English
Related papers

Related papers: Efficiently Escaping Saddle Points for Policy Opti…

200 papers

We develop a new algorithm for non-convex stochastic optimization that finds an $\epsilon$-critical point in the optimal $O(\epsilon^{-3})$ stochastic gradient and Hessian-vector product computations. Our algorithm uses Hessian-vector…

Machine Learning · Computer Science 2021-07-13 Hoang Tran , Ashok Cutkosky

Optimizing non-convex functions is of primary importance in the vast majority of machine learning algorithms. Even though many gradient descent based algorithms have been studied, successive convex approximation based algorithms have been…

Optimization and Control · Mathematics 2019-03-06 Amrit Singh Bedi , Ketan Rajawat , Vaneet Aggarwal

We provide the first study of the problem of finding differentially private (DP) second-order stationary points (SOSP) in stochastic (non-convex) minimax optimization. Existing literature either focuses only on first-order stationary points…

Machine Learning · Computer Science 2026-02-03 Difei Xu , Youming Tao , Meng Ding , Chenglin Fan , Di Wang

Two classes of methods have been proposed for escaping from saddle points with one using the second-order information carried by the Hessian and the other adding the noise into the first-order information. The existing analysis for…

Optimization and Control · Mathematics 2018-03-05 Yi Xu , Rong Jin , Tianbao Yang

While first-order stationary points (FOSPs) are the traditional targets of non-convex optimization, they often correspond to undesirable strict saddle points. To circumvent this, attention has shifted towards second-order stationary points…

Computational Complexity · Computer Science 2026-04-03 Andreas Kontogiannis , Ioannis Panageas , Vasilis Pollatos

The Hessian-vector product has been utilized to find a second-order stationary solution with strong complexity guarantee (e.g., almost linear time complexity in the problem's dimensionality). In this paper, we propose to further reduce the…

Optimization and Control · Mathematics 2017-10-03 Mingrui Liu , Tianbao Yang

In this paper, we develop and analyze sub-sampled trust-region methods for solving finite-sum optimization problems. These methods employ subsampling strategies to approximate the gradient and Hessian of the objective function,…

Optimization and Control · Mathematics 2025-07-24 Max L. N. Goncalves , Geovani N. Grapiglia

Stochastic approximation (SA) algorithms have been widely applied in minimization problems when the loss functions and/or the gradient information are only accessible through noisy evaluations. Stochastic gradient (SG) descent---a…

Optimization and Control · Mathematics 2019-08-26 Jingyi Zhu , Long Wang , James C. Spall

This paper studies the complexity of finding an $\epsilon$-stationary point for stochastic bilevel optimization when the upper-level problem is nonconvex and the lower-level problem is strongly convex. Recent work proposed the first-order…

Optimization and Control · Mathematics 2026-03-10 Lesi Chen , Junru Li , El Mahdi Chayti , Jingzhao Zhang

We propose a novel hybrid stochastic policy gradient estimator by combining an unbiased policy gradient estimator, the REINFORCE estimator, with another biased one, an adapted SARAH estimator for policy optimization. The hybrid policy…

Machine Learning · Computer Science 2020-09-23 Nhan H. Pham , Lam M. Nguyen , Dzung T. Phan , Phuong Ha Nguyen , Marten van Dijk , Quoc Tran-Dinh

We study the sample complexity of reducing reinforcement learning to a sequence of empirical risk minimization problems over the policy space. Such reductions-based algorithms exhibit local convergence in the function space, as opposed to…

Machine Learning · Computer Science 2023-01-26 Naman Agarwal , Brian Bullins , Karan Singh

In this paper, we propose a novel reinforcement- learning algorithm consisting in a stochastic variance-reduced version of policy gradient for solving Markov Decision Processes (MDPs). Stochastic variance-reduced gradient (SVRG) methods…

Machine Learning · Computer Science 2018-06-15 Matteo Papini , Damiano Binaghi , Giuseppe Canonaco , Matteo Pirotta , Marcello Restelli

Efficiently approximating local curvature information of the loss function is a key tool for optimization and compression of deep neural networks. Yet, most existing methods to approximate second-order information have high computational or…

Machine Learning · Computer Science 2021-11-19 Elias Frantar , Eldar Kurtic , Dan Alistarh

Stochastically controlled stochastic gradient (SCSG) methods have been proved to converge efficiently to first-order stationary points which, however, can be saddle points in nonconvex optimization. It has been observed that a stochastic…

Optimization and Control · Mathematics 2021-04-26 Guannan Liang , Qianqian Tong , Chunjiang Zhu , Jinbo Bi

Policy gradient (PG) algorithms have been widely used in reinforcement learning (RL). However, PG algorithms rely on exploiting the value function being learned with the first-order update locally, which results in limited sample…

Machine Learning · Computer Science 2021-07-06 Hao Sun , Ziping Xu , Yuhang Song , Meng Fang , Jiechao Xiong , Bo Dai , Bolei Zhou

In this work, we consider bilevel optimization when the lower-level problem is strongly convex. Recent works show that with a Hessian-vector product (HVP) oracle, one can provably find an $\epsilon$-stationary point within…

Optimization and Control · Mathematics 2026-05-26 Lesi Chen , Yaohua Ma , Jingzhao Zhang

In this paper, we give a sharp analysis for Stochastic Gradient Descent (SGD) and prove that SGD is able to efficiently escape from saddle points and find an $(\epsilon, O(\epsilon^{0.5}))$-approximate second-order stationary point in…

Optimization and Control · Mathematics 2019-06-05 Cong Fang , Zhouchen Lin , Tong Zhang

Finding an approximate second-order stationary point (SOSP) is a well-studied and fundamental problem in stochastic nonconvex optimization with many applications in machine learning. However, this problem is poorly understood in the…

Optimization and Control · Mathematics 2024-03-19 Shuyao Li , Yu Cheng , Ilias Diakonikolas , Jelena Diakonikolas , Rong Ge , Stephen J. Wright

In modern deep learning, highly subsampled stochastic approximation (SA) methods are preferred to sample average approximation (SAA) methods because of large data sets as well as generalization properties. Additionally, due to perceived…

Optimization and Control · Mathematics 2021-08-26 Thomas O'Leary-Roseberry , Nick Alger , Omar Ghattas

We present a comprehensive theoretical analysis of first-order methods for escaping strict saddle points in smooth non-convex optimization. Our main contribution is a Perturbed Saddle-escape Descent (PSD) algorithm with fully explicit…

Machine Learning · Computer Science 2025-08-25 Faruk Alpay , Hamdi Alakkad