English
Related papers

Related papers: Two Time-scale Off-Policy TD Learning: Non-asympto…

200 papers

The convergence of stochastic gradient descent is highly dependent on the step-size, especially on non-convex problems such as neural network training. Step decay step-size schedules (constant and then cut) are widely used in practice…

Optimization and Control · Mathematics 2021-02-19 Xiaoyu Wang , Sindri Magnússon , Mikael Johansson

We study gradient descent (GD) dynamics on logistic regression problems with large, constant step sizes. For linearly-separable data, it is known that GD converges to the minimizer with arbitrarily large step sizes, a property which no…

Machine Learning · Computer Science 2024-11-05 Si Yi Meng , Antonio Orvieto , Daniel Yiming Cao , Christopher De Sa

This paper studies the policy mirror descent (PMD) method, which is a general policy optimization framework in reinforcement learning and can cover a wide range of policy gradient methods by specifying difference mirror maps. Existing…

Optimization and Control · Mathematics 2026-01-01 Wenye Li , Hongxu Chen , Jiacai Liu , Ke Wei

In this paper, we study the Temporal Difference (TD) learning with linear value function approximation. It is well known that most TD learning algorithms are unstable with linear function approximation and off-policy learning. Recent…

Artificial Intelligence · Computer Science 2016-10-06 Dominik Meyer , Hao Shen , Klaus Diepold

We study gradient descent (GD) with a constant stepsize for $\ell_2$-regularized logistic regression with linearly separable data. Classical theory suggests small stepsizes to ensure monotonic reduction of the optimization objective,…

Machine Learning · Statistics 2025-11-04 Jingfeng Wu , Pierre Marion , Peter Bartlett

Differentially private (DP) linear regression has received significant attention in the recent theoretical literature, with several approaches proposed to improve error rates. Our work considers the popular high-dimensional regime with…

Machine Learning · Statistics 2026-04-28 Simone Bombari , Jialei Luo , Inbar Seroussi , Marco Mondelli

Adaptive gradient methods have attracted much attention of machine learning communities due to the high efficiency. However their acceleration effect in practice, especially in neural network training, is hard to analyze, theoretically. The…

Optimization and Control · Mathematics 2020-06-15 Xunpeng Huang , Hao Zhou , Runxin Xu , Zhe Wang , Lei Li

In the framework of Markov Decision Processes, off-policy learning, that is the problem of learning a linear approximation of the value function of some fixed policy from one trajectory possibly generated by some other policy. We briefly…

Artificial Intelligence · Computer Science 2013-04-16 Matthieu Geist , Bruno Scherrer

This paper addresses the issue of policy evaluation in Markov Decision Processes, using linear function approximation. It provides a unified view of algorithms such as TD(lambda), LSTD(lambda), iLSTD, residual-gradient TD. It is asserted…

Machine Learning · Computer Science 2007-05-23 Manuel Loth , Philippe Preux

This work investigates stepsize-based acceleration of gradient descent with {\em anytime} convergence guarantees. For smooth (non-strongly) convex optimization, we propose a stepsize schedule that allows gradient descent to achieve…

Machine Learning · Computer Science 2024-12-10 Zihan Zhang , Jason D. Lee , Simon S. Du , Yuxin Chen

The analysis of Temporal Difference (TD) learning in the average-reward setting faces notable theoretical difficulties because the Bellman operator is not contractive with respect to any norm. This complicates standard analyses of…

Machine Learning · Computer Science 2026-05-05 Haoxing Tian , Zaiwei Chen , Ioannis Ch. Paschalidis , Alex Olshevsky

The problem of on-line off-policy evaluation (OPE) has been actively studied in the last decade due to its importance both as a stand-alone problem and as a module in a policy improvement scheme. However, most Temporal Difference (TD) based…

Machine Learning · Statistics 2017-02-24 Assaf Hallak , Shie Mannor

Temporal-difference learning is a popular algorithm for policy evaluation. In this paper, we study the convergence of the regularized non-parametric TD(0) algorithm, in both the independent and Markovian observation settings. In particular,…

Optimization and Control · Mathematics 2022-05-25 Eloïse Berthier , Ziad Kobeissi , Francis Bach

We propose new continuous-time formulations for first-order stochastic optimization algorithms such as mini-batch gradient descent and variance-reduced methods. We exploit these continuous-time models, together with simple Lyapunov analysis…

Optimization and Control · Mathematics 2020-03-12 Antonio Orvieto , Aurelien Lucchi

Adaptive gradient optimizers (AdaGrad), which dynamically adjust the learning rate based on iterative gradients, have emerged as powerful tools in deep learning. These adaptive methods have significantly succeeded in various deep learning…

Optimization and Control · Mathematics 2024-12-31 Ruinan Jin , Xiaoyu Wang , Baoxiang Wang

We study the convergence behavior of the celebrated temporal-difference (TD) learning algorithm. By looking at the algorithm through the lens of optimization, we first argue that TD can be viewed as an iterative optimization algorithm where…

Machine Learning · Computer Science 2023-11-10 Kavosh Asadi , Shoham Sabach , Yao Liu , Omer Gottesman , Rasool Fakoor

Temporal difference (TD) learning is a popular algorithm for policy evaluation in reinforcement learning, but the vanilla TD can substantially suffer from the inherent optimization variance. A variance reduced TD (VRTD) algorithm was…

Machine Learning · Computer Science 2020-01-13 Tengyu Xu , Zhe Wang , Yi Zhou , Yingbin Liang

Decentralized optimization has become vital for leveraging distributed data without central control, enhancing scalability and privacy. However, practical deployments face fundamental challenges due to heterogeneous computation speeds and…

Machine Learning · Computer Science 2025-05-16 Yijie Zhou , Shi Pu

Gradient descent (GD) is a collection of continuous optimization methods that have achieved immeasurable success in practice. Owing to data science applications, GD with diminishing step sizes has become a prominent variant. While this…

Optimization and Control · Mathematics 2023-06-27 Vivak Patel , Albert S. Berahas

Previous studies on two-timescale stochastic approximation (SA) mainly focused on bounding mean-squared errors under diminishing stepsize schemes. In this work, we investigate {\it constant} stpesize schemes through the lens of Markov…

Systems and Control · Electrical Eng. & Systems 2025-02-25 Jeongyeol Kwon , Luke Dotson , Yudong Chen , Qiaomin Xie