356 results about "Counter propagation" patented technology

The invention relates to a robot decision control method based on gradient rarefaction and a robot. The method comprises the following steps: acquiring multi-modal sensor data of a robot; based on a preset sparsification strategy, generating a dynamic mask corresponding to the gradient matrix of the multi-modal large model; based on the generated dynamic mask, screening an effective gradient in a back propagation process of the dynamic mask; updating parameters corresponding to the effective gradient in real time, and obtaining the output of the multi-modal large model based on the updated parameters; according to the obtained multi-modal sensor data and the output of the multi-modal large model based on the updated parameters, feature fusion is carried out, and a combined state code including an environment state, a robot body state and historical decision information is generated; and according to the determined joint state code and based on a time sequence model, generating an action sequence, a force control parameter and a path planning dynamic decision instruction of the robot, so that the robot can act based on the generated dynamic decision instruction, thereby realizing decision control of the robot.

The invention relates to a vision and GNSS fusion displacement monitoring method and system based on deep learning. The method comprises the following steps: respectively acquiring visual data and GNSS data, preprocessing the visual data and the GNSS data, and synchronizing the visual data and the GNSS data to obtain synchronized visual data and synchronized GNSS data; performing feature extraction and combination to obtain a modal combination feature sequence; constructing a dynamic weight adaptive fusion network; and training the dynamic weight adaptive fusion network to obtain a displacement monitoring model, and inputting the modal joint feature sequence into the displacement monitoring model to complete displacement monitoring. According to the invention, through a feature extraction network of joint learning, time sequence correlation modeling of vision and GNSS in the same embedding space is realized. And dynamic self-adjustment of the modal weight is realized through a dynamic weight adaptive fusion network. All modules can be micro, and end-to-end back propagation training is supported. And when the visual signal is interfered or the GNSS is interrupted for a short time, the system can carry out self-adaptive compensation, and continuous and stable displacement output is kept.

The invention provides a training method and device for a VLA model, and relates to the technical field of intelligent robots with bodies, and the method comprises the steps that the VLA model outputs a joint angle sequence of a mechanical arm of a robot with a body based on visual input and language input; the differentiable forward kinematics module outputs a real-time pose of an end effector of the mechanical arm in a task space; based on joint space loss formed by the joint angle sequence and the joint angle in the teaching data, task space loss formed by the real-time pose and the tail end pose in the teaching data, and task space constraint loss, a multi-objective loss function is constructed, and total loss is calculated; calculating the gradient of the total loss to the joint angle sequence and the real-time pose through back propagation, and optimizing preset parameters of the VLA model; the above steps are repeatedly executed until the total loss meets the expectation, and a trained target VLA model is obtained; the technical problems that when a VLA model is trained in a joint space, the hardware coupling performance is high, and the generalization ability is insufficient are solved.

The invention discloses a crystal oscillator circuit fault classification technology based on a multi-source information fusion network, and belongs to the technical field of analog circuit fault diagnosis. Firstly, a multi-source data set of different measurement points of the crystal oscillator circuit is acquired; carrying out conversion from a time domain to a frequency domain on the data set by utilizing fast Fourier transform; extracting data fault features by using a convolutional neural network, and performing a trust distribution function under each piece of source data by using a softmax classifier; fusing different trust distribution functions under the multi-source data by adopting a D-S evidence theory to obtain a final diagnosis result and diagnosis probability output, and calculating cross entropy loss; and finally, training model parameters through back propagation to obtain a final diagnosis model. According to the method, the time-frequency transformation algorithm, the deep neural network algorithm and the information fusion algorithm are combined, a multi-source information fusion network is constructed, the defects of an existing diagnosis model in crystal oscillator circuit fault classification are overcome, and the accuracy and stability of crystal oscillator circuit fault diagnosis are remarkably improved.

The invention discloses a multi-modal data feature representation optimization method based on comparative learning and two-stage masks. The method comprises the following steps of: firstly, aiming at an input single image-text pair, generating two groups of heterogeneous image-text data with mode missing as two inputs of a model by applying two random mask strategies with different mask areas and proportions; wherein the first group of data applies a high-proportion mask to the image and applies a low-proportion mask to the text; the second group applies a low-scale mask to the image and a high-scale mask to the text. Then, the two groups of data respectively pass through an image encoder and a text encoder which share parameters, and two different multi-modal fusion feature vectors are generated through a cross-modal fusion encoder; according to the method, masked modal information is recovered through a decoder, and the reconstruction loss of the difference between a recovery result and original data is calculated. Meanwhile, two multi-modal fusion feature vectors generated twice are subjected to comparative learning, and the comparative loss of feature representation distances under different enhanced views for approaching the same image-text pair is calculated. And finally, performing weighted summation on the reconstruction loss and the comparison loss to form an overall loss function, and optimizing model parameters through back propagation. When the trained encoder is used for a downstream multi-modal classification task, the classification precision and generalization ability of the model can be effectively improved.

