English
Related papers

Related papers: Logit Dynamics in Softmax Policy Gradient Methods

200 papers

We propose policy gradient algorithms for solving a risk-sensitive reinforcement learning (RL) problem in on-policy as well as off-policy settings. We consider episodic Markov decision processes, and model the risk using the broad class of…

Machine Learning · Computer Science 2024-06-25 Nithia Vijayan , Prashanth L. A

Having a perfect model to compute the optimal policy is often infeasible in reinforcement learning. It is important in high-stakes domains to quantify and manage risk induced by model uncertainties. Entropic risk measure is an exponential…

Machine Learning · Computer Science 2020-06-23 Reazul Hasan Russel , Bahram Behzadian , Marek Petrik

Policy gradient algorithms are widely used in reinforcement learning and belong to the class of approximate dynamic programming methods. This paper studies two key policy gradient algorithms, the Natural Policy Gradient and the Gauss-Newton…

Systems and Control · Electrical Eng. & Systems 2026-05-11 Bowen Song , Sebastien Gros , Andrea Iannelli

This paper aims to overcome a major obstacle in scaling RL for reasoning with LLMs, namely the collapse of policy entropy. Such phenomenon is consistently observed across vast RL runs without entropy intervention, where the policy entropy…

Learning to make decisions from observed data in dynamic environments remains a problem of fundamental importance in a number of fields, from artificial intelligence and robotics, to medicine and finance. This paper concerns the problem of…

Machine Learning · Statistics 2018-06-04 Jack Umenberger , Thomas B. Schön

Two main families of reinforcement learning algorithms, Q-learning and policy gradients, have recently been proven to be equivalent when using a softmax relaxation on one part, and an entropic regularization on the other. We relate this…

Machine Learning · Computer Science 2017-12-27 Pierre H. Richemond , Brendan Maginnis

Managing risk in dynamic decision problems is of cardinal importance in many fields such as finance and process control. The most common approach to defining risk is through various variance related criteria such as the Sharpe Ratio or the…

Machine Learning · Computer Science 2012-07-03 Dotan Di Castro , Aviv Tamar , Shie Mannor

In this paper, we utilize results from convex analysis and monotone operator theory to derive additional properties of the softmax function that have not yet been covered in the existing literature. In particular, we show that the softmax…

Optimization and Control · Mathematics 2018-08-23 Bolin Gao , Lacra Pavel

We propose a metalearning approach for learning gradient-based reinforcement learning (RL) algorithms. The idea is to evolve a differentiable loss function, such that an agent, which optimizes its policy to minimize this loss, will achieve…

Machine Learning · Computer Science 2018-05-01 Rein Houthooft , Richard Y. Chen , Phillip Isola , Bradly C. Stadie , Filip Wolski , Jonathan Ho , Pieter Abbeel

Multi-agent interactions are increasingly important in the context of reinforcement learning, and the theoretical foundations of policy gradient methods have attracted surging research interest. We investigate the global convergence of…

Optimization and Control · Mathematics 2023-03-21 Sarath Pattathil , Kaiqing Zhang , Asuman Ozdaglar

Policy gradient methods in reinforcement learning update policy parameters by taking steps in the direction of an estimated gradient of policy value. In this paper, we consider the statistically efficient estimation of policy gradients from…

Machine Learning · Statistics 2020-02-21 Nathan Kallus , Masatoshi Uehara

We study the problem of computing an optimal policy of an infinite-horizon discounted constrained Markov decision process (constrained MDP). Despite the popularity of Lagrangian-based policy search methods used in practice, the oscillation…

Optimization and Control · Mathematics 2024-01-18 Dongsheng Ding , Chen-Yu Wei , Kaiqing Zhang , Alejandro Ribeiro

Wasserstein policy gradient (WPG) is a policy optimization method for reinforcement learning (RL) that exploits the optimal-transport geometry of action distributions. For the entropy-regularized RL objective, WPG evolves each…

Machine Learning · Computer Science 2026-05-27 Zhaoyu Zhu , Rui Gao , Shuang Li

In this paper, we study the learning of safe policies in the setting of reinforcement learning problems. This is, we aim to control a Markov Decision Process (MDP) of which we do not know the transition probabilities, but we have access to…

Systems and Control · Electrical Eng. & Systems 2022-01-14 Santiago Paternain , Miguel Calvo-Fullana , Luiz F. O. Chamon , Alejandro Ribeiro

Traditional policy gradient methods are fundamentally flawed. Natural gradients converge quicker and better, forming the foundation of contemporary Reinforcement Learning such as Trust Region Policy Optimization (TRPO) and Proximal Policy…

Machine Learning · Computer Science 2022-09-07 W. J. A. van Heeswijk

Entropy regularized algorithms such as Soft Q-learning and Soft Actor-Critic, recently showed state-of-the-art performance on a number of challenging reinforcement learning (RL) tasks. The regularized formulation modifies the standard RL…

Machine Learning · Statistics 2019-10-15 Elena Smirnova , Elvis Dohmatob

Reinforcement learning (RL) has become a key approach for enhancing reasoning in large language models (LLMs), yet scalable training is often hindered by the rapid collapse of policy entropy, which leads to premature convergence and…

Machine Learning · Computer Science 2026-04-14 Ming Lei , Christophe Baehr

We work towards a unifying paradigm for accelerating policy optimization methods in reinforcement learning (RL) by integrating foresight in the policy improvement step via optimistic and adaptive updates. Leveraging the connection between…

Machine Learning · Computer Science 2023-09-07 Veronica Chelu , Tom Zahavy , Arthur Guez , Doina Precup , Sebastian Flennerhag

In this paper, we establish the global optimality and convergence rate of an off-policy actor critic algorithm in the tabular setting without using density ratio to correct the discrepancy between the state distribution of the behavior…

Machine Learning · Computer Science 2025-02-07 Shangtong Zhang , Remi Tachet , Romain Laroche

Direct policy gradient methods for reinforcement learning are a successful approach for a variety of reasons: they are model free, they directly optimize the performance metric of interest, and they allow for richly parameterized policies.…

Machine Learning · Computer Science 2020-08-14 Alekh Agarwal , Mikael Henaff , Sham Kakade , Wen Sun
‹ Prev 1 3 4 5 6 7 10 Next ›