1264 results about "Network generation" patented technology

The present invention relates to the technical field of stomatology. Provided is an auxiliary dental implant generation method based on a diffusion model. The method in the present invention comprises: acquiring oral CBCT image data of historical patients, preprocessing the oral CBCT image data of the historical patients to obtain a CBCT image dataset, using the CBCT image dataset to train a multi-task segmentation network, and using the segmentation network to obtain an intraoral tissue segmentation result; using the intraoral tissue segmentation result to train detection networks from the three dimensions of a cross-sectional plane, a coronal plane and a sagittal plane, respectively; using the detection networks to obtain detection results in the three directions of the cross-sectional plane, the coronal plane and the sagittal plane; fusing the detection results in the three directions of the cross-sectional plane, the coronal plane and the sagittal plane, and using a majority voting algorithm to construct a three-dimensional bounding box, so as to acquire an edentulous area; and using the intraoral segmentation result and the edentulous area as prompt information to guide, by means of an iterative process, a network to generate a post-implantation effect. The implantation effect obtained by the present invention is highly accurate, thereby providing a more precise auxiliary tool for stomatology.

The invention discloses an intelligent substation communication link fault accurate positioning method and system, and the method comprises the steps: obtaining a configuration file and equipment state data, carrying out the processing of the configuration file and the equipment state data, and generating a standardized link feature vector and a marking data set; constructing a hybrid deep learning model, and optimizing parameter configuration of the hybrid deep learning model by adopting an optimization algorithm to obtain a parameter-optimized hybrid deep learning model; training by using a real fault sample in combination with a virtual fault sample generated by a generative adversarial network, optimizing a time sequence prediction capability through an echo state network, and generating a fault positioning model; in combination with the link state data, outputting a fault link positioning result and confidence evaluation through multi-stage confidence evaluation and topological correlation analysis; and carrying out virtual-real corresponding verification in combination with the configuration file, carrying out parameter optimization on the fault positioning model, and outputting a fault positioning system. The problems that the fault positioning precision is low, the response speed is low, and complex fault scenes cannot be processed are solved.

The invention relates to the technical field of safety management and risk prediction, in particular to an intelligent safety management and risk prediction method and system based on cloud computing. The method comprises the following steps: dynamically accessing multi-source heterogeneous data through a cloud platform, and forming unified event representation through time alignment and credibility labeling; constructing a hierarchical mixed probability safety twin model, updating dynamic parameters by adopting credibility weighted online variational Bayesian, and outputting a state interface by combining structural adaptation, cross-object graph regularization and physical constraint projection; mapping the twinborn state into a causal feature, constructing an intervening causal graph, generating causal embedding by using a credibility weighted attention network, simulating an intervention operation in an embedding space, and quantifying a risk probability; and generating a multi-candidate security policy, evaluating and sorting through a multi-objective utility function, executing an optimal policy, collecting feedback data, and updating the model and the policy. According to the method, credibility regulation and control, probability twinning and causal intervention are fused, and real-time intelligent decision making of an industrial safety scene is supported.

The invention discloses a multi-source-domain multi-teacher knowledge distillation method and system based on reinforcement learning, and the method comprises the steps: obtaining target domain sample data, inputting the data into N pre-trained teacher models, and generating the output features of all teacher models; inputting the target domain sample and all teacher model outputs into a reinforcement learning strategy network, generating a dynamic weight of each teacher model, and calculating a knowledge distillation loss function based on the dynamic weights; constructing a total loss function according to the knowledge distillation loss function and the cross entropy loss output by the student model; student model parameters are updated through gradient descent; and calculating a reward value according to student model performance change, and updating reinforcement learning strategy network parameters. According to the method, the reward function based on student model performance improvement is constructed, the strategy network is continuously updated in a strategy gradient optimization mode, the distillation efficiency is effectively improved, knowledge conflicts among teachers are relieved, and the robustness and generalization performance of the student model in a multi-source complex environment are remarkably improved.