The invention belongs to the technical field of target detection, and particularly relates to a teacher tensor knowledge distillation-driven anti-violation target detection method in an electric power safety supervision scene, and the method comprises the following steps: inputting image data to a student model and a pre-trained teacher model; performing forward propagation on the student model to obtain a student prediction tensor; associating differences between the real tags based on the student prediction tensor and the input image data; acquiring an original teacher output tensor by using the same input image data by calling a bottom layer forward propagation method of a teacher model; calculating knowledge distillation loss based on the difference between the student prediction tensor and the original teacher output tensor; the standard detection loss and the knowledge distillation loss are combined to form total loss; and performing back propagation updating on parameters of the student model based on the total loss so as to complete model training and real-time target detection and detection method updating optimization. According to the invention, the performance of the student model can be significantly improved, and the high efficiency of the model is maintained.

The invention relates to the technical field of optics and the like, in particular to a coupled system parameter prediction method based on a physical-data co-driven deep neural network CMT-NN. The method comprises the following steps: taking optical response spectral lines of a double-micro-ring coupling system under different physical parameters as the input of a CMT-NN model, and taking the output as the physical parameters of a coupling resonance system; the step of constructing the trained CMT-NN model specifically comprises the following steps: S1, constructing a physical model of a target coupling system; s2, constructing a CMT-NN model, bringing physical constraints and physical parameters of the coupling model into a loss function of CMT-NN at the same time, calculating physical loss in a back propagation process through the loss function through iterative training and back propagation, and if a preset condition is met, completing training to obtain the CMT-NN model for predicting the physical parameters of the coupling resonance system; and S3, verification of the CMT-NN model is completed, and a trained CMT-NN model is obtained.

The invention relates to the field of constructional engineering, and discloses a dynamic risk prediction system, which generates space-time alignment input through multi-source data fusion, adopts tensor field modeling to embed contract constraint to construct a risk dynamic model, and solves and outputs a continuous risk field through a partial differential equation; a propagation path is analyzed in combination with asymmetric causal analysis, model parameters are adjusted in real time through a dynamic optimization algorithm, and closed-loop optimization of a risk field is achieved; and finally, through four-dimensional thermodynamic diagram interaction early warning and resource intelligent scheduling, a whole-process closed-loop system of data modeling-causal analysis-dynamic optimization-visual management and control is formed. According to the method, dynamic optimization of risk field parameters is realized through adjoint equation back propagation, and the modeling precision of a complex scene is improved; a four-dimensional space-time thermodynamic diagram rendering technology is innovated to solve the problem of fragmentation of multi-modal information expression, and risk disposal response is accelerated; key task resource supply is guaranteed by combining video memory preemption and containerization scheduling strategies, and the system stability bottleneck in a high-load scene is overcome.

The invention belongs to the technical field of wind power plant power prediction and intelligent operation and maintenance, and particularly relates to a wind power plant short-term wind power prediction method based on a space-time attention mechanism. According to the method, a dynamic graph evolved along with time is constructed through multi-source heterogeneous data fusion, and a time-varying adjacency matrix is output; obtaining space-time coupling characteristics through disturbance-sensitive space attention coding and wind speed-power coupling sensitivity driven time attention decoding; multi-scale features are output through dual-scale adaptive fusion; and finally, forming a continuously evolved prediction chain through edge end rolling learning and topological closed-loop correction by combining a loss back propagation optimization model of physical consistency constraint. According to the method, the problems of lack of physical constraints, insufficient dynamic adaptation and low efficiency of edge deployment in the prior art are solved, the prediction physical reasonability and dynamic adaptation are improved, efficient edge deployment is realized, and reliable support is provided for wind power plant scheduling.

