation sequence; and taking the global semantic evolution graph fusion representation sequence as feature input to obtain a policy document analysis result set. According to the method, the timeliness, the accuracy and the interpretability of automatic analysis of the policy document are greatly improved.

The invention relates to an emotion calculation and information processing method based on a deep neural network, and belongs to the technical field of artificial intelligence and natural language processing (NLP). The method comprises the following steps: preprocessing input text data; word embedding is conducted on the preprocessed text data through a Word2Vec model, each word is mapped into a high-dimensional vector, and therefore the semantic relation between words is captured; and processing the word embedding representation by using a bidirectional recurrent neural network model Bi-RNN, a bidirectional long-short-term memory network Bi-LSTM or a bidirectional LSTM model Bi-LSTM CNN combined with a convolutional neural network to obtain a classification result, and realizing an emotion analysis task. According to the method, the precision and robustness of sentiment analysis are remarkably improved.