The invention relates to the technical field of robot training. The robot self-adaptive training method based on reinforcement learning comprises the steps that task sub-target information is generated through a high-level strategy network, the task sub-target information is input into a low-level execution network, an action control instruction is generated according to the task sub-target information, interaction feedback information is collected in the execution process, and the action control instruction is sent to a robot through a robot. Calculating a reward value according to the interaction feedback information, carrying out association processing on the reward value and the scene complexity parameter, executing a dynamic reward shaping operation, generating an adjusted reward signal, generating a strategy model optimized by meta-learning based on the adjusted reward signal, loading the strategy model in a simulation environment, and carrying out dynamic reward shaping. A target strategy model optimized through simulation training is generated, the target strategy model is loaded to the robot, and the robot is controlled to execute task operation in the actual interaction scene. The method has the effect of realizing adaptive task learning of the robot in a multi-interaction scene.

The invention discloses a dynamic scheduling method and system for a satellite-ground cooperative computing task, and the method comprises the following steps: constructing a satellite-ground cooperative computing system comprising a remote sensing satellite cluster, a satellite server node, a ground server / data center and the like, and enabling the satellite cluster to generate an in-orbit task and to describe dependence through DAG; available computing power, task queues and other resource states of a satellite and a ground server are collected in real time, and global observation is formed; on the basis of an MADRL framework, each node is regarded as an independent agent, a task allocation action is generated through a strategy network, and a collaborative decision is modeled through POMDP; optimizing the task execution sequence of the same node by using an EFT algorithm; issuing a task and monitoring execution; task completion time and energy consumption are collected, and a network optimization strategy is updated through a reward function and experience playback; the scheduling strategy is adjusted through loop iteration, satellite offline and link fluctuation are adapted, and the resource utilization rate and the task processing efficiency are improved.

The invention discloses a titanium alloy microscopic structure prediction method and system based on a conditional generative adversarial network and a storage medium, and belongs to the following steps: firstly, constructing a process-structure mapping model, and taking the output of the model as a rule constraint condition; inputting the random noise vector and the rule constraint condition into a conditional generative adversarial network to generate a prediction image; according to the generative adversarial network, thermal dynamic constraints based on physical quantities of microscopic structures are introduced in the training process, so that the interpretability of a prediction result is improved. And carrying out quantitative comparison on the predicted image and the real image, verifying the consistency of the statistical characteristics, and if the verification is passed, outputting a prediction result. According to the method, end-to-end prediction from process parameters to microscopic structure images is realized, the limitation that only symbolization or parameterization prediction can be carried out in a traditional method is broken through, and the intuition, the interpretability and the engineering application value of the method are remarkably enhanced.

The invention belongs to the field of multi-agent reinforcement learning, and discloses a relation-driven multi-agent reinforcement learning method and system based on a mixed game, and the method comprises the steps: enabling a strategy network to generate an agent action, and carrying out the interaction of an environment, so as to collect sample data; the centralized value evaluator calculates marginal influence values of the agents based on samples, and deduces a social influence weight set; assigning group external rewards as individual external social rewards based on the weights; processing the global state by using a random network distillation-driven method, updating a prediction network parameter to minimize a prediction error, and outputting an internal reward set; the external social rewards and the internal rewards are fused to form comprehensive rewards; and updating the strategy network and the value network by using the comprehensive reward, and circularly training until convergence. By adopting the method, accurate modeling of the individual interaction relationship is enhanced, invalid exploration is remarkably reduced, the learning efficiency and the strategy reliability are improved, and the overall performance of the system is enhanced through an explicit social mechanism.