A method, an apparatus, and a non-transitory computer-readable storage medium for video compression using a generative adversarial network (GAN) are provided. The method includes obtaining, by a generator of the GAN, a reconstructed target frame based on a reference frame and a raw target frame to be reconstructed; concatenating, by a transformer-based discriminator of the GAN, the reference frame, the raw target frame and the reconstructed target frame to obtain a paired data; determining, by the transformer-based discriminator of the GAN, whether the paired data is real or fake to guide reconstruction of the raw target frame; and determining a generator loss and a transformer-based discriminator loss, and performing gradient back propagation and updating network parameters of the GAN based on the generator loss and the transformer-based discriminator loss.

The invention provides a text-to-structured query statement method based on a large language model and reinforcement learning, and relates to the technical field of information, the method comprises the following steps: integrating a natural language problem and a database mode into a unified prompt template to obtain a candidate structured query, and generating a reference structured query by a reference model; constructing a multi-dimensional reward framework, and performing weighted aggregation on multi-dimensional rewards to obtain corresponding final reward scores; calculating a KL penalty value between the output of the strategy model and the output of the reference model based on the candidate structured query and the reference structured query to obtain a constraint of stable strategy update; updating the strategy model parameters through back propagation iteration to obtain an updated strategy model; and analyzing the prompt template by using the updated strategy model to obtain a text-to-structured query statement processing result, and completing the process from the text to the structured query statement. The problems that existing structured query generation is low in accuracy and semantic consistency is difficult to guarantee are solved.

The invention discloses a target detection method and device based on domain self-adaption, equipment and a medium. The target detection method based on domain adaptation is realized through a target detection model, the target detection model comprises a dynamic domain adaptation module, and the dynamic domain adaptation module comprises an image level domain classifier with a first dynamic confrontation gradient inversion layer and an object level domain classifier with a second dynamic confrontation gradient inversion layer. In the back propagation of the training process of the target detection model, the gradient inversion intensity of the first dynamic confrontation gradient inversion layer is dynamically adjusted according to the image level domain classification loss, and the gradient inversion intensity of the second dynamic confrontation gradient inversion layer in the back propagation is dynamically adjusted according to the object level domain classification loss; according to the method, the target detection precision under the severe weather condition can be improved, the target missing detection rate and the target false detection rate under the severe weather condition are reduced, and then the safety and the reliability of the automatic driving system are improved.

The invention belongs to the technical field of wind power plant power prediction and intelligent operation and maintenance, and particularly relates to a wind power plant short-term wind power prediction method based on a space-time attention mechanism. According to the method, a dynamic graph evolved along with time is constructed through multi-source heterogeneous data fusion, and a time-varying adjacency matrix is output; obtaining space-time coupling characteristics through disturbance-sensitive space attention coding and wind speed-power coupling sensitivity driven time attention decoding; multi-scale features are output through dual-scale adaptive fusion; and finally, forming a continuously evolved prediction chain through edge end rolling learning and topological closed-loop correction by combining a loss back propagation optimization model of physical consistency constraint. According to the method, the problems of lack of physical constraints, insufficient dynamic adaptation and low efficiency of edge deployment in the prior art are solved, the prediction physical reasonability and dynamic adaptation are improved, efficient edge deployment is realized, and reliable support is provided for wind power plant scheduling.

The invention discloses a TEE-GPU collaborative model credible training method and device based on parameter confusion, and the method comprises the steps: enabling a client to encrypt training data and a model architecture in a preprocessing stage, and uploading the encrypted training data and model architecture to a server; and the trusted execution environment of the server side decrypts the model architecture, carries out confusion processing on a linear layer and then deploys the linear layer on an external GPU (Graphics Processing Unit). In the training stage, initial forward propagation calculation of training data is firstly completed in the TEE, and then intermediate results are confused and then transmitted to the GPU so as to execute subsequent forward propagation calculation. In the back propagation process, extra confusion is applied to the gradient by the TEE, and the gradient is issued to the GPU to calculate a new gradient; and after receiving the confusion gradient, the GPU updates the parameters in a confusion form. And randomly sampling calculation data in the TEE, and carrying out integrity verification to ensure the integrity of a calculation result. Compared with an existing credible training scheme, the method has the advantages that the time overhead can be remarkably reduced while the calculation accuracy, integrity and privacy of the model are ensured.

The invention discloses a slice time sequence large model data prediction method and system based on ridge regression constraint attention, a medium and equipment, and the method comprises the steps: carrying out the normalization processing of multivariate time sequence data of each channel, segmenting the multivariate time sequence data into a plurality of overlapped slice time sequences according to a fixed length and a step length, and obtaining the slice time sequence of each channel; lLM recoding: embedding and recoding a slice time sequence into a text prototype space by using a multi-head cross attention mechanism, and splicing a text and data to obtain time sequence data with prompts and attention scores; constructing an interaction channel code to model an interdependency relationship between channels, and obtaining potential feature representation based on time sequence data with prompts; the LLM outputs projection, the potential feature representation is sent into the LLM for prediction, and a prediction output sequence is generated through linear projection; and carrying out ridge regression attention regularization, including the attention score into a loss function, and optimizing the model through back propagation.

