Search

Robust face clustering enjoys a wide range of applications for gate passes, surveillance systems and security analysis in embedded sensors. Nevertheless, existing algorithms have limitations in finding accurate clusters when data contain noise (e.g., occluded face clustering and recognition). It is known that in subspace clustering, the ℓ 1 - and ℓ 2 -norm regularizers can improve subspace preservation and connectivity, respectively, and the elastic net regularizer (i.e., the mixture of the ℓ 1 - and ℓ 2 -norms) provides a balance between the two properties. However, existing deterministic methods have high per iteration computational complexities, making them inapplicable to large-scale problems. To address this issue, this paper proposes the first accelerated stochastic variance reduction gradient (RASVRG) algorithm for robust subspace clustering. We also introduce a new momentum acceleration technique for the RASVRG algorithm. As a result of the involvement of this momentum, the RASVRG algorithm achieves both the best oracle complexity and the fastest convergence rate, and it reaches higher efficiency in practice for both strongly convex and not strongly convex models. Various experimental results show that the RASVRG algorithm outperformed existing state-of-the-art methods with elastic net and ℓ 1 -norm regularizers in terms of accuracy in most cases. As demonstrated on real-world face datasets with different manually added levels of pixel corruption and occlusion situations, the RASVRG algorithm achieved much better performance in terms of accuracy and robustness. [ABSTRACT FROM AUTHOR]

This paper investigates the stochastic optimization problem focusing on developing scalable parallel algorithms for deep learning tasks. Our solution involves a reformation of the objective function for stochastic optimization in neural network models, along with a novel parallel computing strategy, coined the weighted aggregating stochastic gradient descent (WASGD). Following a theoretical analysis on the characteristics of the new objective function, WASGD introduces a decentralized weighted aggregating scheme based on the performance of local workers. Without any center variable, the new method automatically gauges the importance of local workers and accepts them by their contributions. Furthermore, we have developed an enhanced version of the method, WASGD+, by (1) implementing a designed sample order and (2) upgrading the weight evaluation function. To validate the new method, we benchmark our pipeline against several popular algorithms including the state-of-the-art deep neural network classifier training techniques (e.g., elastic averaging SGD). Comprehensive validation studies have been conducted on four classic datasets: CIFAR-100, CIFAR-10, Fashion-MNIST, and MNIST. Subsequent results have firmly validated the superiority of the WASGD scheme in accelerating the training of deep architecture. Better still, the enhanced version, WASGD+, is shown to be a significant improvement over its prototype. [ABSTRACT FROM AUTHOR]

In recent years, there is massive development of new metaheuristics as stochastic methods. Meanwhile, there is not any metaheuristics is powerful to handle all problems as stated in the no-free-lunch (NFL) theory. Based on this circumstance, this paper introduces a new swarm-based metaheuristics with the main strategy moving toward the resultant of better swarm members and avoiding the resultant of worse swarm members called group better-worse algorithm (GBWA). It consists of five searches: moving toward the best swarm member, moving toward the resultant of better swarm members, moving away from the resultant of worse swarm members, searching locally, and jumping to the opposite area. GBWA is then evaluated in three ways. The first evaluation is a comparative evaluation where GBWA is compared to five recent metaheuristics: coati optimization algorithm (COA), average and subtraction-based optimization (ASBO), clouded leopard optimization (CLO), total interaction algorithm (TIA), and osprey optimization algorithm (OOA). The second evaluation is the individual search evaluation. The third evaluation is hyperparameter test. The collection of 23 classic functions is chosen as the use case in all evaluations. The result of the first evaluation shows that GBWA is better than COA, ASBO, CLO, TIA, and OOA in 20, 21, 20, 21, and 21 functions consecutively. Meanwhile, the result of the second evaluation shows the equal contribution between the motion toward the best swarm member and the motion toward the resultant of better swarm members. [ABSTRACT FROM AUTHOR]

