135results about How to "Accelerated training" patented technology

The invention belongs to the technical field of image classification neural networks, and discloses a lightweight image classification neural network architecture system based on knowledge distillation. The invention provides a lightweight image classification neural network architecture system based on knowledge distillation. The lightweight image classification neural network architecture system comprises a data set preprocessing module, a teacher model preprocessing module, a subnet model construction module, a personalized training module, a real loss calculation module, a distillation loss calculation module, a total loss calculation module and a student model construction module. The invention provides a lightweight neural network architecture system based on knowledge distillation, and designs a brand new training mode by using the superiority of knowledge distillation for teacher model weight extraction: the knowledge distillation of teacher and student models is combined with a deep compression method; the system for constructing the lightweight network can be used for training various complex image classification neural networks so as to improve the network performance.

The invention discloses a data communication method and system based on information security, and belongs to the technical field of data communication, and the data communication method comprises the stages of preparation, transmission, intrusion detection and defense, receiving verification and cleaning ending. The two parties authenticate and exchange keys during preparation; encrypting data by using a symmetric key and an AES algorithm; automatic defense is performed after intrusion detection; the receiver verifies the data; and cleaning the resources at the end, and performing log optimization on the system. According to the invention, through innovatively improving the intrusion detection and defense method, based on real-time network flow monitoring, intrusion analysis based on machine learning and an automatic defense mechanism, the detection accuracy and efficiency are greatly improved, and false alarms and missing alarms are reduced; defense capability and flexibility are enhanced, attacks are intelligently blocked and isolated, and cooperative emergency response is rapid; data communication security technology progress and expansion are promoted, and network security is powerfully guaranteed.

The invention provides a natural language control method for a humanoid robot. By constructing the simple skill library and the complex skill library of the robot, the dependence on a large amount of operation data in a specific field is effectively reduced. According to the method, a large language model is used as a base of a task planner, and the capabilities of scene learning, semantic understanding and common sense reasoning of the task planner are brought into full play. By introducing a visual model in the field of computer vision, high-level semantic information and low-level spatial information of the environment can be obtained, and an accurate scene graph is constructed to serve as environment state representation. According to the voice interaction interface constructed by the invention, the naturalness of man-machine interaction is improved, so that a user can more conveniently communicate with the robot and input instructions, the user experience is effectively improved, and popularization of the humanoid robot in an actual application scene is facilitated.

The invention discloses a high-resolution remote sensing image change detection method based on multi-scale convolution decoding, electronic equipment and a storage medium, and belongs to the technical field of remote sensing image change detection. In order to meet the high-resolution remote sensing image change detection requirement, the method comprises the following steps: constructing a remote sensing image data set; a first stage and a second stage are formed based on a CNN module, a third stage and a fourth stage are formed based on a MambaVision module and a Transform module, and four dual-time-phase multi-scale feature maps are respectively obtained through processing of the four stages; constructing a time feature interaction module TFIM; a multi-scale convolution decoder EMCAD is constructed; and constructing a loss function, carrying out training, testing and evaluation by using the constructed remote sensing image data set, and applying the obtained trained optimal model to high-resolution remote sensing change detection. According to the invention, the detection capability of tiny changes in a high-resolution image is improved.

The invention discloses a blood vessel segmentation method and device based on large model fine tuning, equipment and a medium, and the method comprises the steps: obtaining a blood vessel image, carrying out the preprocessing of the obtained blood vessel image, and obtaining a training set; performing fine tuning on a pre-trained image segmentation model by adopting the training set, so that the image segmentation model learns related features of blood vessel segmentation; wherein the low-rank adaptive fine tuning technology is adopted to perform fine tuning on the image segmentation model: two parallel low-rank matrixes are inserted into a Transform layer in the image encoder, the input is processed by the two low-rank matrixes in sequence, and the processing result and the original output are added to obtain the final output. According to the method, the defects of a large model in a blood vessel segmentation task are overcome through efficient parameter fine tuning and a multi-scale feature fusion module. Accurate blood vessel segmentation can be realized under limited annotation data, the difficulty and cost of medical image segmentation tasks are greatly reduced, and the method can be widely applied to the technical field of medical image processing.