The invention discloses a large language model control fine tuning method and system based on multi-task cooperative regulation and control, and belongs to the technical field of large language models. Generating a gating coefficient and an initial dynamic evaluation signal through the intelligent regulation and control network; a task difficulty index is obtained by combining multi-index fusion and historical moving average, and the sampling probability and the exclusive learning rate are dynamically adjusted; weighting the fusion gradient and carrying out back propagation to update parameters; and closed-loop feedback monitoring is carried out and parameters of the regulation and control network and the scheduling policy device are optimized. The system comprises a data coding module, a collaborative intelligent regulation and control module, a dynamic balance control module, a joint optimization module and a closed-loop feedback module. According to the method, gradient conflicts among tasks are relieved, the problems of convergence instability and performance imbalance are solved, the multi-task training efficiency, convergence stability and generalization ability of a large language model are improved, and the method can be widely applied to multi-class multi-task learning scenes.

The invention discloses a multi-task fixation point estimation method, which relates to the technical field of computer vision and deep learning, and comprises the following steps of: firstly, constructing a data set for providing associated data from tasks to labels and supporting multi-task cooperative training, and then constructing an MTHGaze model, through a semi-structured framework, deep learning, a geometric constraint fitting model and key tasks of cooperative processing are fused, a loss function is designed and trained, defects of each task are solved through targeted loss function design, and a scientific training process enables the model to be iteratively optimized. Finally, the targets of improving the estimation precision of the fixation point, ensuring the physical rationality and enhancing the environmental robustness are achieved; and finally, performing fixation point mapping. According to the estimation method, through data set construction, data consistency can be guaranteed, training noise can be reduced, a real reference can be provided for loss function calculation, error back propagation can be realized, and Euclidean distance errors can be minimized through L2 norms of real fixation point coordinates and prediction coordinates.

The invention provides an HRRP target micro-motion characteristic extraction method based on chaos optimization and polarization synthesis, and aims at the field of high-resolution one-dimensional range profiles. The method comprises the following steps: performing optimal polarization contrast enhancement and optimal polarization fusion based on a maximum signal-to-noise ratio criterion on HRRP multi-channel polarization data; performing period estimation on the polarization fusion data in combination with second harmonic back-off optimization and autocorrelation / amplitude difference enhancement processing; chaotic mapping and back propagation are combined to avoid the problem that traditional optimization is prone to falling into local minimum, and precession angle and size results of the space target are obtained. According to the method, the feature contrast of HRRP data is remarkably improved through polarization fusion, meanwhile, high-precision estimation of the micro-motion features of the space target is achieved under the condition of a low signal-to-noise ratio, the stability of periodic detection within limited observation time and the global search capability of parameter inversion are effectively improved, and the method is suitable for large-scale popularization and application. And the accuracy and robustness of space target micro-motion estimation in a complex environment are obviously enhanced.

The invention discloses a cross-modal anti-attack method and system based on mask weight and random cutting, and belongs to the technical field of deep learning. The method comprises the following steps: extracting a saliency mask of a source image; disturbing the source image by using the disturbance variable to generate an adversarial sample; randomly sampling K cutting frames in combination with a significance mask; respectively extracting clipping characteristics of the clipping frame on the adversarial sample and target embedding selected on the target image according to an alignment mode; calculating the cutting loss of the cutting frame according to the cutting feature and the target embedding; calculating the weighted sum of the cutting loss according to the significance mask to obtain the total loss; performing back propagation according to the total loss so as to update the disturbance variable; and re-perturbing the source image by using the perturbation variable based on the updated perturbation variable until the maximum step number or total loss convergence is achieved. The method gives consideration to attack success rate, mobility and disturbance concealment, and can adapt to quality evaluation of different large language models.

