52 results about "Perception training" patented technology

The invention provides a target recognition model reasoning optimization method and device, and the method comprises the steps: firstly carrying out the structural analysis and sensitivity evaluation of a pre-training model, extracting the structural features of each network layer, activating the distribution features, carrying out the quantitative sensitivity scoring, and constructing a data set reflecting the hierarchical features and fault-tolerant capability; and querying a quantitative configuration knowledge base based on the data set to generate a heterogeneous quantitative strategy. Layered low-bit quantization is executed according to the strategy, and a layered weighted loss function is introduced to carry out quantization perception training, so that precision loss caused by bit width compression is effectively compensated. According to the method, through hierarchical heterogeneous quantification, the model recognition precision is preserved to the maximum extent while high compression ratio and reasoning acceleration are achieved, and particularly, the performance of a high-sensitivity layer is protected. The generated heterogeneous quantitative model remarkably reduces memory occupation and power consumption, is suitable for an edge hardware platform with limited resources, forms a set of complete automatic process from analysis and configuration to training compensation, and has good universality and engineering practical value.

The invention belongs to the technical field of image processing, and discloses a miniature ultrasonic image reconstruction method, device and equipment based on deep learning. The method comprises the following steps: constructing a lightweight U-Net model by a processed signal; a mixed loss function and a quantitative perception optimization and compression lightweight U-Net model are adopted, the processed model is deployed on an FPGA, a layered mapping strategy and a double-buffering mechanism are adopted to execute reasoning, and a reconstructed image is output; an up-sampling strategy is dynamically selected through a closed-loop feedback super-resolution module to optimize a reconstructed image, the ultrasonic imaging quality is improved through a lightweight U-Net architecture, hardware collaborative optimization and an end-to-end imaging pipeline, and a depth separable convolution, multi-scale attention gating and mixed loss function optimization algorithm is adopted; the power consumption is reduced through FPGA accelerated reasoning, dynamic voltage frequency adjustment and on-chip memory optimization, the imaging effect is improved through quantitative perception training, and the system is high in resolution, low in power consumption, high in real-time performance and suitable for portable medical equipment.

The invention provides a federated learning method combining acutance perception training and knowledge distillation, and belongs to the technical field of privacy protection. The method comprises the steps that a central server initializes global model parameters and distributes the initial global model parameters to participating clients; after each participating client receives the global model parameters, acutance perception training is executed based on the local data set, and a local updated global model is obtained and serves as a teacher model; through a knowledge distillation method, on the basis of a local training set and in combination with a soft label of a teacher model, obtaining student model parameter variation; performing privacy protection operation on the parameter variation of the student model, and uploading the parameter variation to a central server; the central server aggregates the parameter variation uploaded by each participating client after privacy protection, and updates a global model; the updated global model is used as an initial global model of the next round of communication. According to the method, privacy and efficiency are balanced, sensitive information leakage can be prevented, and the practicability of a federal learning global model is maintained.

The invention discloses a perception practical training platform and method based on intelligent edge calculation and heterogeneous data fusion. The platform comprises a sensor module, an interaction and display module, a calculation module, a movable base and a supporting rack. The sensing practical training platform adopts a modular design and a standardized interface; the sensor module, the interaction and display module and the calculation module can be conveniently disassembled, assembled and replaced, and convenient indoor and outdoor movement can be realized through a movable base with a stopper; through a data fusion technology of a sensor module, in combination with a GPU hardware acceleration strategy and a time synchronization technology, targets around the practical training platform are accurately detected in real time based on multi-source data, and a sensing result of the intelligent networked automobile to the targets in the surrounding environment is displayed. According to the invention, edge calculation provides a more convenient extensible path, and powerful support is provided for intelligent traffic through data acquisition expansion calculation and communication capability of road end data acquisition equipment and cloud end data acquisition.