This paper focuses on stochastic methods for solving smooth non-convex strongly-concave min-max problems, which have received increasing attention due to their potential applications in deep learning (e.g., deep AUC maximization, distributionally robust optimization). However, most of the existing algorithms are slow in practice, and their analysis revolves around the convergence to a nearly stationary point. We consider leveraging the Polyak-Łojasiewicz (PL) condition to design faster stochastic algorithms with stronger convergence guarantee. Although PL condition has been utilized for designing many stochastic minimization algorithms, their applications for non-convex min-max optimization remain rare. In this paper, we propose and analyze a generic framework of proximal stage-based method with many well-known stochastic updates embeddable. Fast convergence is established in terms of both the primal objective gap and the duality gap. Compared with existing studies, (i) our analysis is based on a novel Lyapunov function consisting of the primal objective gap and the duality gap of a regularized function, and (ii) the results are more comprehensive with improved rates that have better dependence on the condition number under different assumptions. We also conduct deep and non-deep learning experiments to verify the effectiveness of our methods. [ABSTRACT FROM AUTHOR]

Machine learning (ML) is evolving rapidly and has made many theoretical breakthroughs while widely applied in various fields. ML allows systems the ability to access data and use it to enable computers to execute cognitive processes such as learning and improving from previous experiences and solving complicated issues. Many first-order stochastic optimization methods have been used to solve the optimization model of ML. These algorithms adopt Barzilai-Borwein (BB) step size instead of fixed or diminishing step size to improve performance. However, the BB step size format involves fractional calculation, which inevitably leads to a zero denominator, especially when the objective function is non-convex. The BB technique will be violated if the denominator is near 0 or even negative. To improve the computation of the step size, a Positive Defined Stabilized Barzilai-Borwein (PDSBB) approach is introduced in this paper. Integrating PDSBB with the stochastic variance reduced gradient (SVRG) approach, a new method SVRG-PDSBB is proposed. Numerical experiments have shown that the new algorithm has stabilized the performance of the new step size, which successfully avoiding zero denominators and effectively solving the common problems in machine learning. The convergence of SVRG-PDSBB is theoretically and numerically proven, and the effectiveness of the new algorithm is shown by comparison with other algorithms. [ABSTRACT FROM AUTHOR]

While many distributed optimization algorithms have been proposed for solving smooth or convex problems over the networks, few of them can handle non-convex and non-smooth problems. Based on a proximal primal-dual approach, this paper presents a new (stochastic) distributed algorithm with Nesterov momentum for accelerated optimization of non-convex and non-smooth problems. Theoretically, we show that the proposed algorithm can achieve an $\epsilon$ -stationary solution under a constant step size with $\mathcal {O}(1/\epsilon ^2)$ computation complexity and $\mathcal {O}(1/\epsilon)$ communication complexity when the epigraph of the non-smooth term is a polyhedral set. When compared to the existing gradient tracking based methods, the proposed algorithm has the same order of computation complexity but lower order of communication complexity. To the best of our knowledge, the presented result is the first stochastic algorithm with the $\mathcal {O}(1/\epsilon)$ communication complexity for non-convex and non-smooth problems. Numerical experiments for a distributed non-convex regression problem and a deep neural network based classification problem are presented to illustrate the effectiveness of the proposed algorithms. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel adaptive importance sampling (AIS) algorithm for probabilistic inference. The sampler learns a proposal distribution adaptation strategy by framing AIS as a reinforcement learning problem. Under this structure, the proposal distribution of the sampler is treated as an agent whose state is controlled using a parameterized policy. At each iteration of the algorithm, the agent earns a reward that is related to its contribution to the variance of the AIS estimator of the normalization constant of the target distribution. Policy gradient methods are employed to learn a locally optimal policy that maximizes the expected value of the sum of all rewards. Numerical simulations on two different examples demonstrate promising results for the future application of the proposed method to complex Bayesian models. [ABSTRACT FROM AUTHOR]

Decentralization is a promising method of scaling up parallel machine learning systems. In this paper, we provide a tight lower bound on the iteration complexity for such methods in a stochastic non-convex setting. Our lower bound reveals a theoretical gap in known convergence rates of many existing decentralized training algorithms, such as D-PSGD. We prove by construction this lower bound is tight and achievable. Motivated by our insights, we further propose DeTAG, a practical gossip-style decentralized algorithm that achieves the lower bound with only a logarithm gap. While a simple version of DeTAG with plain SGD and constant step size suffice for achieving theoretical limits, we additionally provide convergence bound for DeTAG under general non-increasing step size and momentum. Empirically, we compare DeTAG with other decentralized algorithms on multiple vision benchmarks, including CIFAR10/100 and ImageNet. We substantiate our theory and show DeTAG converges faster on unshuffled data and in sparse networks. Furthermore, we study a DeTAG variant, DeTAG*, that practically speeds up data-center-scale model training. This manuscript is the extended version for (Lu and De Sa, 2021). [ABSTRACT FROM AUTHOR]

