Matrix minimization techniques that employ the nuclear norm have gained recognition for their applicability in tasks like image inpainting, clustering, classification, and reconstruction. However, they come with inher...
详细信息
Matrix minimization techniques that employ the nuclear norm have gained recognition for their applicability in tasks like image inpainting, clustering, classification, and reconstruction. However, they come with inherent biases and computational burdens, especially when used to relax the rank function, making them less effective and efficient in real-world scenarios. To address these challenges, our research focuses on generalized nonconvex rank regularization problems in robust matrix completion, low-rank representation, and robust matrix regression. We introduce innovative approaches for effective and efficient low-rank matrix learning, grounded in generalized nonconvex rank relaxations inspired by various substitutes for the ?0-norm relaxed functions. These relaxations allow us to more accurately capture low-rank structures. Our optimization strategy employs a nonconvex and multi-variable alternating direction method of multipliers, backed by rigorous theoretical analysis for complexity and *** algorithm iteratively updates blocks of variables, ensuring efficient convergence. Additionally, we incorporate the randomized singular value decomposition technique and/or other acceleration strategies to enhance the computational efficiency of our approach, particularly for large-scale constrained minimization problems. In conclusion, our experimental results across a variety of image vision-related application tasks unequivocally demonstrate the superiority of our proposed methodologies in terms of both efficacy and efficiency when compared to most other related learning methods.
The maximum principle has bridged mathematical optimization to optimal control,ushering in significant developments and refinements in optimal control theory,notably during the 1960s with the advent of linear quadrati...
The maximum principle has bridged mathematical optimization to optimal control,ushering in significant developments and refinements in optimal control theory,notably during the 1960s with the advent of linear quadratic (LQ)control and linear quadratic estimation (LQE).This progression propelled optimal control theory into further advancements,encompassing stochastic control,robust/H-infinity control,model predictive control (MPC),networked control,and reinforcement learning *** control,established upon a rigorous mathematical foundation,extends static optimization theory to dynamic systems,exhibiting scientific essence,unity,and ***,since its inception,optimal control theory has served as an indispensable core role across all control-related domains,including communication-constrained control in networked systems,consensus control,cooperative control,and reinforcement learning control.
In the paper, we investigate the optimization problem(OP) by applying the optimal control method. The optimization problem is reformulated as an optimal control problem(OCP) where the controller(iteration updating) is...
详细信息
In the paper, we investigate the optimization problem(OP) by applying the optimal control method. The optimization problem is reformulated as an optimal control problem(OCP) where the controller(iteration updating) is designed to minimize the sum of costs in the future time instant, which thus theoretically generates the “optimal algorithm”(fastest and most stable). By adopting the maximum principle and linearization with Taylor expansion, new algorithms are proposed. It is shown that the proposed algorithms have a superlinear convergence rate and thus converge more rapidly than the gradient descent;meanwhile, they are superior to Newton's method because they are not divergent in general and can be applied in the case of a singular or indefinite Hessian matrix. More importantly, the OCP method contains the gradient descent and the Newton's method as special cases, which discovers the theoretical basis of gradient descent and Newton's method and reveals how far these algorithms are from the optimal algorithm. The merits of the proposed optimization algorithm are illustrated by numerical experiments.
With the rise of artificial intelligence and cloud computing, machine-learning-as-a-service platforms,such as Google, Amazon, and IBM, have emerged to provide sophisticated tasks for cloud applications. These propriet...
详细信息
With the rise of artificial intelligence and cloud computing, machine-learning-as-a-service platforms,such as Google, Amazon, and IBM, have emerged to provide sophisticated tasks for cloud applications. These proprietary models are vulnerable to model extraction attacks due to their commercial value. In this paper, we propose a time-efficient model extraction attack framework called Swift Theft that aims to steal the functionality of cloud-based deep neural network models. We distinguish Swift Theft from the existing works with a novel distribution estimation algorithm and reference model settings, finding the most informative query samples without querying the victim model. The selected query samples can be applied to various cloud models with a one-time selection. We evaluate our proposed method through extensive experiments on three victim models and six datasets, with up to 16 models for each dataset. Compared to the existing attacks, Swift Theft increases agreement(i.e., similarity) by 8% while consuming 98% less selecting time.
Driven by practical applications, the achievement of distributed observers for nonlinear systems has emerged as a crucial advancement in recent years. However, existing theoretical advancements face certain limitation...
