Article
Computer Science, Information Systems
Xie Xie, Chen He, Cunyan Ma, Feifei Gao, Z. Jane Wang
Summary: This letter investigates a relay-aided double-hop reconfigurable intelligent surfaces (RDH-RISs) empowered outdoor-to-indoor communication system, and explores system rates and performance optimization under different settings.
IEEE WIRELESS COMMUNICATIONS LETTERS
(2022)
Article
Operations Research & Management Science
E. G. Birgin, J. M. Martinez
Summary: In this paper, we address the problem of non-simple block constraints in block coordinate descent methods and propose a general algorithm to solve it. The convergence and complexity of the algorithm are proved, and numerical experiments are conducted to verify its effectiveness.
COMPUTATIONAL OPTIMIZATION AND APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Junlong Zhu, Xin Wang, Mingchuan Zhang, Muhua Liu, Qingtao Wu
Summary: This article proposes a distributed randomized block-coordinate projection-free algorithm for solving optimization problems on large datasets, and analyzes its convergence performance. Experimental results validate that the algorithm has lower computational complexity and faster convergence rate, consistent with the theoretical analysis.
COMPLEX & INTELLIGENT SYSTEMS
(2023)
Article
Operations Research & Management Science
Filip Hanzely, Peter Richtarik
Summary: The study draws on ideas from the well-studied field of stochastic convex optimization and proposes two new algorithms for minimizing relatively smooth functions. These algorithms, Relative Randomized Coordinate Descent (relRCD) and Relative Stochastic Gradient Descent (relSGD), generalize famous algorithms in the standard smooth setting. In particular, relRCD corresponds to the first stochastic variant of mirror descent algorithm with linear convergence rate.
COMPUTATIONAL OPTIMIZATION AND APPLICATIONS
(2021)
Article
Automation & Control Systems
Julie Nutini, Issam Laradji, Mark Schmidt
Summary: In this paper, the authors explore the block partitioning strategy, the block selection rule, and the block update rule in the Block Coordinate Descent (BCD) methods. They propose variations for each of these building blocks that can significantly improve the progress made by each BCD iteration. Numerical results are provided for various machine learning problems to support their findings.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Mathematics, Applied
Wei Wei, Tao Shi, Song Nie, Xiaoping Chen
Summary: Two adaptive block coordinate descent methods are proposed for solving large-scale ridge regression problems, outperforming state-of-the-art methods in terms of computing times without the need for predetermining block pavings or solving any subsystems. Convergence theories are established and numerical experiments are conducted to illustrate the significant improvement in performance.
COMPUTATIONAL & APPLIED MATHEMATICS
(2023)
Article
Engineering, Mechanical
Dongyuan Shi, Bhan Lam, Woon-Seng Gan, Shulin Wen
Summary: Multichannel active noise control (MCANC) is an effective solution for achieving large noise-cancellation areas in complicated acoustic fields, but the computational complexity of MCANC algorithms has been a challenge. The proposed BCD McFxLMS algorithm reduces computation costs while maintaining noise reduction performance, and integrates a momentum mechanism to improve convergence speed. Simulation and experimental results demonstrate the effectiveness of the proposed algorithms in realistic environments.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Computer Science, Artificial Intelligence
Lei Yin, Chong Yu, Yuyi Wang, Bin Zou, Yuan Yan Tang
Summary: This paper proposes a novel transformation model that can learn from point correspondences, built using weighted support vector regression and manifold regularization. A probabilistic model to assess correspondences confidence is introduced, guiding model training. Experimental results show that the method is efficient and outperforms other state-of-the-art methods.
APPLIED INTELLIGENCE
(2021)
Article
Computer Science, Information Systems
Shazia Riaz, Saqib Ali, Guojun Wang, Asad Anees
Summary: Deep learning models have revolutionized AI tasks by producing accurate predictions. However, these models rely on precise training with large-scale datasets, which may contain sensitive personal information and pose a risk to privacy. Existing privacy-preserving deep learning models have unsatisfactory performance, leading to limited adoption in the industry. To address this, the authors developed a differentially private version of the block coordinate descent algorithm, which significantly reduces privacy cost while achieving high accuracy. Empirical evaluation shows competitive performance compared to state-of-the-art privacy mechanisms.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2023)
Article
Computer Science, Artificial Intelligence
Tao Sun, Linbo Qiao, Dongsheng Li
Summary: The article investigates the computational complexity and convergence rates of the proximal inertial gradient descent (PIGD) algorithm, demonstrating a nonergodic O(1/k) rate for coercive objective functions and a sublinear rate for non-coercive functions. Linear convergence is proven for objective functions satisfying the Polyak- Lojasiewicz (PL) property. The results are extended to the multiblock version and consider both cyclic and stochastic index selection strategies.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Lin Wang, Mingchuan Zhang, Junlong Zhu, Ling Xing, Qingtao Wu
Summary: This study proposes a privacy-preserving decentralized randomized block-coordinate subgradient projection algorithm, which can reduce the computational burden of each agent while protecting data privacy, and solve a constrained huge-scale optimization problem over time-varying networks. The computational benefit of the algorithm is demonstrated by experimental results, and the theoretical results are verified.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Hao Chen, Yu Ye, Ming Xiao, Mikael Skoglund
Summary: This paper investigates the application of distributed learning in machine learning to alleviate the burden on central servers. Incremental block-coordinate descent (I-BCD) and asynchronous parallel incremental BCD (API-BCD) methods are proposed to reduce communication costs and accelerate convergence speed.
