Efficient detection of defects in fabric plays a dominant role in automating quality control. It significantly enhances efficiency as well as accuracy of quality assurance in textile manufacturing. Skilled humans...
详细信息
The gait abnormality may be the cause of various diseases like foot drop, lower back trembling, and osteoarthritis in the human body. The causes may affect body performance. The problem may be solved if we notice it b...
详细信息
Vehicle-to-everything (V2X) communication is crucial in Intelligent Transportation Systems (ITSs) for enhancing road safety, traffic efficiency, and convenience. To address the need for trust in V2X communications, a ...
详细信息
Deep neural networks (DNNs) having multiple hidden layers are very efficient to learn large volume datasets and applied in a wide range of applications. The DNNs are trained on these datasets using learning algorithms...
详细信息
Deep neural networks (DNNs) having multiple hidden layers are very efficient to learn large volume datasets and applied in a wide range of applications. The DNNs are trained on these datasets using learning algorithms to learn the relationships among different variables. The base method that makes DNNs successful is stochastic gradient descent (SGD). The gradient reveals the way that a function’s steepest rate of alteration is occurring. No matter how the gradient behaves, the key issue with basic SGD is that all parameters must adjust in equal-sized increments. Consequently, creating adaptable step sizes for every parameter is an effective method of deep model optimization. Gradient-based adaptive techniques utilize local changes in gradients or the square roots of exponential moving averages of squared previous gradients. However, current optimizers continue to struggle with effectively utilizing optimization curved knowledge. The novel emapDiffP optimizer suggested in this study utilizes the prior two parameters to generate a non-periodic and non-negative function, and the upgrade parameter makes use of a partly adaptive value to account for learning rate adjustability. Thus, the optimization steps become smoother with a more accurate step size for the immediate past parameter, a partial adapting value, and the largest two momentum values as the denominator of parameter updating. The rigorous tests on benchmark datasets show that the presented emapDiffP performs significantly better than its counterparts. In terms of classification accuracy, the emapDiffP algorithm gives the best classification accuracy on CIFAR10, MNIST, and Mini-ImageNet datasets for all examined networks and on the CIFAR100 dataset for most of the networks examined. It offers the best classification accuracy on the ImageNet dataset with the ResNet18 model. For image classification tasks on various datasets, the suggested emapDiffP technique offers outstanding training speed. With MNIST, CIFAR1
The emergence of the Internet of Things (IoT) has enabled the proliferation of interconnected devices and sensors, generating vast amounts of often complex and unstructured data. Deep learning (DL), a subfield of mach...
详细信息
This paper proposes the Modified Light GBM to classify the Malicious Users (MUs) and legitimate Secondary Users (SUs) in the cognitive-radio network. The proposed method is to avoid the consequences of malicious users...
详细信息
Concerns related to the proliferation of automated face recognition technology with the intent of solving or preventing crimes continue to mount. The technology being implicated in wrongful arrest following 1-to-many ...
详细信息
Collaborative inference(co-inference) accelerates deep neural network inference via extracting representations at the device and making predictions at the edge server, which however might disclose the sensitive inform...
详细信息
Collaborative inference(co-inference) accelerates deep neural network inference via extracting representations at the device and making predictions at the edge server, which however might disclose the sensitive information about private attributes of users(e.g.,race). Although many privacy-preserving mechanisms on co-inference have been proposed to eliminate privacy concerns, privacy leakage of sensitive attributes might still happen during inference. In this paper, we explore privacy leakage against the privacy-preserving co-inference by decoding the uploaded representations into a vulnerable form. We propose a novel attack framework named AttrL eaks, which consists of the shadow model of feature extractor(FE), the susceptibility reconstruction decoder,and the private attribute classifier. Based on our observation that values in inner layers of FE(internal representation) are more sensitive to attack, the shadow model is proposed to simulate the FE of the victim in the blackbox scenario and generates the internal ***, the susceptibility reconstruction decoder is designed to transform the uploaded representations of the victim into the vulnerable form, which enables the malicious classifier to easily predict the private attributes. Extensive experimental results demonstrate that AttrLeaks outperforms the state of the art in terms of attack success rate.
A uniform triangular array (UTA) is proposed for physics-based 2D direction-of-arrival (DOA) estimations of unknown incoming signals. Three capacitively loaded top-hat antennas are used as array elements. Unlike conve...
详细信息
With the exponential increase in data consumption from activities like streaming, gaming, and IoT applications, existing network infrastructure is under significant strain. To address the growing demand for higher dat...
详细信息
暂无评论