The invention relates to an underwater acoustic target recognition system and method based on multi-modal depth feature fusion, and belongs to the field of underwater acoustic target recognition. The method comprises the following steps: firstly, performing preprocessing on an obtained underwater acoustic target original audio and associated metadata, and constructing a multi-modal data set; and then the constructed multi-modal sample is input into an identification model, the model extracts deep representation of each modal through a multi-branch feature coding network, depth alignment and complementary aggregation of different modal features are realized by using a cross-modal cross attention mechanism guided by potential query, and a target identification result is output based on a decision network of mixed experts. According to the method, experimental verification is carried out on two disclosed underwater acoustic data sets, the experimental result verifies the effectiveness of the multi-modal deep fusion and hybrid expert adaptive decision strategy adopted by the method, and through mining the complementary advantages of acoustic features and semantic priori, the multi-modal deep fusion and hybrid expert adaptive decision strategy is obtained. And the robustness and generalization ability of the underwater acoustic target recognition system in the strong-noise and multi-working-condition environment are remarkably improved.

The invention discloses a non-invasive blood pressure prediction method based on signal dimension extension and a lightweight network, which comprises the following steps: firstly collecting PPG blood pressure data and ABP data, then denoising the collected PPG blood pressure data and ABP data, and cleaning the denoised data. The processed data is used for data signal dimension expansion, a PPG blood pressure measurement model is constructed, and the PPG blood pressure measurement model comprises a basic convolution layer and a residual grouping convolution module Group ResidualBock; and then deploying the constructed PPG blood pressure measurement model into the wearable device to obtain the wearable blood pressure monitoring device. And finally, measuring the processed PPG signal by using wearable blood pressure monitoring equipment to obtain a blood pressure value. According to the method, the PPG signal dimension is expanded, the data processing link is optimized, and the model blood pressure prediction accuracy is improved; and meanwhile, a lightweight algorithm architecture is adopted, so that resource occupation is reduced, the system can be deployed in wearable embedded equipment, and the security of personal privacy data is guaranteed.

The invention relates to a medical image processing technology, and discloses a pan cancer species histopathologic full-slice classification method and device, which is a trained multi-modal pan cancer species classification model based on histopathologic full-slice image data and full-slice diagnosis report pairing data of various tissue parts and various cancer species, and has the advantages of wide application range and high efficiency. And full slices of various pathological tissues can be classified. For classification tasks of different tissue parts, only text prompts (prompts) are used, i.e., a candidate cancer species list is set, and zero-sample full-slice classification can be realized without training. And for rare cancer species, full-slice classification can be carried out. Besides, the model has high expansibility, in order to better adapt to downstream tasks, better classification performance can be achieved by combining slice embedding with a linear classifier for use, freezing a slice embedding part and only finely adjusting the linear classifier, and the requirement for the training data size is low.

According to the model training-oriented gradient data transmission method and the related equipment provided by the invention, the compression ratio is dynamically determined according to the change of the transmission rate by monitoring the transmission rate in real time, so that the gradient data can be intelligently compressed according to the compression ratio, the data transmission quantity is remarkably reduced, and the network burden is relieved; and meanwhile, the transmission rate of each working end is adjusted through congestion control according to the bandwidth of a transmission link between the switch and the parameter server so as to relieve congestion. Under the combined action of the two, the transmission time of the gradient data is remarkably shortened, and the data transmission efficiency is improved, so that the training process of the distributed artificial intelligence model is accelerated, and the overall training efficiency is improved.

The invention is applicable to the technical field of knowledge migration, and provides a graph domain adaptive node classification method based on instance learning, which comprises the following steps of: acquiring a source domain data set with a mark and a target domain data set without a mark; the representation of the source domain data set and the representation of the target domain data set are learned through a GNN model, modeling is conducted on the source domain data set and the target domain data set, and embedded vectors of the source domain data set and the target domain data set are obtained; the embedded vectors of the target domain data set are classified through KNN, and target domain category pseudo labels are obtained; performing fine-grained instance contrast learning on embedded vectors of the source domain graph data set and the target domain graph data set; and after classification is finished, displaying a final classification result of the target domain graph data set. According to the method, the defects of an existing graph domain adaptation technology in cross-domain tasks are overcome, and the method has wide applicability and can be applied to graph data analysis tasks in various cross-domain scenes, such as social network analysis, recommendation systems, knowledge graphs and biological networks.