The invention discloses a device for evaluating bladder filling degree before radiotherapy based on perceptual training, and the device comprises a sensing module which is used for obtaining bladder data information; the bladder filling degree prediction module is used for predicting the bladder filling degree according to the bladder data information; the perception training module is used for generating a dynamic perception training plan according to the bladder filling degree predicted by the bladder filling degree prediction module, and the dynamic perception training plan is used for controlling the bladder filling degree of the patient to be kept on the preset bladder filling degree. The technical problem that an effective bladder filling degree self-evaluation and training means is lacked in the prior art is solved.

The invention relates to the technical field of artificial intelligence and neural networks, in particular to a multi-teacher knowledge distillation-based neural network fault perception training method, which comprises a plurality of teacher neural networks and a student neural network, and is characterized in that each neural network comprises a feature extraction sub-network and a classification sub-network; different fault ratios are injected into each teacher neural network, and the teacher model is trained to converge; respectively inputting the training samples into a plurality of teacher models and student models, and calculating loss between each teacher model and each student model; and respectively updating the distillation temperature and the weight of each teacher model according to the corresponding loss function values until the student models converge. According to the method, the technology of combining multi-teacher knowledge distillation and the entropy of the output value is adopted, so that the model has high accuracy under different degrees of noise, and the fault tolerance of the neural network is enhanced.

The invention belongs to the technical field of edge devices, and particularly relates to a model quantitative perception training system on edge devices, which comprises a data acquisition module used for acquiring data from various edge devices; the data preprocessing module is used for cleaning, normalizing and enhancing preprocessing operation on the collected data so as to improve the quality and diversity of the data and provide a better data basis for subsequent model training; and the quantitative perception training module is used for applying a quantitative perception training technology in a training process so as to simulate a quantization operation in forward propagation of the model, performing quantization and inverse quantization on the weight and the activation value, and calculating a gradient by adopting a gradient approximation method in back propagation so as to optimize model parameters. According to the method, through quantitative perception training, the weight and the activation value of the model can be converted into low-precision representation, and the calculation amount when the model runs on edge equipment is greatly reduced.

The invention relates to the technical field of embedded artificial intelligence, and discloses an arc discharge detection model training method and system based on quantitative perception training, and the method comprises the steps: constructing an initial floating point precision neural network model for DC arc discharge fault detection, and defining a plurality of quantization functions for the model to convert floating point parameters; a FakeQuant pseudo quantization strategy is self-defined, a linear calculation layer in an original model is replaced with a quantifiable calculation layer, and a quantization operator is created through the quantifiable calculation layer; performing layer-by-layer differentiation quantitative perception training on the replaced model based on a training data set, and simulating a forward calculation behavior of a low-precision fixed-point type in actual reasoning while keeping floating-point precision back propagation through a quantization operator, so that the weight of the model is adaptive to a quantization error; and converting the trained model parameters into a target low-precision fixed-point type, and generating an end-side deployable arc discharge detection model through model testing and computing power evaluation, thereby realizing high-fidelity conversion from a floating-point model to a low-precision fixed-point model.

The invention discloses a fault suppression type high-reliability CNN quantitative perception training method and system, and the method comprises the steps: calculating the occurrence probability of an expected single event upset event according to the type and scale of a CNN network layer; according to the expected single event upset event occurrence probability, random fault generation is carried out in the quantitative perception training simulation integer inference process; extracting a model fault propagation path, analyzing a fault propagation influence mechanism, performing a model fault-tolerant capability node test, finding out a sensitive node of which the model precision oscillation exceeds a preset threshold value, and performing targeted fault generation and training on the sensitive node; according to the model structure, layer-by-layer fault iteration training is carried out, convergence results are obtained, the training result of each convergence is used as a pre-training parameter of next simulation fault training until no sensitive node appears, and the quantification process of the CNN is completed. According to the method, the reliability and robustness of the model in a high-radiation environment are remarkably improved.