The invention relates to the technical field of industrial automation and intelligent control, in particular to a general control system strategy optimization method based on reinforcement learning, and the method comprises the steps: firstly collecting the operation data of a system, constructing a state vector of reinforcement learning, and enabling a reinforcement learning agent to fully understand the current operation condition of the system; then, the state vector is input into a reinforcement learning strategy network, an action for adjusting a control strategy is generated by the network, and the action can be used for modifying the proportion, integral or differential coefficient of PID and can also be used for adjusting the prediction step length, weight coefficient or constraint strength of model prediction control, so that the adaptive capacity of a controller to external changes is enhanced; then, a reward signal is constructed according to a response result of the reference controller; the reward function comprehensively considers the error size, the steady-state characteristic, the system energy consumption, the control smoothness and the stability requirement, so that the reinforcement learning not only pays attention to the error minimization when optimizing the strategy, but also considers the low energy consumption, the smooth action and the anti-interference performance at the same time.

The invention relates to the field of computer technology and medical image processing, in particular to a medical image segmentation method fusing random region cutting enhancement and a pseudo label semi-supervised mechanism. In order to solve the problems of scarcity of annotation data, weak model generalization ability, inaccurate segmentation boundary and the like in a current medical image segmentation task, the invention provides a medical image segmentation method fusing random region cutting enhancement and a pseudo label semi-supervised mechanism, which is called an RCDE segmentation model. The model adopts a shared encoder and double decoder structure, combines a structure-level disturbance generation strategy, enhances the perception ability of the model for image structure change by mixing and recombining labeled images and unlabeled images, generates pseudo labels by utilizing a teacher network, and introduces a dynamic confidence coefficient screening mechanism, so that low-quality pseudo labels are effectively eliminated, and the robustness of the model is improved. The training stability and the pseudo-supervision effect are improved, preprocessing such as size normalization and image enhancement is carried out on the image, and the consistency and robustness of model input are improved.

The invention relates to the technical field of switch cabinet fault monitoring, and discloses a switch cabinet partial discharge intelligent monitoring system for fault identification. The system comprises a signal dynamic capture layer, a feature reconstruction mapping layer, a heterogeneous data collaboration layer and a self-adaptive diagnosis decision layer. The signal dynamic capture layer obtains a multi-source partial discharge signal based on a spatial perception mechanism, and generates a full-dimensional discharge intensity distribution spectrogram; the feature reconstruction mapping layer deploys a distributed electromagnetic sensing array according to a high-risk area, and generates an insulation defect three-dimensional positioning map through a directional detection pulse and phase analysis algorithm; the heterogeneous data collaboration layer establishes space-time association, and after time reference is aligned, comprehensive risk confidence is generated through a multi-channel fusion network; and the adaptive diagnosis decision layer converts the comprehensive risk confidence into an executable monitoring instruction set, and issues the executable monitoring instruction set to an edge computing unit through an industrial internet of things protocol. According to the system, omnibearing monitoring and accurate diagnosis of partial discharge of the switch cabinet are realized, and the fault identification and response capability is improved.

The invention provides a bioreactor aeration self-adaptive control method and system based on machine learning, and the method comprises the steps: collecting the water inlet flow, ammonia nitrogen concentration, dissolved oxygen concentration, water temperature and chemical oxygen demand data of a bioreactor in real time through a sensor, and carrying out the water balance verification and the Kalman filtering pretreatment containing a water temperature compensation item; generating an aeration instruction by adopting multi-model collaboration; outputting a reference aeration rate by a deep reinforcement learning DRL model; the self-adaptive PID controller dynamically adjusts the proportionality coefficient according to the dissolved oxygen deviation and calculates the correction amount; and the LSTM network generates a fusion weight, and combines the seasonal compensation factor and the safety coefficient to synthesize the final aeration rate. The system passes mutation rate constraint and nitrification efficiency verification, an ASM1 model is triggered to recheck when the deviation exceeds the standard, and after passing, an air blower with the precision of + / -2% executes an instruction and performs closed-loop feedback in a period smaller than or equal to 10 seconds. After the method is implemented, the aeration energy consumption is reduced to 0.38 kWh / m, the effluent ammonia nitrogen standard reaching rate is improved to 98.5%, the dissolved oxygen fluctuation range is reduced to + / -0.3 mg / L, and the impact load response time is shortened to be within 15 minutes.

