In this paper, we present a rigorous analysis of root-exponential convergence of Hermite approximations, including projection and interpolation methods, for functions that are analytic in an infinite strip containing ...
详细信息
Code vulnerability detection is a critical part of ensuring software security, and recent deep learning-based code vulnerability detection methods have proven their effectiveness. However, the currently available meth...
详细信息
In this paper, we investigate dispersive estimates for the time evolution of Hamiltonians (Equation presented) where each j satisfies certain smoothness and decay conditions. We show that, under a spectral assumption,...
In order to resolve the problem of unstable control of force in human–computer interaction based on surface EMG signals, the adaptive neural fuzzy inference system is designed to achieve the grip strength assessment....
详细信息
In the image restoration task, how to make full use of spatial and channel feature information to improve the reconstruction quality of the model without significantly increasing the computational complexity is an imp...
详细信息
ISBN:
(数字)9798331507992
ISBN:
(纸本)9798331508005
In the image restoration task, how to make full use of spatial and channel feature information to improve the reconstruction quality of the model without significantly increasing the computational complexity is an important research issue. To address this challenge, this paper proposes a CS module (Channel Compression and Split-Attention) specifically for optimizing image denoising tasks. The module first compresses the features through a preliminary convolution operation, then introduces the split attention convolution mechanism to split the features by channel and dynamically weight them, thereby enhancing the model's adaptive attention to different feature channels and improving feature expression capabilities. Through experiments on multiple image denoising datasets, the results show that the CS module significantly improves the image reconstruction quality, especially in terms of peak signal-to-noise ratio (PSNR). At the same time, the introduction of the CS module does not lead to a significant increase in computational complexity, maintaining a high computational efficiency, which verifies the practicality and superiority of the module in image denoising tasks.
This paper deals with numerical solutions of nonlinear stiff stochastic differential equations with jump-diffusion and piecewise continuous *** combining compensated split-step methods and balanced methods,a class of ...
详细信息
This paper deals with numerical solutions of nonlinear stiff stochastic differential equations with jump-diffusion and piecewise continuous *** combining compensated split-step methods and balanced methods,a class of compensated split-step balanced(CSSB)methods are suggested for solving the *** on the one-sided Lipschitz condition and local Lipschitz condition,a strong convergence criterion of CSSB methods is *** is proved under some suitable conditions that the numerical solutions produced by CSSB methods can preserve the mean-square exponential stability of the corresponding analytical *** numerical examples are presented to illustrate the obtained theoretical results and the effectiveness of CSSB ***,in order to show the computational advantage of CSSB methods,we also give a numerical comparison with the adapted split-step backward Euler methods with or without compensation and tamed explicit methods.
In recent years,there are numerous studies on chaotic systems with special equilibrium curves having various shapes such as circle,butterfly,heart and *** paper describes a new 3-D chaotic dynamical system with a caps...
详细信息
In recent years,there are numerous studies on chaotic systems with special equilibrium curves having various shapes such as circle,butterfly,heart and *** paper describes a new 3-D chaotic dynamical system with a capsule-shaped equilibrium *** proposed chaotic system has two quadratic,two cubic and two quartic nonlinear *** is noted that the proposed chaotic system has a hidden attractor since it has an infinite number of equilibrium *** is also established that the proposed chaotic system exhibits multi-stability with two coexisting chaotic attractors for the same parameter values but differential initial states.A detailed bifurcation analysis with respect to variations in the system parameters is portrayed for the new chaotic system with capsule equilibrium *** have shown MATLAB plots to illustrate the capsule equilibrium curve,phase orbits of the new chaotic system,bifurcation diagrams and *** an engineering application,we have proposed a speech cryptosystem with a numerical algorithm,which is based on our novel 3-D chaotic system with a capsule-shaped equilibrium *** proposed speech cryptosystem follows its security evolution and implementation on Field Programmable Gate Array(FPGA)*** results show that the proposed encryption system utilizes 33%of the FPGA,while the maximum clock frequency is 178.28 MHz.