The application discloses a deep learning quantization perception training method and device, equipment and medium, and relates to the technical field of deep learning. The method comprises the following steps: loading a pre-trained deep learning network model, and solidifying weight parameters in the pre-trained deep learning network model to obtain a solidified network model; training preset trainable quantization factor parameters in the solidified network model to obtain a trained network model; loading the trained network model, and simultaneously training the trainable quantization factor parameters and the weight parameters in the trained network model to obtain a quantization perception trained model. The application can guarantee the accuracy of a deep learning network model quantized at a low bit, and improve the stability of quantization perception training of a lightweight network model and a pruned network model.

The invention discloses a pet voice recognition translation cloud training method and application, and relates to the technical field of voice recognition and machine learning, and the method comprises the steps of data set construction and preprocessing, model architecture establishment, two-stage training, lightweight optimization, incremental learning iteration, model evaluation and deployment, and application coverage of a multi-terminal scene. The emotion recognition accuracy is improved, the intention recognition accuracy is improved, knowledge distillation and quantitative perception training achieve model lightweight, and incremental learning enables iteration efficiency to be improved. According to the pet voice recognition translation cloud training method and the application, the problems of poor data adaptation, insufficient model optimization and low iteration efficiency of a traditional method are solved, the method is adaptive to various pets and complex environments, multi-terminal deployment is supported, an accurate human-pet communication solution is provided for scenes such as family pet raising and pet medical treatment, and the user experience is improved. And the method has high precision, practicability and expandability.

The invention provides a quantitative perception training method and device, equipment and a storage medium, and relates to the technical field of artificial intelligence, in particular to the technical fields of large models, deep learning, quantitative perception training and the like. According to the specific implementation scheme, quantization operation is performed on a fusion weight of a target linear layer of a pre-training model to obtain a quantization weight of the target linear layer; wherein the fusion weight is obtained by fusing the original weight of the target linear layer and the to-be-optimized weight of the low-rank adapter; and based on the quantization weight of the target linear layer, performing quantization perception training on the pre-training model so as to optimize the weight to be optimized under the condition that the original weight of the target linear layer is frozen.

This invention relates to the field of rehabilitation training technology, and discloses a method for determining amblyopia perception training data and amblyopia perception training visual targets. The method includes: acquiring test data of the target subject, wherein the test data is data obtained from a basic amblyopia perception test of the target subject; determining the initial contrast sensitivity data of the target subject based on the test data; acquiring reference contrast sensitivity data, wherein the reference contrast sensitivity data is obtained from amblyopia perception tests of healthy subjects of the same age as the target subject; determining the target subject's weak data based on the initial contrast sensitivity data and the reference contrast sensitivity data; and determining the target subject's training data based on the weak data. This invention focuses on targeted training in weak areas, reducing ineffective training in normal areas, saving a significant amount of unnecessary training time, and improving the efficiency and targeting of amblyopia perception training.

The inventive concept relates to a stimulus providing device that provides training stimuli for improving pain of a pain patient, a perception training method, and a computer program. The stimulus providing device for providing a training stimulus for improving pain includes one or more mounting units mounted on body portions of a user, and one or more stimulus providing units disposed in the mounting units and which provides stimuli to the body portions, the stimulus providing units are controlled by a control unit included in the stimulus providing device or a control unit physically separated from the stimulus providing device to interwork with the stimulus providing device and included in a terminal of the user, and the control unit controls at least one of a location, a range, and an intensity of the stimuli.

The invention relates to a few-sample visual model optimization system and method based on end-cloud collaboration. The system comprises edge equipment and a cloud server. The edge device collects an environment image of the current deployment environment and sends the environment image to the cloud server; the cloud server is configured to extract an image sample from a pre-training data set after receiving the environment image, and fuse the image sample with the environment image based on a block-level mixing mode to generate an environment perception training sample; adjusting and optimizing the backbone network based on an environment perception training sample and an optimization strategy including mixed perception contrast learning and a multi-task auxiliary target to generate a lightweight adjusting and optimizing module; sending the generated tuning module to the edge device; and the edge device loads the tuning module to perform a few-sample reasoning task. The system provided by the invention has the advantages of low training and communication overhead, no need of extra annotation information, remarkable tuning effect and the like, and can be widely applied to edge intelligent vision application with frequent environment change and limited samples.