The invention discloses a weak supervision image semantic segmentation system and method based on an attention mechanism, and the method comprises the following steps: obtaining to-be-segmented image data and an image-level label, and extracting a multi-level semantic feature map; fusing the multi-level feature map with the space and the channel dimension to generate an attention feature map; combining the attention feature map with an image-level label to obtain a target area positioning map; inputting the attention feature map and the target area positioning map into a double-edge reconstruction network to generate a high-quality pseudo-label map; carrying out progressive decoding structure convolution on the pseudo label graph and the multi-level semantic feature graph to generate a segmentation prediction graph; constructing a supervision signal based on the segmentation prediction map and the pseudo-label map to obtain a supervision training result; dynamically updating the supervised training result to generate an updated pseudo-label graph; and continuously iterating based on the updated pseudo-label graph until the loss function is converged, and outputting an image semantic segmentation result. According to the invention, weak supervision image semantic segmentation based on the attention mechanism is realized.

The invention discloses a multi-modal sentiment analysis method and system based on context enhancement and cross attention, and aims to solve the problem of low sentiment analysis precision caused by insufficient multi-modal feature fusion and insufficient context information utilization in the existing multi-modal sentiment analysis scheme. The system comprises a feature extraction module, an intra-modal context enhancement (ICE) module, a modal alignment and dynamic gating (GCU) weighting module, a time sequence-modal cross attention fusion (TMA) module (cross-modal depth fusion module), and a shared representation generation and multi-task parallel decoding module. The text features are subjected to deep context coding through a BERT model, and the audio and visual features are processed through multi-scale convolution and time sequence Transform; the ICE module enhances audio and visual features by capturing time sequence dependence in a single mode; the GCU module generates a gating weight moment by moment based on a GRU network driven by a global context, and dynamically weights the three modal features; the TMA module realizes deep interaction and fusion in time and modal dimensions through asymmetric time sequence-modal cross attention to generate a shared representation; and the decoding module executes sentiment regression, sentiment classification and modal reconstruction tasks in parallel. Through the structure, the precision and robustness of sentiment analysis in a complex scene are effectively improved.

The invention discloses a household electrical load optimization scheduling method and system based on an MAPPO algorithm. The method comprises the following steps: (1) collecting multi-source heterogeneous data from a user side and an external environment and carrying out preprocessing; (2) estimating the future equipment use probability of the user by collecting historical equipment use data of the user, and outputting an equipment behavior prediction vector; (3) acquiring real-time electricity price information, performing short-term electricity price trend prediction by using the time sequence prediction model based on a historical electricity price sequence, and outputting an electricity price prediction sequence; (4) generating a scheduling strategy by the MAPPO network based on the equipment behavior prediction vector and the electricity price prediction sequence, and outputting a corresponding scheduling result; and (5) sending a scheduling result to the home gateway, controlling the equipment, collecting an execution feedback result in real time, updating internal parameters of the MAPPO network, and realizing strategy iteration and adaptive adjustment of the MAPPO network.

The invention discloses a communication network rapid adaptive control strategy generation method based on meta reinforcement learning, and relates to the field of communication network adaptive control, and the method comprises the steps: collecting and processing a physical layer signal and a network layer index, and obtaining a standardized multi-dimensional time sequence matrix; inputting the standardized multi-dimensional time sequence matrix into a hybrid encoder formed by CNN-LSTM, dynamically weighting through a self-attention mechanism, and outputting an environment feature vector; inputting the environment feature vector into a neural differential equation network, generating a strategy parameter increment and updating a strategy network parameter; and constructing a dynamic reward function based on physical layer features in the environment feature vector, and performing detection to obtain an adjusted dynamic reward value. According to the method, continuous time evolution of strategy parameters is realized through a neural differential equation network, and a dual-channel ResNet-ODE architecture is combined with a gating fusion mechanism, so that the problem of oscillation caused by discrete updating of a traditional reinforcement learning strategy is solved.

