49 results about "Semantic learning" patented technology

The invention relates to the technical field of natural language processing, in particular to a multi-modal interactive intelligent NPC dialogue intention matching method and system. The method comprises the following steps: collecting multi-modal data in real time, and generating multi-modal semantic features through sub-modal preprocessing; a cross-modal fusion module based on semantic association is adopted to integrate each modal semantic feature, and an intention candidate set is generated based on semantic context representation and in combination with a semantic analysis module and a predefined intention template; through a continuous semantic learning mechanism and an interactive memory module, in combination with a Bayesian updating method, tracking user intention change in real time, dynamically adjusting the confidence of each intention in the intention candidate set, and screening a final intention; constructing an NPC semantic cognition model, and performing semantic check on user input and an NPC dialogue state through semantic consistency analysis; generating a dialogue strategy in combination with the final intention and a decision engine, and outputting synchronous response content; according to the invention, the precision of intelligent dialogue intention dynamic matching is improved.

The invention discloses a vulnerability detection method and system based on semantic sensitive contrast learning and graph representation, and belongs to the technical field of vulnerability detection.The method comprises the steps that source codes and transformed codes are input into a vulnerability detection model to be processed, and a vulnerability classification result is obtained; the vulnerability detection model comprises a semantic learning module, a graph representation learning module, a feature fusion module and a support vector machine; wherein the construction process of the vulnerability detection model comprises the following steps: taking a source code and a transformed code as a positive sample pair for comparative learning to obtain code semantic feature embedding; the source code is represented as a graph structure, the graph structure is processed, and then code structure feature embedding is obtained; and performing feature fusion on the code semantic feature embedding and the code structure feature embedding to obtain fusion vectors, and classifying the fusion vectors to obtain a vulnerability classification result. The model is guided to learn semantics related to vulnerabilities through comparative learning, and meanwhile, the accuracy of vulnerability detection is improved in combination with grammatical structure features of codes.

The invention discloses a progressive knowledge graph completion method based on semantic information and structural information. The progressive knowledge graph completion method comprises the following steps: preprocessing and initializing data, defining an input data format, creating and initializing an embedded cache queue, and performing corresponding name and description text information on each entity; entities and relationships are encoded through a text encoder, and names and description information of a head entity and a tail entity are subjected to text representation learning by using a BERT model to obtain semantic-based vector representation; performing structure enhancement on the entity through a structure encoder; a progressive training strategy is adopted, and the method is divided into a semantic learning stage and a structure fusion stage; performing model optimization by using comparative learning, and performing parameter updating by using an InfoNCE loss function by calculating similarity and constructing a negative sample; executing a reasoning prediction process, performing candidate entity generation and sorting on the query, and outputting a knowledge graph completion result; according to the method, the knowledge graph completion performance is remarkably improved.

The invention discloses a cross-modal pedestrian re-identification method and system based on a multi-scale joint learning network, and relates to the technical field of pedestrian re-identification, and the method comprises a four-flow network architecture which enables a model to extract diversified semantic features through a mode of separating an application data enhancement branch from an original branch. Random channel selection and self-adaptive graying are respectively applied to the data enhancement branch, so that the robustness of the model to color change and the adaptive capacity of the model to different thermal imaging conditions are improved. Important channels are enhanced through a channel attention mechanism, irrelevant channels are inhibited, and two substreams are guided to internally enhance modal specific features. Richer semantic information is reserved through features extracted by a four-flow network, a joint semantic learning module is designed, a group of learnable vectors are defined and spatial position codes are added, global features of original branches and color invariant features of channel data enhancement are fully learned under the guidance of loss, and the overall feature of the original branches is optimized; and the features among different modes have higher semantic consistency.

The invention discloses a low-illumination image enhancement method based on learnable semantic prior, and the method comprises the steps: firstly dividing the input into two branches, taking a low-light visible light image as the input of the two branches, enabling one branch to be used for the restoration and reconstruction of image contents, and enabling the other branch to be used for predicting the semantic prior of the low-light image; in the first branch, performing enhancement processing on the input low-light visible light image; in the second branch, learning and predicting semantic prior information of the image; and gradually recovering details of the image by using semantic priori acquired from a semantic learning device, and finally generating an enhanced visible light image. And finally, designing a loss function to guide a training process of a low-light image algorithm. Concealed details in the low-light image are disclosed through the learnable semantic priori prediction task, so that the generation quality of the image is improved.

