115 results about "Web learning" patented technology

The invention discloses an unsupervised domain adaptive semantic segmentation method, system and device based on inter-domain consistency and a medium. The method comprises the following steps: translating source domain image data to a target domain style while keeping the content by using a GAN-based multi-scale generative adversarial network; inputting the source domain image data into an inter-domain consistency training network based on a Transform model for adaptive training, wherein the source domain image data subjected to data enhancement is sent to a student network for source domain semantic segmentation training, and an EMA strategy is used for updating a teacher network weight; inputting the target domain label-free image data into a teacher network to generate a pseudo label so as to further guide students to train the network; randomly combining the source domain image data and the target domain image data by using a hybrid strategy to guide students to learn domain invariant knowledge in a network; and finally, adjusting and optimizing the student network according to the consistency of the attention feature map, outputting an unsupervised field adaptive semantic segmentation model, and completing segmentation of target domain image data under the condition of not using a target domain label. According to the method, a model with excellent performance can be trained by using fewer real image data sets in the unmanned driving field, so that the training time and cost are greatly reduced, and the working efficiency is improved.

The invention relates to the technical field of data processing, in particular to a robot control method, device and system based on a diffusion model and a medium, and the method comprises the steps: collecting the position information of the tail end of an index finger and the state information of a flexible linear object in a human hand teaching process, and constructing a training sample; training the training sample by using a diffusion model, and learning mapping from a high noise state to a clear target state through a noise prediction network in the training diffusion model to obtain a trained diffusion model; the training process of the diffusion model comprises forward noise addition and reverse noise reduction; the state of the flexible linear object and the tail end position of the mechanical arm are collected, a current observation environment is constructed, the current observation environment is input into the trained diffusion model, a target action sequence is generated, and the mechanical arm is controlled to execute control actions based on the target action sequence; according to the invention, the accuracy and real-time performance of the robot to control the flexible linear object can be improved.

The invention relates to a federal location recommendation method based on heterogeneous graph semantics, which comprises the following steps: constructing a corresponding local heterogeneous graph by using a local private data set, learning POI node features in the local heterogeneous graph by using a heterogeneous neural network, and learning meta-path semantic features through an attention mechanism. Fusing the track sequence features and the corresponding meta-path semantic features to train a local model; and fusing the parameters of the local model obtained after the training into global information through a knowledge distillation method. And encrypting and uploading the local model parameter, the local data volume parameter and the training result after the knowledge distillation of this round to a server. And after the server receives the uploads of all the clients, global aggregation is carried out, and global model parameters are updated by using the aggregated parameters, so that one-time training is completed. And for a new user, inputting the historical trajectory data of the new user into the optimal local model on the corresponding local client, and outputting the historical trajectory data as the next point of interest recommended for the new user.

The invention provides a Chinese ancient book entity extraction method based on a deep active learning strategy, and relates to the field of natural language processing. According to the method, the active learning strategy and the entity extraction model based on the deep neural network are subjected to model integration, so that the model fitting capability of a deep network learner can be brought into full play, and the problems of insufficient annotation data and relatively high expert annotation cost are further relieved by fully utilizing the improved active learning strategy. Meanwhile, data pool selection is carried out by adopting a query strategy fusing uncertainty, minimum boundary property and multi-spatial feature characterization, so that the learning efficiency is effectively improved. Furthermore, for an entity extraction network model, a double-encoder structure is designed, a global encoder based on pre-training learning and a local encoder based on character-level embedding are integrated, and fused semantic representation of sentences is obtained, so that more comprehensive context semantic knowledge of the sentences is comprehensively captured; therefore, the performance of the entity extraction model based on the deep neural network is improved.