The invention relates to a computing power resource dynamic scheduling method, device and equipment based on deep reinforcement learning and a medium thereof, and the method comprises the steps: constructing a joint state vector through real-time fusion of a network layer channel state and computing layer node load data, and driving a strategy network to generate transmission parameters and resource allocation actions of cooperative control; the code modulation parameters of the wireless transmission module and the computing resource proportion of the target node are synchronously configured in the execution layer, and dynamic task scheduling in the channel decay environment is achieved; a multi-target reward mechanism is designed to couple transmission bit error rate penalty, resource utilization efficiency and task timeliness evaluation indexes, and a reinforcement learning agent is guided to balance communication stability and computing power demand conflicts; according to the method, strategy network parameters are optimized through time difference errors, closed-loop feedback is formed in combination with channel state prediction and node load updating, the problems of network and calculation layer splitting decision, insufficient dynamic adaptability and multi-target optimization imbalance in the prior art are effectively solved, and the task scheduling success rate in the time-varying wireless environment is improved.

A set of images of a scene re received. Each image includes temporal data and spatial data relating to the scene. Based on the spatial data of each image, three-dimensional (3D) Gaussian splatting data is generated. The temporal data of each image and the 3D Gaussian splatting data are inputted to a neural network to generate spatial-temporal 3D Gaussian embeddings. Offset data based on the spatial-temporal 3D Gaussian embeddings is generated. The video of the scene is rendered based on the 3D Gaussian splatting data and the offset data, allowing for improved rendering of video of the scene.

A system is provided for dynamic analysis and verification of smart contracts. The system includes an autoencoder configured to preprocess smart contract code to reduce noise and highlight critical features; a capsule network configured to analyze the preprocessed smart contract code, capturing hierarchical relationships and dependencies within the code; a generative adversarial network (GAN) configured to generate optimal routing coefficients for the capsule network, enhancing the efficiency and accuracy of the analysis; and a blockchain-based platform for deploying and executing smart contracts, wherein the platform utilizes the capsule network to continuously monitor the smart contracts for anomalies during execution.

The invention provides a safety key scene enhancement generation method fusing traffic rule hard constraints and a dynamic feasible region, and relates to the technical field of automatic driving safety tests.The method comprises the steps that a multi-strategy initial data set is constructed, pre-training an optimal feasible value network of the automatic driving vehicle through offline reinforcement learning by utilizing the multi-strategy initial data set, and generating a dynamic feasible region; a traffic rule is represented in a formalized mode through linear sequential logic, and a rule hard constraint and feasibility guide dual-action mask mechanism is constructed; and controlling the confrontation background vehicle to generate a compliance confrontation scene by maximizing a rule-feasible region dual-guided confrontation objective function, feeding back the generated effective key scene to an initial data set, and realizing data-driven feasible region dynamic updating. According to the method, by combining rule formalization constraint and feasible region reachability analysis, the scene is ensured to be compliant and close to a system capability boundary.

The invention provides a spinning threading action optimization method and system based on path deep learning, and the method comprises the steps: firstly obtaining a continuous three-dimensional point cloud sequence of a yarn broken end region, yarn physical attributes and mechanical arm dynamic parameters, and then constructing a tension-moment constraint related to the tensile strength of yarn according to the physical attributes and the dynamic parameters, meanwhile, on the basis of the three-dimensional point cloud sequence, a motion acceleration constraint associated with the anti-jitter capacity and an obstacle avoidance safety distance constraint associated with the obstacle distribution density are constructed, and then a loss function of a pre-constructed mechanical arm path generation neural network is optimized according to the three kinds of dynamic constraints linked with a scene; according to the invention, the first success rate and the track stability of the threading action can be improved, the threading failure caused by snapping, throwing or collision is reduced, the threading time is shortened on the premise of ensuring the safety of yarns, and the reliability of spinning operation is improved.