详细信息
Driven by practical applications, the achievement of distributed observers for nonlinear systems has emerged as a crucial advancement in recent years. However, existing theoretical advancements face certain limitations: They either fail to address more complex nonlinear phenomena, rely on hard-to-verify assumptions, or encounter difficulties in solving system ***, this paper aims to address these challenges by investigating distributed observers for nonlinear systems through the full-measured canonical form(FMCF), which is inspired by full-measured system(FMS) theory. To begin with, this study addresses the fact that the FMCF can only be obtained through the observable canonical form(OCF) in existing FMS *** paper demonstrates that a class of nonlinear systems can directly obtain FMCF through state space equations, independent of OCF. Also, a general method for solving FMCF in such systems is provided. Furthermore, based on the FMCF, A distributed observer is developed for nonlinear systems under two scenarios: Lipschitz conditions and open-loop bounded *** paper establishes their asymptotic omniscience and demonstrates that the designed distributed observer in this study has fewer design parameters and is more convenient to construct than existing approaches. Finally, the effectiveness of the proposed methods is validated through simulation results on Van der Pol oscillators and microgrid systems.
Dear Editor,The letter deals with the distributed state and fault estimation of the whole physical layer for cyber-physical systems(CPSs) when the cyber layer suffers from DoS attacks. With the advancement of embedded...
详细信息
Dear Editor,The letter deals with the distributed state and fault estimation of the whole physical layer for cyber-physical systems(CPSs) when the cyber layer suffers from DoS attacks. With the advancement of embedded computing, communication and related hardware technologies, CPSs have attracted extensive attention and have been widely used in power system, traffic network, refrigeration system and other fields.
Backdoor attacks pose great threats to deep neural network models. All existing backdoor attacks are designed for unstructured data(image, voice, and text), but not structured tabular data, which has wide real-world a...
详细信息
Backdoor attacks pose great threats to deep neural network models. All existing backdoor attacks are designed for unstructured data(image, voice, and text), but not structured tabular data, which has wide real-world applications, e.g., recommendation systems, fraud detection, and click-through rate prediction. To bridge this research gap, we make the first attempt to design a backdoor attack framework, named BAD-FM, for tabular data prediction models. Unlike images or voice samples composed of homogeneous pixels or signals with continuous values, tabular data samples contain well-defined heterogeneous fields that are usually sparse and discrete. Tabular data prediction models do not solely rely on deep networks but combine shallow components(e.g., factorization machine, FM) with deep components to capture sophisticated feature interactions among fields. To tailor the backdoor attack framework to tabular data models, we carefully design field selection and trigger formation algorithms to intensify the influence of the trigger on the backdoored model. We evaluate BAD-FM with extensive experiments on four datasets, i.e.,HUAWEI, Criteo, Avazu, and KDD. The results show that BAD-FM can achieve an attack success rate as high as 100%at a poisoning ratio of 0.001%, outperforming baselines adapted from existing backdoor attacks against unstructured data models. As tabular data prediction models are widely adopted in finance and commerce, our work may raise alarms on the potential risks of these models and spur future research on defenses.
In this study, we consider a single-link flexible manipulator in the presence of an unknown Bouc-Wen type of hysteresis and intermittent actuator faults. First, an inverse hysteresis dynamics model is introduced, and ...
详细信息
In this study, we consider a single-link flexible manipulator in the presence of an unknown Bouc-Wen type of hysteresis and intermittent actuator faults. First, an inverse hysteresis dynamics model is introduced, and then the control input is divided into an expected input and an error compensator. Second,a novel adaptive neural network-based control scheme is proposed to cancel the unknown input hysteresis. Subsequently,by modifying the adaptive laws and local control laws, a fault-tolerant control strategy is applied to address uncertain intermittent actuator faults in a flexible manipulator system. Through the direct Lyapunov theory, the proposed scheme allows the state errors to asymptotically converge to a specified interval. Finally,the effectiveness of the proposed scheme is verified through numerical simulations and experiments.
This letter proposes a reliable transfer learning(RTL)method for pre-fault dynamic security assessment(DSA)in power systems to improve DSA performance in the presence of potentially related unknown *** takes individua...
详细信息
This letter proposes a reliable transfer learning(RTL)method for pre-fault dynamic security assessment(DSA)in power systems to improve DSA performance in the presence of potentially related unknown *** takes individual discrepancies into consideration and can handle unknown faults with incomplete *** experiment results demonstrate high DSA accuracy and computational efficiency of the proposed RTL *** analysis shows RTL can guarantee system performance.
Legged robots have always been a focal point of research for scholars domestically and *** to other types of robots,quadruped robots exhibit superior balance and stability,enabling them to adapt effectively to diverse...
详细信息
Legged robots have always been a focal point of research for scholars domestically and *** to other types of robots,quadruped robots exhibit superior balance and stability,enabling them to adapt effectively to diverse environments and traverse rugged *** makes them well-suited for applications such as search and rescue,exploration,and transportation,with strong environmental adaptability,high flexibility,and broad application *** paper discusses the current state of research on quadruped robots in terms of development status,gait trajectory planning methods,motion control strategies,reinforcement learning applications,and control algorithm *** highlights advancements in modeling,optimization,control,and data-driven *** study identifies the adoption of efficient gait planning algorithms,the integration of reinforcement learning-based control technologies,and data-driven methods as key directions for the development of quadruped *** aim is to provide theoretical references for researchers in the field of quadruped robotics.
暂无评论