English
Related papers

Related papers: A Generalized Bootstrap Target for Value-Learning,…

200 papers

Multi-step temporal difference (TD) learning is an important approach in reinforcement learning, as it unifies one-step TD learning with Monte Carlo methods in a way where intermediate algorithms can outperform either extreme. They address…

Machine Learning · Computer Science 2018-09-10 Kristopher De Asis , Richard S. Sutton

Reinforcement learning systems have the potential to enable continuous improvement in unstructured environments, leveraging data collected autonomously. However, in practice these systems require significant amounts of instrumentation or…

Robotics · Computer Science 2022-03-30 Abhishek Gupta , Corey Lynch , Brandon Kinman , Garrett Peake , Sergey Levine , Karol Hausman

To estimate the value functions of policies from exploratory data, most model-free off-policy algorithms rely on importance sampling, where the use of importance sampling ratios often leads to estimates with severe variance. It is thus…

Machine Learning · Computer Science 2017-02-13 Ashique Rupam Mahmood , Huizhen Yu , Richard S. Sutton

Temporal difference learning (TD) is a simple iterative algorithm used to estimate the value function corresponding to a given policy in a Markov decision process. Although TD is one of the most widely used algorithms in reinforcement…

Machine Learning · Computer Science 2018-11-07 Jalaj Bhandari , Daniel Russo , Raghav Singal

Quantifying the value of data is a fundamental problem in machine learning. Data valuation has multiple important use cases: (1) building insights about the learning task, (2) domain adaptation, (3) corrupted sample discovery, and (4)…

Machine Learning · Computer Science 2019-09-27 Jinsung Yoon , Sercan O. Arik , Tomas Pfister

Meta-learning empowers artificial intelligence to increase its efficiency by learning how to learn. Unlocking this potential involves overcoming a challenging meta-optimisation problem. We propose an algorithm that tackles this problem by…

Machine Learning · Computer Science 2022-03-17 Sebastian Flennerhag , Yannick Schroecker , Tom Zahavy , Hado van Hasselt , David Silver , Satinder Singh

Bootstrapping is behind much of the successes of deep Reinforcement Learning. However, learning the value function via bootstrapping often leads to unstable training due to fast-changing target values. Target Networks are employed to…

Reinforcement learning has been successful across several applications in which agents have to learn to act in environments with sparse feedback. However, despite this empirical success there is still a lack of theoretical understanding of…

Machine Learning · Statistics 2023-11-08 Blake Bordelon , Paul Masset , Henry Kuo , Cengiz Pehlevan

We study the problem of temporal-difference-based policy evaluation in reinforcement learning. In particular, we analyse the use of a distributional reinforcement learning algorithm, quantile temporal-difference learning (QTD), for this…

Machine Learning · Computer Science 2023-05-31 Mark Rowland , Yunhao Tang , Clare Lyle , Rémi Munos , Marc G. Bellemare , Will Dabney

Sample efficiency is central to developing practical reinforcement learning (RL) for complex and large-scale decision-making problems. The ability to transfer and generalize knowledge gained from previous experiences to downstream tasks can…

Machine Learning · Computer Science 2024-07-23 Parvin Malekzadeh , Ming Hou , Konstantinos N. Plataniotis

We propose a method for meta-learning reinforcement learning algorithms by searching over the space of computational graphs which compute the loss function for a value-based model-free RL agent to optimize. The learned algorithms are…

Machine Learning · Computer Science 2022-11-11 John D. Co-Reyes , Yingjie Miao , Daiyi Peng , Esteban Real , Sergey Levine , Quoc V. Le , Honglak Lee , Aleksandra Faust

Off-policy learning allows us to learn about possible policies of behavior from experience generated by a different behavior policy. Temporal difference (TD) learning algorithms can become unstable when combined with function approximation…

Machine Learning · Computer Science 2021-06-23 Ray Jiang , Tom Zahavy , Zhongwen Xu , Adam White , Matteo Hessel , Charles Blundell , Hado van Hasselt

Reinforcement learning (RL) tackles sequential decision-making problems by creating agents that interacts with their environment. However, existing algorithms often view these problem as static, focusing on point estimates for model…

Machine Learning · Statistics 2024-03-21 Frank Shih , Faming Liang

We investigate the statistical properties of Temporal Difference (TD) learning with Polyak-Ruppert averaging, arguably one of the most widely used algorithms in reinforcement learning, for the task of estimating the parameters of the…

Machine Learning · Statistics 2026-02-25 Weichen Wu , Gen Li , Yuting Wei , Alessandro Rinaldo

Solving a reinforcement learning (RL) problem poses two competing challenges: fitting a potentially discontinuous value function, and generalizing well to new observations. In this paper, we analyze the learning dynamics of temporal…

Machine Learning · Computer Science 2022-06-07 Clare Lyle , Mark Rowland , Will Dabney , Marta Kwiatkowska , Yarin Gal

Value functions derived from Markov decision processes arise as a central component of algorithms as well as performance metrics in many statistics and engineering applications of machine learning techniques. Computation of the solution to…

Machine Learning · Computer Science 2020-03-02 Adithya M. Devraj , Ioannis Kontoyiannis , Sean P. Meyn

In reinforcement learning, it is typical to use the empirically observed transitions and rewards to estimate the value of a policy via either model-based or Q-fitting approaches. Although straightforward, these techniques in general yield…

Machine Learning · Computer Science 2020-07-28 Ilya Kostrikov , Ofir Nachum

Bootstrapping and rollout are two fundamental principles for value function estimation in reinforcement learning (RL). We introduce a novel class of Bellman operators, called subgraph Bellman operators, that interpolate between…

Machine Learning · Computer Science 2024-12-02 Wenlong Mou , Jian Qian

We prove that the combination of a target network and over-parameterized linear function approximation establishes a weaker convergence condition for bootstrapped value estimation in certain cases, even with off-policy data. Our condition…

Bootstrapping is behind much of the successes of Deep Reinforcement Learning. However, learning the value function via bootstrapping often leads to unstable training due to fast-changing target values. Target Networks are employed to…