The invention discloses a COD (Chemical Oxygen Demand) prediction method based on a two-channel convolution time sequence mechanism, which comprises the following steps of: acquiring COD original data for preprocessing to obtain COD input characteristics; constructing a prediction model, and importing data into a dual-channel convolution module; importing the first double-channel convolution output feature vector and the second double-channel convolution output feature vector into a double attention memory network to obtain a time sequence dependence feature; the time sequence dependence characteristics are imported into an output module, and predicted COD output is obtained; training the model, and in the training process, optimizing hyper-parameters of the model based on an improved pelican algorithm; and importing real-time COD original data into the trained model, and obtaining real-time COD output. According to the method, a high-efficiency effluent COD content prediction model of the sewage treatment plant is constructed by creatively fusing the two-channel convolution and the double-attention memory network.

The invention provides a high-precision attitude estimation method for processing limited data based on a pre-training diffusion model, and the method comprises the steps: carrying out the preprocessing of an input image, and calling a function to combine with noise to generate a key point prompt; mapping the input image through an encoder to obtain a potential code, and increasing Gaussian noise of different amplitudes on the potential code to obtain a noise image; performing attention mechanism processing on the noise image to obtain a semantic corresponding point between the input image and the reference image; training the image pair, integrating training results of the basic class image and the new class image, and inputting an integration result into an Adam optimizer for optimization to obtain an optimization result; evaluating the image pair; according to the method, the key points of the unlabeled image data can be learned in a cross-category mode, a good attitude estimation effect is achieved under the condition that the data is limited, and the actual application value of attitude estimation is improved.

The invention relates to a warehouse logistics scheduling method and system based on cloud computing, and relates to the technical field of warehouse logistics scheduling, and the method comprises the steps: jointly building a logistics scheduling model through a plurality of edge cloud computing nodes, and the logistics scheduling model comprises an edge scheduling scheme generation model and a comprehensive scheduling scheme generation model; determining at least one first optimal edge cloud computing node based on the computing power resource demand of the warehouse logistics scheduling task; each first optimal edge cloud computing node generates a model through an edge scheduling scheme, generates a plurality of candidate warehouse logistics scheduling schemes based on the warehouse logistics scheduling task, the warehouse node state information and the logistics vehicle state information, screens the plurality of candidate warehouse logistics scheduling schemes, and determines a plurality of target warehouse logistics scheduling schemes; through the comprehensive scheduling scheme generation model, the optimal warehouse logistics scheduling scheme is generated based on the multiple target warehouse logistics scheduling schemes, warehouse logistics scheduling is carried out, and the method has the advantage of improving the intelligent level of warehouse logistics scheduling.

The invention provides a virtual power plant baseline load measurement and calculation method based on multi-dimensional factors, and the method specifically comprises the steps: screening load prediction correlation factors based on historical demand response information, and determining the correlation degree between each load prediction correlation factor and a load demand; setting a weight coefficient of each load prediction association factor based on the association degree, and constructing a prediction training data set according to the corresponding weight coefficient; an LSTM model is constructed, training is performed through the prediction training data set, model parameters are optimized through a sparrow optimization algorithm, and a baseline load measurement and calculation model is obtained; and calculating the baseline load of the virtual power plant in the target time period according to the baseline load measurement and calculation model. Baseline load prediction can be performed through the LSTM model, and parameter optimization is performed by using a sparrow optimization algorithm in the model training process so as to efficiently and accurately set LSTM network parameters, and at the same time, the measurement and calculation precision and the measurement and calculation efficiency of the baseline load are ensured.

The invention provides a motion control method of a curling robot. According to the method, the position information of the curling ball is accurately obtained through the deep learning technology, the self-adaptive optimization algorithm is combined, the throwing strategy is dynamically generated based on the real-time scene, and the control precision and the strategy decision-making capacity of the curling robot are improved. According to the method, the target position obtained through the optimization process has the higher optimality probability, the training and competition ability of the curling robot can be remarkably improved, a foundation is laid for precise throwing of the curling robot, more effective throwing strategy learning is provided for curling players, the curling players are helped to throw curling balls with the higher score rate in the competition, and the experience of the curling robot is improved. The curling sports development is promoted.