The invention discloses a semantic-guided lightweight three-dimensional reconstruction method based on three-dimensional Gaussian, and the method comprises the steps: inputting a multi-view two-dimensional image, and extracting two-dimensional semantic features; projecting the two-dimensional semantic features to the three-dimensional Gaussian primitives; learning three-dimensional semantic distribution in a knowledge distillation mode, and predicting three-dimensional semantic embedding features; dividing a large scene into a plurality of sub-regions, and performing divide-and-conquer three-dimensional Gaussian optimization and semantic learning; performing semantic-guided lightweight modeling on the three-dimensional Gaussian primitives in each sub-region; executing sub-scene fusion, and performing quantitative compression on the three-dimensional data in the fused scene; and a rasterization renderer is used to carry out efficient real-time rendering under any visual angle, and a lightweight large-scene three-dimensional reconstruction result is output. According to the method provided by the invention, the real-time rendering performance of a complex scene can be remarkably improved, the model volume is effectively compressed, meanwhile, the model training process is accelerated, and the method is suitable for efficient three-dimensional visual reconstruction tasks.

The invention provides a wetland semantic change detection system and method based on frequency domain information and multi-scale feature extraction, the system is a dual-domain adaptive frequency sensing wetland semantic change detection network structure, and the network structure comprises a BSFE module, an MASPP module, an AFSM module and an SCG module which are connected in sequence; the BSFE module adopts an AFT module based on frequency domain information to extract high-level and low-level features of an image, and improves the perception capability of subtle changes through frequency enhancement; the MASPP module enhances the sensing ability of the network under different scales through dense void factor combination and multi-scale pooling operation; the AFSM module enhances a specific frequency component through frequency domain convolution; and the SCG module utilizes time sequence consistency constraint to guide semantic learning, fuses the change features and the spatial-temporal features, and generates a final semantic change graph through mask operation. According to the invention, the precision and efficiency of wetland change detection can be improved, and the calculation overhead is reduced.

This invention relates to the field of natural language processing technology, specifically to a method and system for matching the intent of intelligent NPC dialogues in multimodal interaction. The method involves real-time acquisition of multimodal data, generating multimodal semantic features through submodal preprocessing; integrating the semantic features of each modality using a cross-modal fusion module based on semantic association, generating a candidate intent set based on semantic context representation and combined with a semantic parsing module and predefined intent templates; tracking changes in user intent in real time through a continuous semantic learning mechanism and an interactive memory module, combined with a Bayesian update method, dynamically adjusting the confidence level of each intent in the candidate intent set, and filtering the final intent; constructing an NPC semantic cognition model, and performing semantic consistency analysis to perform semantic checks on user input and NPC dialogue state; combining the final intent and a decision engine to generate a dialogue strategy and output synchronized response content; this invention improves the accuracy of dynamic matching of intelligent dialogue intents.

The application relates to a multi-dimensional time sequence recognition method, device and equipment based on time-frequency semantic learning. The method comprises the following steps: constructing a task-oriented segmented basic unit, performing semantic segmentation sampling on an input multi-dimensional time sequence to obtain a plurality of initial segments, obtaining a basic segment after weighting by a weight calculation network, constructing a time domain and frequency domain component generation unit based on the basic unit, inputting multi-dimensional time sequence data monitored by different sensors in the same scene and time period into the two units for parallel processing to obtain time-frequency domain segments, inputting the time-frequency domain segments into a downstream sequence recognition model for prediction, calculating time-frequency domain consistency loss and classification loss, adjusting parameters of the two units until the loss converges, completing training, extracting time-frequency domain segments of a multi-dimensional time sequence by using the trained unit, and inputting the time-frequency domain segments into the model for recognition. The method can effectively improve the classification accuracy of multi-dimensional time sequence.