作者：
Xiaofan BaiChaoxiang HeXiaojing MaBin Benjamin ZhuHai JinSchool of Cyber Science and Engineering
Huazhong University of Science and Technology National Engineering Research Center for Big Data Technology and System and Services Computing Technology and System Lab and Hubei Engineering Research Center on Big Data Security and Hubei Key Laboratory of Distributed System Security MicrosoftSchool of Computer Science and Technology
Huazhong University of Science and Technology and National Engineering Research Center for Big Data Technology and System and Services Computing Technology and System Lab and Cluster and Grid Computing Lab.
Cloud-based AI services offer numerous benefits but also introduce vulnerabilities, allowing for tampering with deployed DNN models, ranging from injecting malicious behaviors to reducing computing resources. Fingerpr...
Cloud-based AI services offer numerous benefits but also introduce vulnerabilities, allowing for tampering with deployed DNN models, ranging from injecting malicious behaviors to reducing computing resources. Fingerprint samples are generated to query models to detect such tampering. In this paper, we present Intersecting-Boundary-Sensitive Fingerprinting (IBSF), a novel method for black-box integrity verification of DNN models using only top-1 labels. Recognizing that tampering with a model alters its decision boundary, IBSF crafts fingerprint samples from normal samples by maximizing the partial Shannon entropy of a selected subset of categories to position the fingerprint samples near decision boundaries where the categories in the subset intersect. These fingerprint samples are almost indistinguishable from their source samples. We theoretically establish and confirm experimentally that these fingerprint samples' expected sensitivity to tampering increases with the cardinality of the subset. Extensive evaluation demonstrates that IBSF surpasses existing state-of-the-art fingerprinting methods, particularly with larger subset cardinality, establishing its state-of-the-art performance in black-box tampering detection using only top-1 labels. The IBSF code is available at: https://***/CGCL-codes/IBSF.
Background: In order to reduce radiation dose and improve the diagnostic accuracy of computed tomography (CT), numerous studies in recent years have explored deep learning-based image denoising methods to suppress noi...
详细信息
With the evolution of self-supervised learning, the pre-training paradigm has emerged as a predominant solution within the deep learning landscape. Model providers furnish pre-trained encoders designed to function as ...
详细信息
ISBN:
(数字)9798350331301
ISBN:
(纸本)9798350331318
With the evolution of self-supervised learning, the pre-training paradigm has emerged as a predominant solution within the deep learning landscape. Model providers furnish pre-trained encoders designed to function as versatile feature extractors, enabling downstream users to harness the benefits of expansive models with minimal effort through fine-tuning. Nevertheless, recent works have exposed a vulnerability in pre-trained encoders, highlighting their susceptibility to downstream-agnostic adversarial examples (DAEs) meticulously crafted by attackers. The lingering question pertains to the feasibility of fortifying the robustness of downstream models against DAEs, particularly in scenarios where the pre-trained encoders are publicly accessible to the *** this paper, we initially delve into existing defensive mechanisms against adversarial examples within the pre-training paradigm. Our findings reveal that the failure of current defenses stems from the domain shift between pre-training data and downstream tasks, as well as the sensitivity of encoder parameters. In response to these challenges, we propose Genetic Evolution-Nurtured Adversarial Fine-tuning (Gen-AF), a two-stage adversarial fine-tuning approach aimed at enhancing the robustness of downstream models. Gen-AF employs a genetic-directed dual-track adversarial fine-tuning strategy in its first stage to effectively inherit the pre-trained encoder. This involves optimizing the pre-trained encoder and classifier separately while incorporating genetic regularization to preserve the model’s topology. In the second stage, Gen-AF assesses the robust sensitivity of each layer and creates a dictionary, based on which the top-k robust redundant layers are selected with the remaining layers held fixed. Upon this foundation, we conduct evolutionary adaptability fine-tuning to further enhance the model’s generalizability. Our extensive experiments, conducted across ten self-supervised training methods and six d
暂无评论