The invention relates to the technical field of AI visual detection, and discloses a glass fiber filament flying and breaking early warning and disposal method based on AI visual analysis, which comprises the following steps: collecting a normal operation video of a coil base, processing the video according to frames to form a plurality of pictures, manually marking the pictures containing broken filaments, and training a model by using the marked pictures to obtain a broken filament early warning and disposal result; the coil base is detected in real time through the model, when it is detected that the number of broken wires reaches a threshold value, the audible and visual alarm of the coil base where the broken wires are located is started, alarm information of the coil base where the broken wires are located is obtained, the alarm information is processed through a definition rule, a hexadecimal instruction is generated, and the instruction is sent to a corresponding worker. And meanwhile, the work number information of the worker is judged through an optical character recognition model.

The invention relates to a credit distribution method in multi-agent cooperation, and the method comprises the following steps: obtaining joint data of a multi-agent system, the joint data comprising observation information and execution actions of each agent; the joint data are input into a discriminator model, and the discriminator model carries out modeling on an interaction dependency relationship among multiple agents based on an interaction dependency relationship modeling module of an attention mechanism; generating an auxiliary reward signal corresponding to each agent through a discriminator model; generating a fused reward signal based on the auxiliary reward signal and a global reward signal from the environment; and training and updating the multi-agent strategy model by using the fused reward signal so as to optimize the cooperation strategy of the multi-agent system. By introducing a credit distribution mechanism based on a discriminator model, automatic decomposition and optimization generation of individual reward signals are realized, so that the learning efficiency and stability of a multi-agent system in a complex cooperative task are remarkably improved.

The invention belongs to the technical field of power distribution networks, and particularly relates to a power distribution network load transfer method and system based on data knowledge driving. The method comprises the following steps: constructing a mathematical model of a power distribution network load transfer problem; constructing a multi-agent space according to the mathematical model; combining the multi-agent space with the mathematical model to obtain a multi-agent initial model; obtaining a teaching experience library, and training the multi-agent initial model in combination with a learning algorithm to obtain a pre-trained multi-agent model; and generating a load transfer strategy by using the pre-trained multi-agent model. The load transfer decision can be generated according to the observation information through the multi-agent system, generation of the load transfer decision is assisted, and safe operation of the power system is ensured.

The invention discloses an outdoor scene-oriented efficient multi-view three-dimensional reconstruction method and system, and belongs to the field of multi-view three-dimensional reconstruction, and the method specifically comprises the steps: preparing a multi-view three-dimensional reconstruction data set for model training; constructing a multi-view three-dimensional reconstruction network comprising an adaptive feature extraction converter module based on big kernel attention, a cost regularization module and a depth estimation module; performing training and fine tuning on the network model to obtain a final model; building a multi-view three-dimensional reconstruction system and equipment; according to the method, the problems of poor accuracy and low integrity of a reconstruction result when an existing multi-view three-dimensional reconstruction method faces an outdoor large-scale scene can be solved, and the training reasoning time and the calculation cost of multi-view three-dimensional reconstruction can be reduced; and an important foundation is laid for application of outdoor large-scale scene image data in the field of multi-view three-dimensional reconstruction and development of a three-dimensional reconstruction technology.

The invention discloses a viewpoint planning method based on deep reinforcement learning. The method comprises the following steps: acquiring a sheltered image and a camera attitude and obtaining state information; inputting the state information into a trained DQN network to obtain a camera motion action at the next moment; and controlling the mechanical arm to move to adjust the camera position. In order to improve the training efficiency of the DQN network, a training strategy of global trend guide learning is creatively provided. The global trend guided learning strategy refers to a method for forming a global trend by recording viewpoint areas of successfully detected fruit stems in the same scene in a learning process, and then guiding a network learning process by using the global trend. According to the invention, viewpoint planning is realized by a brand-new technical means, so that the problem of shielding of the grape picking robot in the picking operation process is solved. According to the method, the viewpoint planning time is greatly shortened under the condition that the high picking success rate is kept, and a new technical scheme is provided for the field of active vision technology application.