The embodiment of the invention discloses an FPGA (Field Programmable Gate Array) model reasoning hardware adaptation method based on a Qwen-2 low-bit quantization technology. According to the embodiment of the invention, quantitative perception training is carried out on a Transform layer of a Qwen-2 model so as to keep reasoning precision; converting the quantized weight and activation value into a fixed-point format supported by the FPGA and generating a quantization parameter table; according to DSP and BRAM resource distribution of the FPGA, mapping low-bit matrix operation to a DSP unit, storing quantization parameters in the BRAM, and planning a task scheduling sequence; a parallel pipeline is designed for an attention layer and a feed-forward layer, low-bit storage optimization is carried out on key value cache, and dynamic quantization adjustment is carried out on an activation value; and finally, deploying to an FPGA for testing, and performing closed-loop optimization on a hardware mapping strategy or quantization granularity based on delay and resource data. According to the method, the model precision is effectively maintained while the storage and calculation overhead is remarkably reduced.

The invention discloses a generalized quantitative perception training method and system based on knowledge feature distillation, and belongs to the technical field of intelligent computing. The method comprises the following steps: constructing a full-precision floating-point teacher model and a mixed bit width quantification student model based on a neural network architecture and a target task scene; carrying out pre-training on the full-precision floating point teacher model; taking the pre-trained full-precision floating-point teacher model as a knowledge transmission subject and the mixed bit width quantification student model as a knowledge receiving subject to carry out quantification perception training; the quantitative perception training is integrated with an attention feature fusion module and a multi-level context knowledge module; integrating a knowledge feature distillation strategy in the quantitative perception training process to realize knowledge extraction, migration and fusion of multilayer features; and a knowledge distillation loss function is calculated, the knowledge distillation loss function is taken as an optimization target, training optimization of the mixed bit width quantification student model is completed, and universal quantification perception training is realized.

The invention relates to a visual model closed-loop enhancement method based on physical simulation and parameter perception, and the method comprises the steps: testing a failure mask of a reference visual model through a test set, extracting a physical feature vector of a physical parameter of the failure mask, carrying out the clustering of the physical feature vector of the failure mask, and carrying out the recognition of the physical feature vector of the failure mask; concluding a physical scene mode in which the reference visual model is invalid; generating physical perception training data in a targeted manner according to the obtained physical scene mode; carrying out enhanced training on the improved reference vision model by utilizing physical perception training data; adding a modulation parameter generation branch and a modulation module which are adaptive to the reference visual model to the reference visual model, wherein the modulation parameter generation branch generates adaptive modulation parameters according to physical parameters of a simulation image in the physical perception training data; and the modulation module modulates the image feature map of the reference visual model by using the adaptive modulation parameter, enhances the physical understanding ability of the model, and improves the model effect.

The application discloses a visual language model construction method and system for stroke image analysis, relates to the technical fields of artificial intelligence, computer vision and medical image processing, and comprises the following steps: constructing prompt words corresponding to modality labels through key points in a stroke image review guide; performing perception training on a general medical visual language large model through a first data set and the prompt words corresponding to the modality labels; constructing a second data set based on positive sample texts of multiple training samples and multiple negative sample texts; constructing a third data set based on first negative sample images and second negative sample images; training the general medical visual language large model after perception training through the second data set and the third data set, and obtaining a visual language model for stroke image analysis. The application reduces the incidence of cross-modal confusion and medical hallucinations, and ensures that the automatically generated structured radiology interpretation report has high reliability and safety required for actual clinical application.