The invention relates to a low-carbon energy-saving operation method and system for an edge node of the Internet of Things, and relates to the technical field of data processing, and the method comprises the steps: constructing a state vector of space-time correlation through dynamic weighted fusion based on the state information of the edge node collected in real time; constructing a comprehensive energy efficiency evaluation index based on the state vector; taking an energy efficiency evaluation index as an optimization target, and generating an action strategy through a deep Q network learning state and action mapping relation; according to the action strategy, dynamically adjusting the operation frequency of the node equipment, the task queue priority and the node working mode; based on the deviation between the actual energy efficiency and the target value, adaptively adjusting the weight and learning rate of each model; and performing dynamic attenuation correction on the execution strategy of each edge node through each updated model to obtain a target strategy, and controlling node equipment to operate in a low-carbon and energy-saving manner based on the target strategy. According to the invention, the low-carbon energy-saving capability and the operation safety of node equipment can be improved.

The embodiment of the invention provides a big data analysis visualization method and device based on a big model, a medium and a product. The method comprises the following steps: receiving a natural language demand input by a user; performing semantic layer, logic layer and task layer decomposition on the natural language demand by adopting a preset hierarchical analysis model to generate an executable data analysis instruction; obtaining target data required by the data analysis instruction; according to the data analysis instruction, the target data, a preset chart recommendation algorithm and an aesthetic evaluation optimization algorithm, generating a basic visual chart; and learning the user preference characteristics through the visual generative adversarial network, dynamically adjusting the basic visual chart according to the user preference characteristics, and generating a personalized visual result corresponding to the natural language demand. According to the scheme, the defects of an existing dragging type visualization system and an existing programming type visualization scheme in the visualization process are overcome.

The invention discloses a black box large language model-oriented knowledge-guided reinforcement learning discrete cue word optimization method, and belongs to the technical field of natural language processing and artificial intelligence optimization. According to the method, a KPE-RL (Known-gued Prompt Evolution with Reforming Learning) algorithm is put forward, editing operation is constrained by using a structured phrase-level knowledge base, cue words are optimized and modeled into a Markov decision process, and a discrete editing strategy is learned through a deep Q network. According to the method, a mixed state coding mode fusing semantic embedding and task statistical characteristics is designed, and a diversity regular reward is designed, so that efficient exploration is encouraged, and the optimization capability of the model in a limited API calling scene is improved. The method is suitable for automatic prompt optimization of the large language model in the API black box scene, and has generalization, robustness and practical application value.

The invention discloses a weak supervision semantic segmentation method and system based on a component perception learning segmentation network. The method comprises the following steps: step 1, constructing the component perception learning segmentation network; step 2, acquiring an SAM mask set of the training set image; step 3, obtaining a mask pseudo label set of the training set image and a relation matrix between a target component in a mask and the whole; 4, obtaining an initial part perception learning segmentation network model; step 5, obtaining and optimizing a CAM set of a target mask in the training set image to obtain a final pseudo label set, and obtaining a semantic segmentation result in the to-be-segmented image by using a segmentation decoder; removing redundant masks through a part-whole refining module in the network, and capturing a relationship between a target part and a whole in the masks; through a component perception learning module in the network, the network is encouraged to learn the relationship between each component of the target in the pseudo tag set and the whole, and accurate and complete segmentation of the target under a weak supervision condition is realized.

The invention discloses an uncertainty action quality evaluation method based on video interaction enhancement, which is based on a visual framework and comprises the following steps of: firstly, establishing a single-person movement video data set, preprocessing a video, extracting features, exploring information interaction among video clips by using a self-attention mechanism, and evaluating the motion quality of the single-person movement video data set; the method comprises the following steps of: selecting a plurality of fragments, changing the fragments into mutually related groups instead of independent groups, endowing adaptive weights to the fragments, simulating a real scoring scene in which the completeness of the actions is greatly influenced by important actions, and finally, learning predictable score probability by using an uncertainty perception score distribution learning network, according to the method, the most advanced performance is realized on the action quality evaluation data set.