The invention relates to a method of performing a map learning via a unified framework comprising the steps on a computing device (10): - Retrieving data comprising features that are geospatially referenced and / or localized with respect to geospatially referenced; (S1) - Arranging the features in relation to each other on the basis of their localization and / or their geospatially referenced features, whereby a lane is determined and / or aligned; (S2) - Aggregate and / or cluster geometric features; (S3) - Determine a lane divider and / or lane segments based on the geometric features and / or based on lanes traveled; (S4) - Deriving a lane topology based on a connectivity of the lane segments; (S5) and - Semantic learning based on the lane topology, whereby features are assigned a function and the map is thus created or expanded. (S6) Furthermore, the invention relates to computer program and a computing device (10).

The present invention discloses a rumor detection method using a self-attention generator and a BiLSTM discriminator, comprising the following steps: collecting rumor text data to form a rumor dataset; constructing a generative adversarial network generator including a self-attention layer based on a self-attention mechanism; constructing a discriminator network to perform rumor detection and classification on the original rumor text and the text decoded by the generator; training the generative adversarial network to adjust the model parameters of the generator and the model parameters of the discriminator; extracting the discriminator network of the generative adversarial network to perform rumor detection on the text to be detected. The rumor detection method of the present invention has high detection accuracy and good robustness. The self-attention generator is used to construct key features through semantic learning of rumor samples, and generate text samples rich in performance features to simulate the information loss and confusion in the rumor propagation process. The discriminator's semantic feature recognition ability is enhanced through adversarial training.

The invention discloses an unsupervised domain adaptive method based on a cascade learning strategy, and aims to decouple the training process of a source domain and a target domain, prevent semantic information confusion and realize fine-grained adaptation to the target domain. Specifically, the proposed two-stage cascade learning framework mainly comprises two stages of learning processes, in the first stage, a low-rank adaptation fine tuning module is used for performing fine tuning of a source field on a visual language pre-training model so as to learn feature information which is related to categories and invariable in field; and in the second stage, learning knowledge of a target domain by freezing the fine-tuned visual language pre-training model and a low-rank adaptive fine-tuning module and introducing text prompt, and refining a pseudo tag by means of knowledge of a source domain. According to the method, semantic learning and a domain-specific adaptation process are effectively decoupled, so that the performance of the target domain is improved.

The application provides an embedding generation method and device of a binary function, electronic equipment and a medium, and relates to the technical field of computers. The method comprises the following steps: inputting a target binary function into a pre-trained control flow semantic learning model to obtain a control flow semantic embedding, wherein the control flow semantic learning model is used for converting a relationship control flow graph corresponding to the binary function into a vector and outputting; inputting the target binary function into a pre-trained global sequential semantic learning model to obtain a global sequential semantic embedding, wherein the global sequential semantic learning model is used for converting machine code corresponding to the binary function into a vector and outputting; and integrating the control flow semantic embedding and the global sequential semantic embedding, and taking the integration result as an embedding corresponding to the target binary function. The scheme of the application can make the embedding of the binary function retain code information to the greatest extent, has good generalization performance, and improves the quality of the binary function embedding.

The application relates to a pedestrian attribute recognition method based on attribute semantic similarity matching, and relates to the technical field of computer vision. First, an adaptive semantic query module is designed, adaptive queries combined with visual features are used to learn attribute-specific spatial distribution, and semantic information of each attribute is captured. In addition, pedestrian attribute recognition is reconstructed as a semantic matching task, attribute text features are used as semantic anchors, and the distance between the semantic information of the query and the anchors is used to predict the pedestrian attribute. Finally, a dynamic negative semantic learning strategy is proposed, the spatial prior information of the attribute is combined with the learnable parameters to generate the negative semantic information of the attribute, the attention area of the semantic query module to the attribute is constrained, and more accurate pedestrian attribute recognition is realized.

The invention provides a task planning method, a control device and a storage medium, and belongs to the technical field of long-range planning. The method comprises the following steps: constructing a corresponding rich semantic learning graph according to a target planning problem of a PDDL problem format; and inputting the rich semantic learning graph into the trained hybrid planning network, calculating a heuristic value corresponding to each state, guiding to search a transfer solution from an initial state to a target state, obtaining a planning action sequence of a target planning problem, and constructing the hybrid planning network based on node attention and a relational graph neural network. According to the method, semantic information is extracted by adopting a large language model, and a rich semantic learning graph and a hybrid planning network which are used for fully capturing PDDL problem information to assist task planning are constructed; the success rate of the planning problem can be improved by utilizing the rich semantic learning graph and the hybrid network architecture, and the reliability and the high efficiency of long-range planning are effectively guaranteed.