The invention belongs to the related technical field of biped robot gait planning, and discloses a method, application and equipment for generating a biped robot anthropomorphic foot end trajectory based on a full-connection network, and the method comprises the steps: (1) grouping the collected foot end trajectory data according to a starting step, an intermediate circulation step and a termination step; (2) mapping each group of foot end track data into foot end track data which can be used by a robot based on a kinematics mapping relation between a human body and the robot, and carrying out conformal interpolation; (3) training a full-connection neural network model by adopting the interpolated foot end trajectory data, further generating a foot end trajectory of each intermediate step by adopting the full-connection neural network model, and mapping the directly generated foot end trajectories of the starting step and the ending step to jointly form an anthropomorphic foot end trajectory; and the main loss of back propagation of the full-connection neural network model is punished by introducing the trajectory smoothness and the prediction precision of the start end and the tail end. According to the method, the anthropomorphism of the foot end track is improved.

The invention provides a fine tuning training method and device of MoE, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, in particular to the technical fields of distributed training and optimization of a large-scale machine learning model, compression and acceleration of a deep learning model and the like. According to the implementation scheme, model structure analysis is conducted on a MoE model to be finely adjusted, and an expert module and a non-MoE module in the MoE model are recognized; quantizing the weight parameter of the expert module by adopting first low-bit precision, and quantizing the weight parameter of the non-MoE module by adopting second low-bit precision; performing model forward calculation and back propagation calculation based on the quantized weight by using a mixed precision calculation operator; and based on an LoRA fine tuning technology, performing fine tuning training on the quantized MoE model by using a distributed training framework to obtain a target MoE model. The scheme can significantly reduce video memory occupation, effectively improve training efficiency, break through model scale limitation, and maintain excellent model performance.

The invention discloses a food plastic package seal online detection method and system, and relates to the technical field of food package quality nondestructive detection.The method comprises the steps that based on a multi-physical-quantity data set, amplitude attenuation parameters in acoustic data are extracted, and food plastic package thickness distribution is scanned; dynamically correcting the amplitude attenuation parameter through a sound wave energy absorption model to obtain calibrated acoustic data; inputting the physical quantity matrix into a pre-trained convolutional neural network model, outputting a multi-scale feature map, analyzing multi-modal feature association of the physical quantity matrix through an attention fusion mechanism, and outputting a defect probability score; and according to the defect probability score, detecting the sealing state of the seal, removing the invalid food plastic package, integrating multi-source data to generate a structured detection log, and optimizing a convolutional neural network model through back propagation to generate a detection report. According to the invention, the recognition capability and classification precision of complex defects are enhanced, and precise judgment and efficient sorting of the sealing state of the seal are realized.

The invention relates to the technical field of artificial intelligence, in particular to a large-model multi-chain reasoning optimization method and system based on truncation-overwriting, and provides a multi-chain fusion reasoning method combining thinking chain score feedback, a truncation-overwriting module and an MCTS search strategy for solving the problem that a current thinking chain lacks effective monitoring and intervention. The method comprises the steps of obtaining an initial thinking chain of a text input by a user; taking the initial thinking chain as a root node, calling a thinking chain evaluator to obtain a score of the root node and positioning weak steps; in each round of search, utilizing a truncation-overwriting module to reconstruct a thinking chain according to the weak step of the selected node with the maximum confidence upper bound, and taking the thinking chain as an expansion node; obtaining scores of the expansion nodes and positioning weak steps; modifying the scores of all ancestor nodes of the expansion node through the score of the back propagation expansion node; the thinking chain of the node with the highest score is selected as an answer to be output; the invention further provides a two-section type basic model fine tuning method.

The invention discloses a power transmission line defect detection method and device and a medium, and the method comprises the steps: carrying out the recognition and positioning of a power transmission line defect in image data according to a multi-mode large model, and obtaining a defect detection result; wherein the multi-modal large model is obtained by freezing a target trunk parameter in an initial multi-modal model, training a newly added low-rank parameter in the initial multi-modal model according to preset data, and performing back propagation optimization in combination with joint loss. According to the power transmission line defect detection method and device and the medium, model stability can be kept by freezing trunk parameters, precision can be improved and overfitting can be reduced through low-rank parameter training, multi-task performance is balanced through joint loss optimization, model robustness can be enhanced, and the detection accuracy is improved. Therefore, the defects of the power transmission line can be accurately identified by using the multi-modal large model, false detection and missing detection are reduced, and the problem that the precision and robustness of defect detection of the power transmission line are difficult to improve can be solved.