To address the problem of providing means for building a curriculum on the basis of question answers of users and generating problems, etc. according to the situation so as to differentiate from existing contents, in an online web learning system.SOLUTION: Provided is a system including: means for utilizing generative AI, building a curriculum on the basis of question answers of users, and generating problems, etc. according to the situation in an online web learning system; and a web service that can be accessed any time, 24 hours a day, 365 days a year.SELECTED DRAWING: Figure 1

The invention provides a collaborative distillation method and device based on semantic prompt, and belongs to the technical field of industrial defect detection. The method comprises the steps that a coupling teacher collaborative distillation structure composed of a main teacher network and an auxiliary teacher network is constructed, the main teacher network is used for modeling local detail information of an image, the auxiliary teacher network is used for extracting an abstract semantic relation of the image, and the main teacher network and the auxiliary teacher network complement each other and collaboratively guide students to learn representation characteristics with better distinction; the student network receives the feature guidance of the two teacher networks at the same time, and the modeling ability of the student network to a normal mode and the identification ability to an abnormal area are improved. In order to alleviate the problem of insufficient prior knowledge, a hierarchical cross-modal alignment mechanism based on semantic prompt is further introduced, the image is guided to be transited layer by layer from pixel-level details to image-level semantics through predefined text semantic prompt, effective fusion of multi-scale features and semantic information is realized, and effective defect detection is carried out. According to the invention, defects of different sizes and different types can be accurately detected, and the robustness is high.

The invention provides an AI-based underground pipe network electrical drawing generation method and system, and relates to the technical field of artificial intelligence, and the method comprises the steps: carrying out the feature extraction and fusion of a design demand text, a reference drawing and an engineering file; generating a three-dimensional position distribution thermodynamic diagram through the layout generation network; generating a network learning space connection relationship through the connection relationship; generating an initial drawing through the detail optimization network; and generating a final drawing by adopting multi-discriminator collaborative verification. According to the method, the underground pipe network electrical drawing conforming to the specification can be automatically and efficiently generated, and the design efficiency and quality are improved.

The invention relates to the technical field of artificial intelligence, can be applied to business scenes of financial science and technology, medical health, intelligent robots and the like, and discloses a decision-making method, device, equipment and medium based on an environment semantic atlas. Constructing and updating an environment semantic map based on the initial multi-modal fusion feature, fusing the environment semantic map and the initial multi-modal fusion feature to generate a deep fusion feature, determining current action distribution based on the deep fusion feature, and learning transmission mapping through a flow matching network to generate post-mapping action distribution; and generating an action sequence, adjusting the action sequence according to the environment interaction feedback, and updating the flow matching network parameters in combination with the environment interaction feedback and the reward function. According to the method, the environment semantic map and the multi-modal features are fused, the action distribution is optimized and the action sequence is dynamically adjusted in combination with the flow matching network, and the multi-modal decision-making performance and the environment adaptability are improved.

The invention discloses an e-commerce platform product selection and pricing method based on multi-agent reinforcement learning, aiming at the game decision problem between an e-commerce platform and a supplier, a platform agent and a supplier agent are constructed, the platform agent uses a deep Q network DQN to learn a product selection strategy, and the supplier agent uses an Actor-Critic algorithm to learn a pricing strategy. Through Stackelberg master-slave game modeling, a platform is used as a leader to make a decision for selection, and a supplier is used as a follower to make a decision for pricing after observation. A state space (including commodity sales volume, inventory, user score, supplier fulfillment rate and the like), an action space (discrete commodity selection and continuous pricing) and a multi-target reward function (balance profit, user satisfaction, inventory turnover and the like) suitable for an e-commerce scene are designed. The strategies of the two parties are converged to game equilibrium through alternate training, and online continuous learning is supported to adapt to a dynamic market environment. Experiments show that compared with an existing method, the method has the advantages that the overall income can be improved by 18-30%, the win-win situation of the platform and the suppliers is achieved, and the blank of multi-agent reinforcement learning in e-commerce product selection and pricing scenes is filled.

The invention relates to the technical field of physiological signal measurement, and discloses a light-weight electrocardiogram diagnosis method and device based on a generative adversarial network and distillation learning. The method specifically comprises the following steps: converting a 12-lead electrocardiogram signal to obtain multiple views, learning the characteristics of each view in the multiple views by utilizing a teacher network, and fusing to obtain real characteristics; the student network is used as a generator, features are extracted from the single-lead electrocardiogram signals, and generated features which are as vivid as possible are obtained; the input features are classified through a discriminator, and whether the input features are from the teacher network or the student network is determined through the classification result. Through combination of multi-view feature extraction and generative adversarial learning, the knowledge migration ability of the teacher model is effectively improved, so that the student network maintains similar performance to the teacher network while the model is lightweight.