IEEE TRANSACTIONS ON BIG DATA
(2023)
Article
Computer Science, Artificial Intelligence
Chunming Zhang, Lixing Zhu, Yanbo Shen
Summary: In this paper, a new family of robust Bregman divergences called robust-BD is proposed, which is less sensitive to data outliers. The performance of the proposed penalized robust-BD estimate is evaluated through extensive numerical experiments and compared with classical approaches, showing improvements on existing methods.
Article
Mathematics, Applied
Jianheng Chen, Wen Huang
Summary: This paper proposes an optimization model and a block coordinate descent algorithm for low-rank tensor completion problems. Numerical experiments show that the proposed model with the algorithm is feasible and effective.
NUMERICAL LINEAR ALGEBRA WITH APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Jinyao Ma, Haibin Zhang, Shanshan Yang, Jiaojiao Jiang, Gaidi Li
Summary: Convex clustering, a method to convert clustering problem into convex optimization problem, has been widely studied to overcome the limitations of traditional clustering algorithms. However, it is susceptible to outlier features. To address this issue, this paper proposes a model that decomposes data into clustering structure component and normalized component to accurately identify and remove outlier features. The proposed method outperforms existing approaches in convex clustering according to experiments on synthetic and UCI datasets.
TSINGHUA SCIENCE AND TECHNOLOGY
(2023)
Article
Operations Research & Management Science
Andrea Cristofari, Marianna De Santis, Stefano Lucidi, Francesco Rinaldi
JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS
(2017)
Article
Environmental Sciences
Armando Pelliccioni, Andrea Cristofari, Mafalda Lamberti, Claudio Gariazzo
INTERNATIONAL JOURNAL OF ENVIRONMENT AND POLLUTION
(2017)
Article
Operations Research & Management Science
Andrea Cristofari
OPTIMIZATION LETTERS
(2017)
Article
Operations Research & Management Science
Andrea Cristofari, Tayebeh Dehghan Niri, Stefano Lucidi
OPTIMIZATION LETTERS
(2019)
Article
Operations Research & Management Science
Andrea Cristofari
COMPUTATIONAL OPTIMIZATION AND APPLICATIONS
(2019)
Article
Operations Research & Management Science
Andrea Cristofari, Marianna De Santis, Stefano Lucidi, Francesco Rinaldi
COMPUTATIONAL OPTIMIZATION AND APPLICATIONS
(2020)
Article
Operations Research & Management Science
Andrea Cristofari, Gianni Di Pillo, Giampaolo Liuzzi, Stefano Lucidi
Summary: This paper considers nonlinear optimization problems with nonlinear equality constraints and bound constraints on the variables. It proposes a modified algorithm to incorporate second-order information into the augmented Lagrangian framework, using an active-set strategy. Numerical results confirm that the proposed algorithm is a viable alternative to ALGENCAN with greater robustness.
JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS
(2022)
Article
Management
Andrea Cristofari
Summary: This paper proposes a novel algorithm to solve LASSO problems with zero-sum constraint in high-dimensional spaces. The algorithm combines a tailored active-set technique with a 2-coordinate descent scheme and uses two different strategies at each iteration. Global convergence to optimal solutions is proved and numerical results demonstrate the effectiveness of the proposed method on synthetic and real datasets. The software is publicly available.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2023)
Article
Neurosciences
Andrea Cristofari, Marianna De Santis, Stefano Lucidi, John Rothwell, Elias P. Casula, Lorenzo Rocchi
Summary: The combination of TMS and EEG allows for the study of cortical physiology by examining brain electrical responses, known as TEPs. However, these TEPs can be contaminated by AEPs due to the TMS click, making it difficult to differentiate between the two using common statistical methods. In this study, machine learning algorithms were used to successfully distinguish TEPs with masked TMS clicks, AEPs, and non-masked TEPs. The classification accuracy was higher when using averaged TEPs and lower at the group level when comparing different stimulation conditions and subjects.
Article
Mathematics, Applied
Andrea Cristofari, Francesco Rinaldi
Summary: This paper presents an optimization method for structured optimization problems with sparse solutions, demonstrating good scalability in handling dimensionality and number of atoms. Global convergence to stationary points is analyzed, and it is shown that the algorithm can identify a specific subset of atoms with zero weight in the final solution under suitable assumptions.
SIAM JOURNAL ON OPTIMIZATION
(2021)
Article
Mathematics, Applied
Andrea Cristofari, Francesco Rinaldi, Francesco Tudisco
SIAM JOURNAL ON APPLIED MATHEMATICS
(2020)
Article
Multidisciplinary Sciences
A. Cristofari, M. De Santis, S. Lucidi, F. Rinaldi