The invention discloses a robot remote operation and maintenance task scheduling and optimization method based on artificial intelligence, and the method comprises the following steps: collecting data, executing data cleaning, formatting and feature extraction, and forming a standardized feature vector set; constructing a state space, an action space and a reward function; the method comprises the following steps of: inputting an improved Distributional SAC (Software Consensus Consensus) task scheduling model; constructing a task demand matrix and a robot state matrix; generating a plurality of subtask sets, and establishing a high-level scheduling strategy according to task priority weights; in the low-layer execution network, a subtask strategy network is trained, and an action instruction set is generated to control the robot to execute tasks; updating strategy parameters based on environment feedback; establishing a self-adaptive hierarchy switching mechanism; and a hierarchical parameter updating mechanism is adopted to complete task scheduling and optimization. The improved reinforcement learning model is adopted, intelligent scheduling of robot operation and maintenance tasks is achieved, and the method has the advantages of being efficient, self-adaptive and high in stability.

The invention discloses a crane operation parameter self-learning optimization control system oriented to complex working conditions, which is characterized in that a multi-dimensional physical quantity signal is processed through a hierarchical mixed data fusion module, and a working condition feature vector is generated by using a lightweight mode recognition network; the working condition identification and risk assessment module outputs a working condition type identifier and calculates a dynamic safe operation boundary; a physical model enhanced reinforcement learning decision engine is built in the parameter self-learning optimization module, a parameter adjustment instruction is generated through an actor network, and evaluation optimization is carried out through a reviewer network; the optimized parameter adjusting instruction is issued to a self-adaptive fuzzy motion controller to drive an execution mechanism; and after execution, collecting system response data as an empirical sample to be stored and used for periodically updating network parameters. Accurate perception and safety evaluation of complex working conditions are achieved, control parameters can be autonomously optimized through fusion learning and a physical model, and the self-adaptive control capacity, stability and safety of the crane in a changeable environment are improved.

The invention discloses a voltage regulation and control method and system suitable for a distributed energy storage photovoltaic transformer area of a power distribution network. The method comprises the following steps: establishing a transformer area active power distribution network model for carrying out voltage regulation and control by considering distributed photovoltaic, wind power, an energy storage system, a static var compensator and a load node; a voltage regulation problem is modeled as a Markov decision process, a state space comprises multi-source power information and energy storage state information, and an action space is controlled quantities of a static var compensator and an energy storage system; a Markov decision is solved by adopting a depth deterministic strategy gradient algorithm based on Actor-Critic dual networks, a control action is generated through a strategy network Actor, and a current strategy is evaluated through a value network Critic for training; and inputting the trained strategy network based on the current observation state, outputting an optimal energy storage active / reactive power output and SVC reactive power compensation control instruction, realizing online adjustment of node voltage, and realizing stable control of transformer area voltage.

The invention discloses a disturbance of consciousness stimulation regulation and control system and method fused with electroencephalogram connection recognition. The system comprises a simulated electroencephalogram signal data acquisition stage, a connection recognition analysis stage, a stimulation parameter optimization stage and an executable stimulation instruction conversion stage. The method has the following advantages and effects that a whole-electroencephalogram activity distribution diagram is generated by simulating an electroencephalogram signal data acquisition stage, a key connection area is identified, and a brain function connection map is generated by utilizing a function connection analysis network and a phase synchronization algorithm in a connection identification analysis stage; in the stimulation parameter optimization stage, space-time correlation between a whole electroencephalogram activity distribution map and a brain function connection map is established, a multi-objective optimization algorithm is adopted to generate an optimized stimulation parameter set through a fusion network, and finally, in the executable stimulation instruction conversion stage, the optimized stimulation parameter set is converted into an executable stimulation instruction based on a self-adaptive control model. Therefore, the accuracy and the self-adaptive capability of electroencephalogram signal stimulation regulation and control are remarkably improved.