The invention relates to the field of computer vision processing, in particular to a video retrieval method based on a deep neural network model and multi-instance learning, which comprises the following steps: pre-training a query text to obtain an initial feature, and extracting an I 3D-RGB feature, an ROI feature and a connection feature from a video; updating to obtain frame-level visual features and word-level text features; constructing a training graph, and performing graph attention network learning to obtain word-level text features; calculating feature residual errors of the word-level text and the word-level text, wherein the mean value of the residual errors is sentence-level text features; performing segment dimension average operation on the frame-level visual features to obtain pipeline-level visual features; calculating alignment scores of the sentence-level text features and the pipeline-level visual features, constructing a positive sample pair and a negative sample pair, and training a video retrieval network; according to the method, discriminability features in a multi-query text are obtained through deep learning features to construct the graph neural network, so that text features with more representation meanings and multi-modal alignment supervision signals are provided under weak supervision setting.

Disclosed herein is the framework for Causal Reinforcement Learning which combines Causal Cooperative Networks with the Actor-Critic algorithm, introducing reverse dynamics, and invertible-reasoning policy within the framework to enable bidirectional transitions while maximizing accumulative rewards. The framework involves a redesigned Critic and Actor, as well as a newly developed Reverse-environment network. During the iterative training and exploration phases, the cooperative network learns a policy that identifies actions capable of reversing a future state back to its prior state, through the reverse-environment network. It allows agents to consider the consequences of their actions, which facilitates deeper decision-making and exploration strategies.

Embodiments of the present application provide a deep learning-based phase unwrapping method and apparatus, a device, and a medium. The method comprises: acquiring a wrapped phase map; and inputting the wrapped phase map into a pre-trained deep learning network model and performing phase unwrapping, such that the deep learning network model outputs, on the basis of a non-linear mapping relationship between wrapped phases and absolute phases, an absolute phase map corresponding to the wrapped phase map, wherein the deep learning network model adopts a U-Net framework and comprises an encoder, a decoder, a deep-supervision feature fusion module, and an atrous spatial pooling pyramid module, and the encoder and the decoder comprise Inception modules. On this basis, even in high-noise environments and in the presence of phase discontinuity, the embodiments of the present application enable a trained deep learning network model to predict a high-quality absolute phase map in just one step, thereby achieving accurate phase unwrapping.

The invention relates to a holographic semantic content efficient transmission method based on hierarchical reinforcement learning, and belongs to the technical field of wireless communication. The method comprises the following steps: S1, establishing a target-oriented holographic content interactive semantic transmission framework; s2, establishing an SSP optimization problem mathematical model aiming at minimizing the total accumulated resource cost; s3, the SSP optimization problem is decomposed into a two-stage hierarchical decision-making structure, and the two-stage hierarchical decision-making structure comprises a high-layer element controller used for strategic semantic planning and a low-layer controller used for tactical resource allocation; and S4, an algorithm based on hierarchical reinforcement learning is designed to solve an optimization problem, a high-layer element controller adopts a deep Q network to learn a long-term semantic sorting strategy, and a low-layer controller executes an instant optimal power distribution strategy. According to the method, the total cumulative cost can be remarkably reduced, and the excellent efficiency of layered semantic transmission in a dynamic wireless environment is proved.

This application provides a big data analytics visualization method, apparatus, medium, and product based on a large model. The method includes: receiving a natural language input from a user; using a preset hierarchical parsing model to decompose the natural language input into semantic, logical, and task layers, generating executable data analysis instructions; acquiring the target data required by the data analysis instructions; generating a basic visualization chart based on the data analysis instructions, target data, a preset chart recommendation algorithm, and an aesthetic evaluation and optimization algorithm; learning user preference features through a visualization generation adversarial network, and dynamically adjusting the basic visualization chart according to the user preference features to generate a personalized visualization result corresponding to the natural language input. This application's solution addresses the shortcomings of existing drag-and-drop visualization systems and existing programmatic visualization solutions in the visualization process.

