A Stochastic Quasi-Newton Method for Non-convex Optimization with Non-uniform Smoothness

• URL: http://arxiv.org/abs/2403.15244v1
• Date: Fri, 22 Mar 2024 14:40:29 GMT
• Title: A Stochastic Quasi-Newton Method for Non-convex Optimization with Non-uniform Smoothness
• Authors: Zhenyu Sun, Ermin Wei,
• Abstract summary: We propose a fast quasi-Newton method when there exists non-uniformity in smoothness. Our algorithm can achieve the best-known $mathcalO(epsilon-3)$ sample complexity and enjoys convergence speedup. Our numerical experiments show that our proposed algorithm outperforms the state-of-the-art approaches.
• Score: 4.097563258332958
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Classical convergence analyses for optimization algorithms rely on the widely-adopted uniform smoothness assumption. However, recent experimental studies have demonstrated that many machine learning problems exhibit non-uniform smoothness, meaning the smoothness factor is a function of the model parameter instead of a universal constant. In particular, it has been observed that the smoothness grows with respect to the gradient norm along the training trajectory. Motivated by this phenomenon, the recently introduced $(L_0, L_1)$-smoothness is a more general notion, compared to traditional $L$-smoothness, that captures such positive relationship between smoothness and gradient norm. Under this type of non-uniform smoothness, existing literature has designed stochastic first-order algorithms by utilizing gradient clipping techniques to obtain the optimal $\mathcal{O}(\epsilon^{-3})$ sample complexity for finding an $\epsilon$-approximate first-order stationary solution. Nevertheless, the studies of quasi-Newton methods are still lacking. Considering higher accuracy and more robustness for quasi-Newton methods, in this paper we propose a fast stochastic quasi-Newton method when there exists non-uniformity in smoothness. Leveraging gradient clipping and variance reduction, our algorithm can achieve the best-known $\mathcal{O}(\epsilon^{-3})$ sample complexity and enjoys convergence speedup with simple hyperparameter tuning. Our numerical experiments show that our proposed algorithm outperforms the state-of-the-art approaches.

Related papers

• On the Stochastic (Variance-Reduced) Proximal Gradient Method for Regularized Expected Reward Optimization [12.244251361123396]
  We consider a regularized expected reward optimization problem in the non-oblivious setting that covers many existing problems in reinforcement learning (RL) In particular, the method has shown to admit an $O(epsilon-4)$ sample to an $epsilon-stationary point, under standard conditions. We show that the sample complexity can be improved from $epsilon-4)$ to $O(epsilon-3)$ under additional conditions.
  arXiv  Detail & Related papers  (2024-01-23T06:01:29Z)
• Strictly Low Rank Constraint Optimization -- An Asymptotically $\mathcal{O}(\frac{1}{t^2})$ Method [5.770309971945476]
  We propose a class of non-text and non-smooth problems with textitrank regularization to promote sparsity in optimal solution. We show that our algorithms are able to achieve a singular convergence of $Ofrac(t2)$, which is exactly same as Nesterov's optimal convergence for first-order methods on smooth convex problems.
  arXiv  Detail & Related papers  (2023-07-04T16:55:41Z)
• Stochastic Inexact Augmented Lagrangian Method for Nonconvex Expectation Constrained Optimization [88.0031283949404]
  Many real-world problems have complicated non functional constraints and use a large number of data points. Our proposed method outperforms an existing method with the previously best-known result.
  arXiv  Detail & Related papers  (2022-12-19T14:48:54Z)
• Convergence of First-Order Methods for Constrained Nonconvex Optimization with Dependent Data [7.513100214864646]
  We show the worst-case complexity of convergence $tildeO(t-1/4)$ and MoreautildeO(vareps-4)$ for smooth non- optimization. We obtain first online nonnegative matrix factorization algorithms for dependent data based on projected gradient methods with adaptive step sizes and optimal convergence.
  arXiv  Detail & Related papers  (2022-03-29T17:59:10Z)
• Mirror Descent Strikes Again: Optimal Stochastic Convex Optimization under Infinite Noise Variance [17.199063087458907]
  We quantify the convergence rate of the Mirror Descent algorithm with a class of uniformly convex mirror maps. This algorithm does not require any explicit gradient clipping or normalization. We complement our results with information-theoretic lower bounds showing that no other algorithm using only first-order oracles can achieve improved rates.
  arXiv  Detail & Related papers  (2022-02-23T17:08:40Z)
• Near-Optimal High Probability Complexity Bounds for Non-Smooth Stochastic Optimization with Heavy-Tailed Noise [63.304196997102494]
  It is essential to theoretically guarantee that algorithms provide small objective residual with high probability. Existing methods for non-smooth convex optimization have complexity with bounds dependence on the confidence level that is either negative-power or logarithmic. We propose novel stepsize rules for two gradient methods with clipping.
  arXiv  Detail & Related papers  (2021-06-10T17:54:21Z)
• Leveraging Non-uniformity in First-order Non-convex Optimization [93.6817946818977]
  Non-uniform refinement of objective functions leads to emphNon-uniform Smoothness (NS) and emphNon-uniform Lojasiewicz inequality (NL) New definitions inspire new geometry-aware first-order methods that converge to global optimality faster than the classical $Omega (1/t2)$ lower bounds.
  arXiv  Detail & Related papers  (2021-05-13T04:23:07Z)
• Single-Timescale Stochastic Nonconvex-Concave Optimization for Smooth Nonlinear TD Learning [145.54544979467872]
  We propose two single-timescale single-loop algorithms that require only one data point each step. Our results are expressed in a form of simultaneous primal and dual side convergence.
  arXiv  Detail & Related papers  (2020-08-23T20:36:49Z)
• Stochastic Optimization with Heavy-Tailed Noise via Accelerated Gradient Clipping [69.9674326582747]
  We propose a new accelerated first-order method called clipped-SSTM for smooth convex optimization with heavy-tailed distributed noise in gradients. We prove new complexity that outperform state-of-the-art results in this case. We derive the first non-trivial high-probability complexity bounds for SGD with clipping without light-tails assumption on the noise.
  arXiv  Detail & Related papers  (2020-05-21T17:05:27Z)
• Towards Better Understanding of Adaptive Gradient Algorithms in Generative Adversarial Nets [71.05306664267832]
  Adaptive algorithms perform gradient updates using the history of gradients and are ubiquitous in training deep neural networks. In this paper we analyze a variant of OptimisticOA algorithm for nonconcave minmax problems. Our experiments show that adaptive GAN non-adaptive gradient algorithms can be observed empirically.
  arXiv  Detail & Related papers  (2019-12-26T22:10:10Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.