The invention relates to the technical field of unmanned aerial vehicle flight path planning, in particular to a public route network generation method for a small unmanned aerial vehicle, which comprises the following steps: reading urban road network data, removing expressways and branches, and generating a two-dimensional route network; determining the lifting height according to the height of the tree, and generating a three-dimensional air route network; constructing a space matrix by using the surface feature passenger flow data and rasterizing a road network; rasterizing a coherent feasible region of a road network through morphological opening operation, and identifying an intersection and generating a topological structure by openCV; establishing a layered traffic line for the intersection to realize two-way traffic; the airspace utilization rate and the flight safety can be improved, and the problem of large-scale unmanned aerial vehicle path conflict is solved.

The invention relates to a large model fine tuning method, device and equipment of a power system, a storage medium and a program product, and relates to the field of power systems. Comprising the following steps: acquiring multi-modal time sequence data of a power system, inputting the multi-modal time sequence data into a pre-trained multi-modal large model, and acquiring a prediction sequence of each signal output by the multi-modal large model in a time dimension; scoring the importance of each signal at each time point based on the volatility index of the prediction sequence; screening the signal time sequence correlation points with the importance scores exceeding a preset threshold value; for each signal time sequence correlation point, calculating a deviation between a predicted value and a historical steady-state mean value, fusing deviation correlation characteristics of different signals at the same time point, and constructing a state vector; inputting the state vector into a reinforcement learning strategy network to generate a fine tuning action; and performing parameter updating on the corresponding model substructure in the multi-modal large model according to the fine tuning action. According to the method, the fine tuning efficiency can be improved, and the response efficiency and reliability of the multi-modal large model in a complex power scene are improved.

The invention provides an enterprise business-oriented flow node intelligent optimization and scheduling method and system, and relates to the technical field of business processing, and the method comprises the steps: obtaining a business target description and an executed flow step sequence, retrieving historical flow nodes based on depth vector embedding, and generating business intention features through a bidirectional long-short term memory network; and constructing context state representation to perform structure matching, and finally constructing a dynamic heterogeneous graph structure to generate a weight recommendation sequence. According to the invention, accurate recommendation and intelligent scheduling of the enterprise process are realized, and the process execution efficiency and the business objective achievement rate are improved.

The invention provides a text classification method and a deep learning model training method and device, and relates to the technical field of artificial intelligence, in particular to the technical field of deep learning, natural language processing, content security and intelligent search. According to the specific implementation scheme, the method comprises the steps of generating semantic features of an input text by using a shared base network; according to the semantic features and a splicing weight matrix of the plurality of tower networks, determining respective classification results of a plurality of text classification tasks respectively corresponding to the plurality of tower networks, the splicing weight matrix being obtained by splicing weight matrixes of the same processing layer of the plurality of tower networks; and according to respective classification results of the plurality of text classification tasks, determining a target classification result of the input text.

The invention discloses a multi-modal robot control method and system based on space-time fusion and a liquid neural network. The method comprises the following steps: collecting asynchronous time sequence data of a multi-source heterogeneous sensor, and establishing a unified reference clock and time grid; through a neural phase-locked loop and soft-DTW, learnable synchronization is realized, and an alignment sequence, confidence and a residual error are output; after modal feature coding, fusion features are formed through confidence-gated layered space-time cross attention; a nominal control instruction is generated by the liquid neural network of the self-adaptive time constant; and the command is optimized and corrected in combination with CLF / CBF and QP, and a safe and feasible control command is output. According to the method, uncertainty caused by asynchronization, shielding and the like is effectively inhibited, the path deviation and the collision rate are remarkably reduced, and high precision, high robustness and provable safety are kept in a dynamic unstructured environment.