The application relates to the field of computer vision processing, in particular to a video retrieval method based on a deep neural network model and multi-example learning, which comprises the following steps: obtaining initial features by pre-training a query text, extracting I 3D-RGB features, ROI features and connection features from a video; updating frame-level visual features and word-level text features; constructing a graph for training, learning word-level text features by using a graph attention network; calculating the residual error of the word-level text features and the word-level text features, and taking the mean value of the residual error as a sentence-level text feature; performing segment dimension average operation on the frame-level visual features to obtain pipeline-level visual features; calculating the alignment score of the sentence-level text features and the pipeline-level visual features, constructing positive sample pairs and negative sample pairs, and training a video retrieval network; and the application constructs a graph neural network by acquiring discriminative features in multiple query texts through deep learning features, so as to provide text features with more representation meanings and multi-modal alignment supervision signals under weak supervision settings.

The invention relates to the technical field of computer vision, artificial intelligence and multimedia security, in particular to a 3D field speaking face generation defense method, system and device and a medium, and relates to a reinforcement learning-based anti-interference method and system which can effectively protect a reference video from the source and enable the reference video not to be used for generating a high-quality speaking face video. Modeling a reference video protection problem into a sequence decision process by constructing a reinforcement learning environment; generating an antagonistic interference action by utilizing strategy network learning, designing a dual-objective reward function, and optimizing an interference effect and visual concealment at the same time; and finally, iterative training is carried out through a deep reinforcement learning algorithm, so that the model autonomously masters the optimal interference strategy.

The unsupervised domain self-adaptive semantic segmentation method, system, device and medium based on inter-domain consistency use a multi-scale generative adversarial network based on GAN to translate source domain image data to target domain style while preserving its content; then input the inter-domain consistency training network based on the Transformer model for adaptive training, including sending the source domain image data after data enhancement to the student network for source domain semantic segmentation training and updating the teacher network weight using the EMA strategy; input the target domain unlabeled image data into the teacher network to generate pseudo labels to further guide the student network training; use a hybrid strategy to randomly combine the source domain image data and the target domain image data to guide the student network to learn domain invariant knowledge; finally, the student network is optimized according to the attention feature map consistency, and an unsupervised domain self-adaptive semantic segmentation model is output, which can complete the segmentation of the target domain image data without using the target domain label; the present application can train a model with excellent performance using less real image data set in the field of unmanned driving, greatly reducing the time and money cost of training, and improving the work efficiency.

The application discloses a lightweight adaptive network learning method for multi-modal multi-task learning, comprising the following steps: 1, constructing a downstream task dataset, 2, constructing a deep self-attention network model, 3, pre-training weight pruning, 4, constructing a task adapter, 5, adapting a pre-training model, 6, designing a progressive guided distillation training and training a model. The method prunes part of the weights of the pre-training model and adapts it with an efficient task adapter. The application proposes a progressive guided distillation training algorithm to better fill the gap between the pre-training task and the downstream task, ensuring the performance of the model on the downstream task. The application can be combined with any existing pre-training model based on a deep self-attention network, and through training, an adapter model with superior performance in the downstream task, total storage overhead during model deployment, calculation overhead during model inference and model configuration flexibility is obtained.

This invention relates to a scalable industrial vision task offloading method and scheduling device based on energy consumption awareness, belonging to the field of task scheduling and resource management technology. To address the problems of current task offloading methods failing to simultaneously consider reliability, latency, and energy consumption, and lacking flexibility in adjusting latency and energy consumption, this invention first introduces energy consumption concern weights to establish a hierarchical transmission and computation energy consumption model, and constructs a utility function that comprehensively considers reliability, latency, and energy consumption. Then, task offloading scheduling is modeled as a Markov decision process using deep reinforcement learning, with the utility function as the immediate reward. Through the interaction between the deep reinforcement learning agent and the environment, a policy network is trained to learn the optimal joint decision on execution location and offloading layer under different channel and computing power conditions, thereby maximizing long-term average utility and satisfying latency constraints. Finally, intelligent scheduling is achieved based on deep reinforcement learning.