Stochastic optimization (SO) algorithms based on the Powerball function, namely powered stochastic optimization (PoweredSO) algorithms, have been confirmed, effectively, and demonstrated great potential in the context of large-scale optimization and machine learning tasks. Nevertheless, the issue of how to determine the learning rate for PoweredSO is a challenge and still unsolved problem. In this paper, we propose a class of adaptive PoweredSO approaches that are efficient, scalable and robust. It takes advantage of the hypergradient descent (HD) technique to automatically acquire an online learning rate for PoweredSO-like methods. In the first part, we study the behavior of the canonical PoweredSO algorithm, the Powerball stochastic gradient descent (pbSGD) method, with HD. The existing PoweredSO algorithms also suffer from the high variance because they take the similar algorithmic framework to SO algorithms, arising from sampling tactics. Therefore, the second portion develops an adaptive powered variance-reduced optimization method via utilizing both variance-reduced technique and HD. Moreover, we present the convergence analysis of the proposed algorithms and explore their iteration complexity on non-convex cases. Numerical experiments are conducted on machine learning tasks, verifying the superior performance over modern SO algorithms. • We propose a type of adaptive powered stochastic optimization algorithms. • The theoretical analysis of the resulting algorithms is established. • Experiments on machine learning tasks verify the efficacy of the algorithms. [ABSTRACT FROM AUTHOR]

There has been a surge of works bridging MCMC sampling and optimization, with a specific focus on translating non-asymptotic convergence guarantees for optimization problems into the analysis of Langevin algorithms in MCMC sampling. A conspicuous distinction between the convergence analysis of Langevin sampling and that of optimization is that all known convergence rates for Langevin algorithms depend on the dimensionality of the problem, whereas the convergence rates for optimization are dimension-free for convex problems. Whether a dimension independent convergence rate can be achieved by the Langevin algorithm is thus a long-standing open problem. This paper provides an affirmative answer to this problem for the case of either Lipschitz or smooth convex functions with normal priors. By viewing Langevin algorithm as composite optimization, we develop a new analysis technique that leads to dimension independent convergence rates for such problems. [ABSTRACT FROM AUTHOR]

This paper studies some asymptotic properties of adaptive algorithms widely used in optimization and machine learning, and among them Adagrad and Rmsprop, which are involved in most of the blackbox deep learning algorithms. Our setup is the non-convex landscape optimization point of view, we consider a one time scale parametrization and the situation where these algorithms may or may not be used with mini-batches. We adopt the point of view of stochastic algorithms and establish the almost sure convergence of these methods when using a decreasing step-size towards the set of critical points of the target function. With a mild extra assumption on the noise, we also obtain the convergence towards the set of minimizers of the function. Along our study, we also obtain a "convergence rate" of the methods, in the vein of the works of Ghadimi and Lan (2013). [ABSTRACT FROM AUTHOR]

We consider stochastic optimization of a smooth non-convex loss function with a convex non-smooth regularizer. In the online setting, where a single sample of the stochastic gradient of the loss is available at every iteration, the problem can be solved using the proximal stochastic gradient descent (SGD) algorithm and its variants. However in many problems, especially those arising in communications and signal processing, information beyond the stochastic gradient may be available thanks to the structure of the loss function. By constructing a strongly convex surrogate of the loss function at every iteration, the stochastic SCA algorithms can exploit the structural properties of the loss function. In this work, we take a closer look at the stochastic SCA algorithm and develop its asynchronous variant which can be used for resource allocation in wireless networks. While the stochastic SCA algorithm is known to converge asymptotically, its iteration complexity has not been well-studied, and is the focus of the current work. The insights obtained from the non-asymptotic analysis allow us to develop a more practical asynchronous variant of the stochastic SCA algorithm which allows the use of surrogate calculated in earlier iterations. We characterize precise bound on the maximum delay the algorithm can tolerate, while still achieving the same convergence rate. We apply the algorithm to the problem of linear precoding in wireless sensor networks, where it can be implemented at low complexity but is shown to perform well in practice. [ABSTRACT FROM AUTHOR]