The invention discloses a text-image pedestrian re-identification method oriented to semantic fuzziness and excessive decision confidence, and belongs to the technical field of pedestrian re-identification. The method mainly comprises the steps that based on a distributed feature alignment module, image and text modal features are coded into Gaussian distribution, a feature center and an intra-class semantic change range are jointly expressed through a mean value and a variance, and the modeling ability of cross-modal semantic difference and visual diversity is improved; the uncertainty penalty alignment module based on evidence deep learning explicitly models the uncertainty in the image-text matching process by introducing Dirichlet distribution and subjective logic theories, and applies adaptive penalty to a high-uncertainty matching relationship, so that the excessive confidence of the model on a mismatching result is inhibited. According to the method, a probability distribution-based feature expression mode and an uncertainty perception training mechanism are adopted, so that the matching accuracy and generalization ability of a cross-modal retrieval system in view angle change, semantic fuzziness and background interference scenes are remarkably improved.

The invention discloses an intelligent teaching aid for child perception training in nursing practical training of secondary job. The intelligent teaching aid comprises a base, function buttons, a label card and a recognition device. The base is provided with a containing cavity, a loudspeaker and an operation platform, a control system is arranged in the containing cavity, and the control system is electrically connected with the loudspeaker; the function button is arranged on the operation platform, and the side face of the function button is provided with an insertion groove. The number of the label cards is multiple, and the multiple label cards are respectively provided with different identification labels. And the identification device is arranged in the plugging groove and is in communication connection with the control system. In this way, durability and operation safety of the teaching aid in the environment where children frequently use the teaching aid are ensured, and a stable bearing platform is provided for all interaction functions. A modularized label card is adopted as a content carrier, training themes are allowed to be freely replaced and combined according to teaching requirements or child cognitive development stages, and the application range of the teaching aid is greatly widened, and the teaching life cycle of the teaching aid is greatly prolonged.

The present application discloses a quantization-aware training method and related apparatus, comprising: obtaining a trained neural network model; calculating the second-order derivative corresponding to each network layer in the neural network model; constructing a linear programming problem for the neural network model, wherein the linear programming problem is used to determine the target bit position corresponding to each network layer in the neural network model; quantizing each network layer according to the target bit position corresponding to each network layer to obtain a mixed-precision quantization model; and performing quantization-aware training on the mixed-precision quantization model based on the neural network model. The present application embodiment is advantageously used to improve the optimization time used to determine the mixed-precision bit position setting, and is advantageously advantageous for better performance adaptation to hardware deployment requirements such as power.

The application discloses to the technical field of edge device, specifically for model quantization perception training system on edge device, including: data acquisition module, for collecting data from various edge devices; data preprocessing module, for cleaning, normalizing, enhancing preprocessing operation to the collected data, to improve the quality and diversity of data, to provide better data basis for subsequent model training; quantization perception training module, for being responsible for applying quantization perception training technology in the training process, to simulate quantization operation in the forward propagation of the model, quantizing and dequantizing the weight and activation value, adopting gradient approximation method to calculate gradient in the back propagation, to optimize model parameters. The application can convert the weight and activation value of the model into low-precision representation through quantization perception training, greatly reducing the calculation amount of the model when running on the edge device.

The invention provides a large language model binarization compression method and system. The method comprises the following steps: carrying out post-training binarization compression processing on a large language model to obtain a 1-bit structured weight initial model; carrying out Hadamard normalization preprocessing on a weight matrix of each layer of the structured weight initial model; constructing a quantitative perception training initial point based on the normalized model; and based on the initial point, adopting a layer-by-layer freezing-unfreezing strategy to carry out quantitative perception training on the normalized model layer by layer, outputting a final stably trained 1-bit model, and completing trainable binarization compression processing of the large language model. The method can be applied to medical semantic annotation tasks, embedded deployment tasks without training corpora, mobile phone end NLP reasoning acceleration tasks and other tasks, the trainability and convergence speed of the low-bit model are remarkably improved, and the integrated compression method high in training efficiency and model precision is formed.