The invention belongs to the technical field of data processing and generation, and particularly relates to a federal cross-domain retrieval entity recommendation method and system for mineral resource retrieval, and the method comprises two stages of federal cross-domain semantic learning and behavior prediction based on a large language model. The method comprises the following steps: firstly, locally extracting semantic features of texts of knowledge entities from each data domain, encrypting the semantic features and uploading the encrypted semantic features to a server, and mining a cross-domain public semantic structure by the server through clustering to generate a shared semantic vector and issuing the shared semantic vector; and obtaining the user and entity representation of the ID modal based on the user-entity interaction sequence, and embedding and fusing the issued shared semantic vector and the ID modal entity through knowledge distillation to obtain an enhanced local representation. And projecting the user and enhanced entity representation into a soft prompt which can be understood by a large language model through a mapping network, forming a mixed prompt in combination with a task instruction, inputting the mixed prompt into the large language model for reasoning, and recommending a next possible retrieval entity to the user. And on the premise of protecting user privacy, knowledge migration and fusion among non-overlapping fields are realized.

The application discloses a kind of unified visual and semantic learning multimodal image collaborative fusion method, it is related to infrared and visible light image fusion field. Including: based on sample scene visible light image and infrared image generate shared semantic representation;Through hierarchical interactive attention module and fusion specific enhancement module respectively to shared semantic representation and the visible light feature of visible light image is processed, generates the spatial domain feature of segmentation task and fusion task;Through frequency perception task router, shared semantic representation is processed to generate the frequency domain feature of segmentation task and fusion task;Based on the spatial domain feature and frequency domain feature corresponding to segmentation task and fusion task respectively, the task feature of segmentation task and fusion task is generated;Based on the task feature of segmentation task, preliminary segmentation image is obtained, based on the task feature of fusion task, preliminary fusion image is obtained, to realize in unified framework, visual fidelity, significant target expression and availability of high-level task are considered.

The present invention provides a method and system for moving target segmentation based on satellite video. The method first determines a reference frame image based on a query frame image to be segmented, and determines a corresponding true mask image based on the reference frame image. Then, shape prior extraction is performed based on the true mask image to determine an edge mask image. Finally, a moving target segmentation model is used to perform feature extraction, spatiotemporal semantic relationship modeling, and semantic affinity constraint processing on the query frame image, reference frame image, true mask image, and edge mask image, and outputs a target mask image for representing the moving target segmentation result. In this way, the method can achieve efficient interactive learning of semantic information within and between satellite video frames, improve the robustness and accuracy of semantic learning of the moving target segmentation model, and significantly enhance the edge feature expression capability, thereby effectively improving the efficiency and accuracy of moving target segmentation in satellite videos.

The invention discloses a metaphor recognition method based on anchor-based double example and multi-theory semantic interaction. The method comprises the following steps: step 1, normalizing a metaphor text; step 2, anchor foundation double example group construction; step 3, performing context semantic learning; 4, multi-theory semantic interactive learning is carried out; and 5, performing model joint loss training. According to the method, the deep semantics of the sentences can be accurately understood, the existence of the metaphor semantics in the sentences can be accurately identified, and the limitation of a traditional metaphor identification method on the complex metaphor processing capacity and context understanding is improved.

The invention relates to a fraud number recognition method and system, and the method comprises the steps: carrying out the voice abnormality recognition, text abnormality recognition and behavior abnormality recognition of the obtained multi-mode communication data of each number; weighted dynamic scoring is carried out on the voice abnormity recognition result, the text abnormity recognition result and the behavior abnormity recognition result obtained by each number through a weighted fusion mechanism, so that current case-related numbers are obtained, and each case-related number is used as a core node. Constructing a case-related association network based on a core node according to the multi-mode communication data of each case-related number, performing semantic learning on the case-related association network by adopting a GAT map attention network, and generating a fraud prediction model based on the captured fraud mode and the active period of the core node; and carrying out identification and early warning on future case-related numbers through the fraud prediction model. Therefore, according to the method and the system, the fraud number is accurately identified, and meanwhile, early warning can be performed on the occurrence of future fraud behaviors.