The invention relates to the technical field of computer vision and mechanical arm automation, and discloses a mechanical arm control method and system based on improved Swinin-Unet. Comprising the following steps: S0, acquiring an RGB color image and a depth image of a target working scene; s1, converting the depth image into a normal vector angle image, and splicing the depth image and the normal vector angle image to obtain a space image; s2, analyzing the RGB color image and the space image by using a grabbing model to obtain grabbing parameters; the capturing model is constructed based on improved Swinin-Unet, and comprises a coding part which is used for respectively extracting features of a plurality of set scales of the RGB color image and the space image and carrying out feature fusion; in the decoding part, the fusion features of the minimum scale are sampled upwards step by step to the input resolution, the fusion features of the corresponding scales are fused through jump connection, and then capture parameters are obtained through prediction; and S3, the mechanical arm is controlled to grab the target object based on the grabbing parameters. According to the method, the position information of the target object can be accurately judged in a multi-object disordered scene, and grabbing is completed.

The invention belongs to the technical field of cooking fume emission supervision, and particularly relates to an online cooking fume concentration monitoring system based on the Internet of Things, which comprises a cooking fume emission port monitoring module, an emission harmfulness analysis module, an emission port control classification module, a regional emission evaluation module and a background monitoring end, according to the invention, the emission harmfulness analysis module judges the oil smoke emission harmfulness of the corresponding oil smoke emission port in real time based on the oil smoke emission information, and the emission port control classification module analyzes the emission pollution performance of the corresponding oil smoke emission port in the detection period to realize the grade division of the emission port, so that the subsequent formulation of matched management and control measures is facilitated; the area emission evaluation module accurately analyzes the oil smoke emission management performance of all areas needing oil smoke emission supervision one by one, subsequently strengthens the investigation of the areas corresponding to key investigation signals, supervises and urges the improvement of oil smoke emission, and significantly reduces the management difficulty of managers. The atmospheric environment can be protected; and the living quality of residents can be improved.

The invention relates to a temperature and strain bimodal flexible sensor and a preparation method thereof, and the sensor achieves the efficient detection of temperature and strain signals through the combination of an ink direct writing process and a screen printing process. The sensor is unique in design and comprises a rigid area substrate and a soft area substrate which are respectively provided with a thermal resistance type temperature sensor and a piezoresistive strain sensor, so that effective separation and accurate measurement of temperature and strain signals are ensured. A flexible material and an improved FA-BP neural network algorithm based on a firefly algorithm are adopted, so that the stability and the measurement precision in a variable environment are improved. In addition, the sensor can still work normally in a high-temperature environment, and is superior to a common flexible thin film material in the market. The sensor disclosed by the invention shows huge application potential in the fields of intelligent wearable equipment, health monitoring, industrial automation and the like, and a new solution is provided for realizing high-precision and high-stability temperature and strain monitoring.

The application discloses a low-latency computing task unloading method and system based on an artificial fish school algorithm, and comprises the following steps: establishing a MEC system for unloading computing tasks of multiple user devices to MEC servers; the MEC system comprises K MEC servers and N user devices; the user device computing task is completed through the pre-optimized MEC system, and the process comprises the following steps: dividing the computing task of each user device into K+1 subtasks, compressing part of the data of the subtasks by using a data compression method to obtain compressed data, dividing the subtasks into compressed data and uncompressed data, transmitting the K subtasks of the user device to each MEC server in a non-orthogonal multiple access mode, and respectively calculating the K+1 subtasks of the user device and each MEC server, and returning the calculation results of the MEC servers to the user device; the maximum user computing task completion latency is significantly reduced, and the user experience is improved.

The present invention belongs to the technical field of automated target detection, and discloses an unmanned aerial vehicle target detection system, a target detection method, and a computer-readable storage medium. The unmanned aerial vehicle target detection system includes an unmanned aerial vehicle system for carrying an optoelectronic payload system to an observation position and hovering at a fixed point; an optoelectronic payload system for performing fixed-point observation tasks and performing fixed observation of a target range area; a wireless image transmission system for transmitting observation video images using an HHLM type transmission module; a wireless remote control system for controlling the unmanned aerial vehicle system and the optoelectronic payload system, and a target detection integrated task station for receiving, storing, forwarding, and processing video images. The present invention provides a marine target detection system, which is of great significance to improving the level of testing technology and testing equipment, and to improving the actual combat capability and training level of destroyer and frigate naval gun firing. At the same time, it has a strong practical significance for ensuring the quality of naval gun live-fire firing and adapting to battlefield needs.