This invention relates to the field of data processing technology, specifically to a robot manipulation method, device, system, and medium based on a diffusion model. The method includes: collecting the position information of the index finger tip and the state information of a flexible linear object during a human hand teaching process to construct training samples; training the training samples using a diffusion model, and learning the mapping from a high-noise state to a clear target state through a noise prediction network in the training diffusion model to obtain a trained diffusion model; the training process of the diffusion model includes forward noise addition and backward noise reduction; collecting the state of the flexible linear object and the position of the robotic arm end effector to construct the current observation environment, inputting the current observation environment into the trained diffusion model to generate a target action sequence, and controlling the robotic arm to perform manipulation actions based on the target action sequence; this invention can improve the accuracy and real-time performance of robot manipulation of flexible linear objects.

The present invention provides a method for extracting Chinese ancient book entities based on a deep active learning strategy, which relates to the field of natural language processing. In the present invention, the active learning strategy is integrated with an entity extraction model based on a deep neural network, which can not only give full play to the model fitting ability of the deep network learner, but also make full use of the improved active learning strategy to further alleviate the problems of insufficient labeled data and high expert annotation costs. At the same time, a query strategy that combines uncertainty, minimum boundary, and multi-space feature representation is used to select the data pool, effectively improving the learning efficiency. Further, for the entity extraction network model, a dual encoder structure is designed, which integrates a global encoder based on pre-trained learning and a local encoder based on character-level embedding to obtain the fused semantic representation of the sentence, so as to comprehensively capture the more comprehensive context semantic knowledge of the sentence, thereby improving the performance of the entity extraction model based on the deep neural network.

The application discloses a kind of based on information bottleneck's debiasing recommendation method, comprising:1. Construct original data: user-product interaction matrix of user interaction to product, user biased attribute matrix;2. Using biased attribute encoder learns user biased representation in user-product interaction data;3. Based on deep graph neural network learns user and product representation matrix;4. Based on information bottleneck theory minimizes the mutual information between user representation and biased representation, user subgraph representation and biased representation, calculates loss function;5. Recomputes interaction matrix based on user and product representation matrix, calculates loss function;6. The loss function of joint step 4-step 5 is carried out information bottleneck learning, and model parameter is updated to model convergence.The application is based on the thought of information bottleneck, learns unbiased user representation under the demand of meeting recommendation task, so as to effectively reduce recommendation bias, guarantee the accuracy of recommendation.

A method for cooperating packing actions and unpacking actions of mechanical arm based on deep reinforcement learning (DRL), wherein a synergistic effect between a packing action and an unpacking action is learnt by means of a packing-unpacking network (PUN), thereby improving the packing precision and efficiency; and a packing heuristic algorithm and an unpacking heuristic algorithm are provided on the basis of human experience and are combined with the PUN, thereby further improving the packing precision and efficiency. Also provided are a system for cooperating packing actions and unpacking actions of mechanical arm based on DRL, a computer-readable storage medium, and an electronic device.

The application discloses a kind of based on AI model's place name classification method, system, equipment and storage medium, design surveying and mapping technical field, method includes: convergence multi-source professional surveying and mapping place name data, execute data preprocessing, based on the structured place name text feature vector, generate the semantic vector that can be directly identified by general AI big model, input big model extraction place name deep layer semantic structure feature, while extracting the geographical attribute characteristics corresponding to place name;Structural place name field knowledge system library is built;Introduce the basic reasoning framework built by Transformer big language model and traditional deep network learning technology, adopt knowledge enhancement strategy to optimize model, cross-modal reasoning is carried out by fusing place name semantic structure feature and geographical attribute characteristics, preliminarily determine the category of place name;Adopt prompt engineering driven zero sample and few sample learning mode, generate multiple candidate classification results, output final place name classification result after screening verification.