The invention relates to the technical field of enterprise information management systems, and discloses an enterprise information management method and an enterprise information management system.The method comprises the steps that interaction demand information of a first system in a second system in enterprise information management is calculated; constructing an interaction strategy optimization semantic learning model in the interaction module, and mapping a first system internal field and a corresponding second system internal field in the interaction demand information to determine a plurality of interaction nodes in the second system; corresponding interaction mapping nodes are established in the interaction module; and when it is detected that data update exists in the interaction demand information, determining an interaction node of the to-be-synchronized data to the interaction mapping node, synchronizing the updated interaction demand information in the interaction node to the interaction mapping node, and generating feedback data sent to the first system. According to the invention, the problem of data islands in the existing enterprise information management can be solved.

The present invention discloses a method for reconstructing fine-grained tactile signals with audio-visual assistance. First, the tactile signals used for training are passed through a tactile autoencoder, and tactile features are extracted based on a clustering task. Then, these features are transferred to audio and image feature extraction networks to achieve feature extraction for audio and image signals. Next, a triplet constraint is used to optimize the multi-modal fusion mapping function for tactile, audio, and image, so as to obtain fused features from the extracted audio and image features. Finally, the fused features are input into a tactile generation network to achieve the reconstruction of fine-grained tactile signals. The present invention well solves the problems of weak supervision and weak pairing existing between multi-modal signals, realizes cross-modal shared semantic learning, enhances the clustering characteristics while ensuring the structural and semantic integrity of the generated tactile signals, thereby significantly improving the reconstruction quality of tactile signals.

A method, computer system, and computer program product for leveraging semantic learning-enhanced joint learning are provided. An aggregator can receive cluster information from a distributed computing device. The cluster information may relate to identified clusters in sample data from the distributed computing device. The aggregator can integrate the cluster information to define categories. This integration may include identifying any redundant clusters among the identified clusters. The number of categories may correspond to the total number of clusters from the distributed computing device minus any redundant clusters. A deep learning model can be sent from the aggregator to the distributed computing device. The deep learning model may include an output layer with nodes that may correspond to the defined categories. The aggregator can receive the results of joint learning performed by the distributed computing device. Joint learning can train the deep learning model.

An embodiment of the present application proposes an artificial intelligence-based image processing method, device and storage medium, the method comprising: obtaining a training sample set, the training sample set comprising a first image sample set with category labels and a second image sample set consisting of triplets; using the training sample set to perform semantic learning and metric learning training on an original model, the original model comprising a first branch network and a second branch network, the first branch network and the second branch network comprising shared network parameters; determining an image feature extraction model based on the trained original model, the image feature extraction model being used to extract feature vectors of an image, enabling the model to achieve metric learning while having semantic extraction capabilities, and the image features extracted based on the model can improve the accuracy of image retrieval.

The application discloses a text classification method based on label semantic learning and attention adjustment mechanism, and mainly comprises the following steps: preprocessing text data, extracting text semantic features, text label graph embedding, using a multi-head adjustment attention mechanism to measure the semantic relationship between words and labels, then multi semantic integration and network training, thereby realizing multi-label text classification, training the model, and then using the trained model to predict the category of a text. The application proposes a multi-head adjustment attention hybrid BERT model for a multi-label text classification framework, which can effectively extract useful features from text content, establish semantic connection between labels and words, obtain label-specific word representation, and thus improve the performance of multi-label text classification.

This invention discloses a 3D point cloud analysis method and apparatus based on Transformer that enhances local semantic learning capabilities. The method constructs a deep learning model, including a local semantic self-attention learning module, which can perceive global context information and acquire finer-grained local semantic features in parallel, thereby enhancing the perception capability of the entire network framework. Local and global features of the point cloud are obtained from different modules through four local semantic learning modules, and these features are converged to obtain global joint features. The joint features obtained in the learning stage are then sent to subsequent point cloud classification and segmentation stages to obtain semantic classification results and final segmentation results. This invention can acquire key local geometric semantic information in 3D data and has significant advantages in various 3D point cloud analysis applications, such as point cloud classification, point cloud segmentation, and semantic segmentation of large indoor scenes.