This invention relates to an intelligent decision support system for rectal cancer prognosis integrating digital pathological images, belonging to the field of medical image processing technology. The system includes a data acquisition module for collecting multiple sets of sample data; a feature selection module for identifying multiple key medical imaging features of rectal cancer based on the multiple sets of sample data; a model building module for acquiring a pre-trained deep learning model and constructing a rectal cancer prognosis prediction model based on transfer learning, the pre-trained deep learning model, the multiple sets of sample data, and the multiple key medical imaging features of rectal cancer; an image acquisition module for acquiring pre- and post-operative medical images of the rectal cancer patient to be evaluated; and a recurrence prediction module for predicting the recurrence probability of the rectal cancer patient to be evaluated based on the pre- and post-operative medical images of the patient using the rectal cancer prognosis prediction model. This system has the advantage of improving the accuracy of non-invasive prediction of postoperative recurrence of rectal cancer.

According to the training method for the table understanding ability of the large language model, effective information of the table can be compressed, meanwhile, row and column information capable of improving the effect is reserved, the effect can be improved at the same time, the effect improvement is embodied in the fact that the same table content is expressed in the form, the number of required tokens is generally small, and the method is easy to implement. The method has the advantages that the number of tokens participating in training and reasoning is smaller, computing power is utilized more efficiently, three major capabilities, namely table reading, numerical reasoning and data analysis, of table understanding are considered, the generation probability of correct answers is further increased, and generation of wrong answers caused by the probability is reduced. According to the method, the accuracy of table understanding is improved, and the information density after token conversion is effectively represented, optimized and improved through the table data, so that the efficiency of table data training and reasoning is improved.

The invention discloses a few-channel motor imagery electroencephalogram signal recognition method and device and a storage medium. Few-channel motor imagery electroencephalogram signals are obtained and filtered through a second-order Butterworth low-pass filter, and continuous wavelet transform is conducted on the filtered signals to generate a time-frequency graph; connecting the time-frequency diagrams of different channels in channel dimensions to form a channel-dependent multi-layer time-frequency diagram, and dividing the channel-dependent multi-layer time-frequency diagram into a training verification set and a test set; the method comprises the following steps of: pre-training an OfficientNet-B0 network, then deleting a full connection layer of the pre-trained network, adding a channel-dependent multi-layer time-frequency graph classification head, and carrying out transfer learning on a training verification set to obtain a final classifier for motor imagery electroencephalogram signal recognition. According to the method, the time-frequency diagram of the motor imagery electroencephalogram signals is generated through continuous wavelet transformation, the channel-dependent multi-layer time-frequency diagram is designed, the time and frequency characteristics of the motor imagery electroencephalogram signals are reserved, and the recognition accuracy of the motor imagery electroencephalogram signals is improved under the limitation of few channels.

The invention discloses an intelligent robot end-to-end motion planning method. The method comprises the following steps: constructing a simulation environment for extraterrestrial detection; the simulation environment is used for simulating motion characteristics of the robot in the extraterrestrial detection environment and wheel soil contact dynamic characteristics, and outputting sensing information of the robot based on a received action instruction; an end-to-end motion planning model is constructed, and the motion planning model takes perception information of the robot as input and takes the action instruction as output; the motion planning model is provided with a reward function, and the reward function is fused with artificial priori knowledge and actual engineering constraints; training the motion planning model based on a simulation environment to obtain a trained model; and in response to deployment of the trained model to an upper computer of the extraterrestrial exploration robot, enabling the robot to output a motion planning result based on real sensing information measured by a sensor of the robot. According to the invention, computing resources are saved, and the end-to-end motion planning capability of the robot under the constraint of limited resources can be improved.

The invention relates to the technical field of aviation, in particular to a device and method for remotely controlling a flap to retract and release in a rear cabin. The flap teaching and training efficiency and effect are improved. Structurally, pneumatic electromagnetic valves are additionally arranged on corresponding pipelines for air supply, retraction and release of a rear cabin flap switch; the controller is composed of a voltage supply power source, a valve control assembly, a control program and the like and is in electric signal connection with each electromagnetic valve. The invention provides a device and a method for remotely controlling a flap folding and unfolding function of a rear cabin when a rear cabin flap switch handle is in a neutral state. The original control function of the front cabin flap and the rear cabin flap is reserved, the control function of the rear cabin flap in the machine can be achieved through ground remote control, and the teaching and training efficiency and effect are improved.