The invention relates to the technical field of deep learning, and discloses a hysteresis quantitative perception training method and device for a neural network, and a storage medium, and the method comprises the steps: training a pre-constructed neural network to obtain a floating point network meeting a first preset convergence condition; inserting a pseudo quantization node into the network, and inputting pre-acquired training data to carry out PTQ calibration to obtain an initial quantization network; performing QAT training on the initial quantization network according to a preset weight updating strategy; and if the model parameter meets a second preset convergence condition in the QAT training process, fixing a quantization parameter of the initial quantization network as a target quantization parameter, adding a hysteresis strategy in a weight updating strategy to obtain a weight hysteresis updating strategy, and performing QAT training on the initial quantization network according to the weight hysteresis updating strategy to obtain a target quantization parameter. And obtaining a target quantization network meeting a third preset convergence condition. Therefore, by implementing the method, the network precision can be improved, and the network training stability and reliability can be improved.

The invention relates to the technical field of rehabilitation training, and discloses an amblyopia perception training data determination method and an amblyopia perception training sighting mark, the method comprises the following steps: obtaining test data of a target object, the test data being data obtained by carrying out an amblyopia perception basic test on the target object; determining initial contrast sensitivity data of the target object based on the test data; obtaining reference contrast sensitivity data, wherein the reference contrast sensitivity data is obtained by performing an amblyopia perception test on a healthy object of the same age as the target object; determining weak data of the target object based on the initial contrast sensitivity data and the reference contrast sensitivity data; and based on the weak data, determining training data of the target object. The amblyopia perception training method focuses on a weak area to carry out key training, reduces invalid training in a normal area, saves a large amount of unnecessary training time, and improves the efficiency and pertinence of amblyopia perception training.

The invention provides a field-generalization-oriented gradient adaptive quantitative perception training method, device and equipment, relates to the technical field of computers, and aims to improve the generalization ability of a quantitative model in an unknown field. The method comprises the following steps: acquiring a training data set and a quantitative convolutional neural network model; inputting each sample image in the training data set into the quantitative convolutional neural network model for processing to obtain a task gradient and a flatness of each scale factor in different training domains; evaluating the gradient confusion degree of the task gradients of each scale factor in different training domains, wherein the gradient confusion degree represents the consistency degree of the task gradient direction; and according to the gradient confusion degree, determining a learning gradient of each scale factor in different training domains, and updating parameters of the quantized convolutional neural network model based on the learning gradient to obtain a trained quantized convolutional neural network model.

An image detection model training method and device, and a readable storage medium, the image detection model training method comprises: acquiring a preset image detection model; performing quantization perception training on the preset image detection model to obtain a teacher model; performing feature distillation on the teacher model to obtain a student model; the data feature distribution of the student model is consistent with the data feature distribution of the teacher model; and performing dynamic offline quantization on the student model to obtain a trained image detection model. By using the above scheme, the precision and speed of the image detection model can be improved.

The invention relates to a method, a system and equipment for compressing a speech synthesis model, and relates to the field of artificial intelligence. According to the technical scheme, the problem of traditional model deployment is solved, storage in an ONNX format occupies 112 MB originally, the hardware requirement is high, embedded equipment is difficult to bear, the model is only 30 MB after compression, reasoning is accelerated, and smooth operation can be achieved on edge equipment such as an ARM chip and a low-power-consumption MCU; moreover, the scheme is superior to a traditional compression means, solves the problem of proneness to voice distortion caused by static quantization and violent modification of a model structure, adopts quantization perception training to control errors, and ensures that voice details and tone quality are basically flush with those of an original model; in addition, according to the scheme, the training cost is reduced, 1000 epochs can reach the target on the basis of the original pre-training model, the GPU investment is reduced, the GPU cycle is shortened, and the threshold of enterprises and developers is reduced.

The invention provides a sensing method, device and system, a storage medium and a program product, in the method, a first device represented by terminal equipment or access network equipment and the like can determine first sensing measurement data based on a first signal used for sensing an environment; and sending perceptual training data to a second device represented by the perceptual network element, the perceptual training data including the first perceptual measurement data and the perceptual result of the environment, and the perceptual training data being used for training the perceptual neural network. Thus, the perception training data used for training the perception neural network is collected, and due to the fact that the perception training data is collected based on the perception measurement process, the number of times of information interaction in the system is not increased too much.