The invention discloses a scalable cross-modal tactile signal generation method applied to an emergency rescue scene, which comprises the following steps: encoding audio and video signals at a device end, reducing the difference between modals by utilizing comparative learning, sensing network bandwidth information by utilizing an edge end, and timely feeding back the network bandwidth information to the device end; the method comprises the steps that a device end is guided to complete scalable semantic coding, and finally, after an edge end receives corresponding cross-modal fused semantic information or different hierarchical representations of fused semantics, a tactile signal with corresponding granularity is generated by utilizing a designed adaptive tactile signal generation strategy. According to the invention, the problem that the experience of an operator executing remote control is finally influenced due to the fact that the tactile signal is difficult to directly acquire and the transmission bandwidth is dynamically fluctuated in emergency rescue is solved, related semantic learning and cross-modal generation of a multi-modal signal are realized, and the experience of the operator executing remote control is ensured to be greatly improved under the condition that the network bandwidth is dynamically fluctuated. Tactile signals are reliably obtained in real time, and the rescue efficiency of operators is finally improved.

The invention relates to a multidimensional time sequence recognition method, device and equipment based on time-frequency semantic learning. The method comprises the following steps: constructing a task-oriented word segmentation basic unit, carrying out semantic word segmentation sampling on an input multi-dimensional time sequence to obtain a plurality of initial segmented words, weighting by a weight calculation network to obtain basic segmented words, constructing a time domain and frequency domain component generation unit based on the basic unit, and generating a time domain and frequency domain component according to the time domain and frequency domain component generation unit. Multi-dimensional time sequence data monitored by different sensors in the same scene and the same time period are respectively input into two units for parallel processing to obtain time-frequency domain segmented words, the time-frequency domain segmented words are input into a downstream sequence recognition model for prediction, time-frequency domain consistency loss and classification loss are calculated, parameters of the two units are adjusted until loss convergence, and training is completed. And extracting time-frequency domain segmented words of the multi-dimensional time actual measurement sequence by using the trained unit, and inputting the time-frequency domain segmented words into a model to realize recognition. According to the method, the multi-dimensional time sequence classification accuracy can be effectively improved.

The disclosure provides a quality evaluation model training method and device and a multi-round dialogue quality evaluation method, relates to the fields of artificial intelligence such as natural language processing and deep learning, and comprises the following steps: obtaining an initial quality evaluation model, training the initial quality evaluation model, obtaining a trained candidate quality evaluation model, labeling a first sample multi-round dialogue to obtain a sample label vector set, training the candidate quality evaluation model according to the sample label vector set, and obtaining a trained target quality evaluation model. The semantic learning effect of the candidate quality evaluation model on the first sample multi-round dialogue is optimized, the quality evaluation capability of the candidate quality evaluation model for the first sample multi-round dialogue is improved, the quality evaluation efficiency and precision of the multi-round dialogue are improved, the degree of artificial dependence and the artificial cost are reduced compared with the quality evaluation of the multi-round dialogue relying on artificial implementation, and accurate data support is provided for a downstream task.

The application relates to the technical field of enterprise information management systems, and discloses an enterprise information management method and an enterprise information management system, the method comprising the following steps: calculating interaction demand information of a first system in a second system in enterprise information management; constructing an interaction strategy optimization semantic learning model in an interaction module, mapping internal fields of the first system and corresponding internal fields of the second system in the interaction demand information, so as to determine a plurality of interaction nodes in the second system; establishing corresponding interaction mapping nodes in the interaction module; when it is detected that data updating exists in the interaction demand information, determining an interaction node of the interaction mapping node to which to-be-synchronized data are synchronized, synchronizing updated interaction demand information in the interaction node to the interaction mapping node, and generating feedback data sent to the first system. The application is beneficial to solving the problem of data islands existing in the current enterprise information management.