The invention discloses a distributed AI training-oriented RDMA transmission mode intelligent selection method, and the method comprises the steps: obtaining multi-dimensional data features in a deep learning training process, the multi-dimensional data features comprising a tensor structure feature, a data access mode feature, a data timeliness feature, a data dependency feature and a data change rate feature; identifying a current training stage, wherein the training stage comprises a forward propagation stage, a back propagation stage and a parameter updating stage; based on the multi-dimensional data features and the training stage, an optimal RDMA transmission mode matched with the current training scene is selected through an adaptive decision matrix, and the optimal RDMA transmission mode comprises one or more combinations of RDMA Write operation, RDMA Read operation, Send / Receive operation and Atomic operation; and switching the RDMA transmission mode in real time according to the dynamic change of a network environment, ensuring that a training data flow is not interrupted in a switching process by adopting a progressive flow migration strategy, and optimizing a subsequent mode selection decision through a historical performance learning mechanism.

The invention provides a model training method, system and device and a storage medium, and belongs to the technical field of artificial intelligence. According to the method, the back propagation process of a first model layer group of a model is executed in a mode of accessing a video memory of a GPU, the back propagation process of a second model layer group is executed in a mode of direct memory access, and due to the fact that the direct memory access mode does not need weight loading, weight loading is not needed, and therefore weight loading is not needed. The time for waiting for weight loading in the model iteration process can be shortened, in addition, the process of generating the gradient of the second model layer group and the process of writing the generated gradient of the first model layer group into the memory of the CPU are executed in parallel, and the processing efficiency is improved. The process of writing the gradient of the generated second model layer group into the memory of the CPU and the process of updating the weight of the first model layer group are executed in parallel, so that the time required for writing the generated gradient into the memory of the CPU can be masked, the time of each iteration is shortened, and the efficiency of model training is improved.

Method comprising: monitoring whether a MTLF receives a first state of an environment on which a RL training is to be performed; performing a ML model forward propagation on a first model of the environment having the first state for each of plural actions to obtain a respective expected reward for each of the plural actions: informing a service consumer on the plural actions and their respective expected reward; supervising whether the MTLF receives a RL training result information after the informing the service consumer on the plural actions, wherein the RL training result information comprises an indication of one of the plural actions, a second state of the environment, and a reward feedback; conducting a ML model backward propagation on the first model of the environment having the second state for the one of the plural actions using the reward feedback to obtain a second model of the environment.

The invention relates to a neural network weight sparse training method and device, a medium and a program product. The invention discloses a neural network weight sparse training method, and the method comprises the steps: configuring a weight sparse training parameter which comprises sparse granularity with a sparse group as a basic unit; in the forward propagation process, generating a sparse mask according to the absolute value sequence of the original weight of the neural network in each sparse group, and applying the sparse mask to obtain a sparse weight; in the back propagation process, the gradient used for updating the original weight is calculated, and the calculated gradient comprises the gradient of a loss function on the sparse weight and a regularization penalty term applied to the pruned original weight determined according to the sparse mask; and updating the original weight according to the calculated gradient.

The invention discloses a differential privacy protection method and system based on dynamic gradient, and the method comprises the steps: 1, monitoring the gradient features of back propagation in real time in the training process of a target neural network model, and calculating a dynamic sensitivity zoom factor based on the gradient features; 2, identifying the type of the network hierarchy in the target model based on a three-level progressive judgment method; and step 3, setting a noise scaling factor according to the network hierarchy, selecting a noise type according to the intensity requirement of the network hierarchy for privacy protection, obtaining a noise scale corresponding to the noise type by using the dynamic sensitivity factor and the noise scaling factor, constructing noise according to the noise scale, and injecting the constructed noise into the corresponding network hierarchy to obtain the privacy protection. Obtaining a noise-added target model; and step 4, predicting input target data based on the neural network model after noise addition. According to the method, the protection effect of the model is improved, and stealing attacks aiming at genetic privacy and high-risk group identities can be effectively defended.

The application discloses a transformer substation intelligent operation and maintenance method based on an adaptive path aggregation network, relates to the technical field of target detection, and is used for realizing transformer substation operation and maintenance multi-target detection. The method comprises the following steps: a backbone network is used for extracting feature maps of different scales from a transformer substation operation and maintenance target detection image; in a neck network, an adaptive path aggregation pyramid network fuses multi-scale features extracted by the backbone network, a channel attention fusion module for relieving semantic information conflicts caused by directly fusing different scale features, and a cross-level path aggregation module for distributing additional feature levels to obtain an optimal instance gradient back propagation path; and a head network adopts a decoupling head to predict the multi-scale features fused by the adaptive path aggregation pyramid network. The method can improve the robustness of the network to scales and plays a key role in coping with large-scale scale changes.