English
Related papers

Related papers: Safe Policy Improvement with Soft Baseline Bootstr…

200 papers

This paper considers Safe Policy Improvement (SPI) in Batch Reinforcement Learning (Batch RL): from a fixed dataset and without direct access to the true environment, train a policy that is guaranteed to perform at least as well as the…

Machine Learning · Computer Science 2019-06-11 Romain Laroche , Paul Trichelair , Rémi Tachet des Combes

We study the problem of Safe Policy Improvement (SPI) under constraints in the offline Reinforcement Learning (RL) setting. We consider the scenario where: (i) we have a dataset collected under a known baseline policy, (ii) multiple reward…

Machine Learning · Computer Science 2021-11-01 Harsh Satija , Philip S. Thomas , Joelle Pineau , Romain Laroche

In an offline reinforcement learning setting, the safe policy improvement (SPI) problem aims to improve the performance of a behavior policy according to which sample data has been generated. State-of-the-art approaches to SPI require a…

Machine Learning · Computer Science 2023-05-16 Patrick Wienhöft , Marnix Suilen , Thiago D. Simão , Clemens Dubslaff , Christel Baier , Nils Jansen

Safe Policy Improvement (SPI) aims at provable guarantees that a learned policy is at least approximately as good as a given baseline policy. Building on SPI with Soft Baseline Bootstrapping (Soft-SPIBB) by Nadjahi et al., we identify…

Machine Learning · Computer Science 2022-08-02 Philipp Scholl , Felix Dietrich , Clemens Otte , Steffen Udluft

Safe Policy Improvement (SPI) is an important technique for offline reinforcement learning in safety critical applications as it improves the behavior policy with a high probability. We classify various SPI approaches from the literature…

Machine Learning · Computer Science 2022-08-02 Philipp Scholl , Felix Dietrich , Clemens Otte , Steffen Udluft

Previous work has shown the unreliability of existing algorithms in the batch Reinforcement Learning setting, and proposed the theoretically-grounded Safe Policy Improvement with Baseline Bootstrapping (SPIBB) fix: reproduce the baseline…

Machine Learning · Computer Science 2021-01-01 Thiago D. Simão , Romain Laroche , Rémi Tachet des Combes

In offline reinforcement learning (RL), we learn policies from fixed datasets without environment interaction. The major challenges are to provide guarantees on the (1) performance and (2) safety of the resulting policy. A technique called…

Machine Learning · Computer Science 2026-05-12 Maris F. L. Galesloot , Thomas Rhemrev , Nils Jansen

Within batch reinforcement learning, safe policy improvement (SPI) seeks to ensure that the learnt policy performs at least as well as the behavior policy that generated the dataset. The core challenge in SPI is seeking improvements while…

Machine Learning · Computer Science 2024-10-15 Abhishek Sharma , Leo Benac , Sonali Parbhoo , Finale Doshi-Velez

Batch reinforcement learning (RL) is important to apply RL algorithms to many high stakes tasks. Doing batch RL in a way that yields a reliable new policy in large domains is challenging: a new decision policy may visit states and actions…

Machine Learning · Computer Science 2020-07-23 Yao Liu , Adith Swaminathan , Alekh Agarwal , Emma Brunskill

Safe policy improvement (SPI) offers theoretical control over policy updates, yet existing guarantees largely concern offline, tabular reinforcement learning (RL). We study SPI in general online settings, when combined with world model and…

Machine Learning · Computer Science 2026-01-29 Florent Delgrange , Raphael Avalos , Willem Röpke

When modifying existing policies in high-risk settings, it is often necessary to ensure with high certainty that the newly proposed policy improves upon a baseline, such as the status quo. In this work, we consider the problem of safe…

Machine Learning · Computer Science 2024-08-23 Brian M Cho , Ana-Roxana Pop , Kyra Gan , Sam Corbett-Davies , Israel Nir , Ariel Evnine , Nathan Kallus

We propose a policy improvement algorithm for Reinforcement Learning (RL) which is called Rerouted Behavior Improvement (RBI). RBI is designed to take into account the evaluation errors of the Q-function. Such errors are common in RL when…

Machine Learning · Computer Science 2019-07-12 Elad Sarafian , Aviv Tamar , Sarit Kraus

Safe policy improvement (SPI) is an offline reinforcement learning problem in which a new policy that reliably outperforms the behavior policy with high confidence needs to be computed using only a dataset and the behavior policy. Markov…

Artificial Intelligence · Computer Science 2025-08-20 Kasper Engelen , Guillermo A. Pérez , Marnix Suilen

We study safe policy improvement (SPI) for partially observable Markov decision processes (POMDPs). SPI is an offline reinforcement learning (RL) problem that assumes access to (1) historical data about an environment, and (2) the so-called…

Artificial Intelligence · Computer Science 2023-01-13 Thiago D. Simão , Marnix Suilen , Nils Jansen

Safe reinforcement learning (RL) aims to learn policies that satisfy certain constraints before deploying them to safety-critical applications. Previous primal-dual style approaches suffer from instability issues and lack optimality…

Machine Learning · Computer Science 2022-06-20 Zuxin Liu , Zhepeng Cen , Vladislav Isenbaev , Wei Liu , Zhiwei Steven Wu , Bo Li , Ding Zhao

The recent success of reinforcement learning's (RL) in solving complex tasks is most often attributed to its capacity to explore and exploit an environment where it has been trained. Sample efficiency is usually not an issue since cheap…

Computation and Language · Computer Science 2023-03-16 Govardana Sachithanandam Ramachandran , Kazuma Hashimoto , Caiming Xiong

Safety is the priority concern when applying reinforcement learning (RL) algorithms to real-world control problems. While policy iteration provides a fundamental algorithm for standard RL, an analogous theoretical algorithm for safe RL…

Machine Learning · Computer Science 2025-03-14 Yujie Yang , Zhilong Zheng , Shengbo Eben Li , Wei Xu , Jingjing Liu , Xianyuan Zhan , Ya-Qin Zhang

In optimal control problem, policy iteration (PI) is a powerful reinforcement learning (RL) tool used for designing optimal controller for the linear systems. However, the need for an initial stabilizing control policy significantly limits…

Optimization and Control · Mathematics 2024-11-13 Zhen Pang , Shengda Tang , Jun Cheng , Shuping He

We consider the problem of reinforcement learning when provided with (1) a baseline control policy and (2) a set of constraints that the learner must satisfy. The baseline policy can arise from demonstration data or a teacher agent and may…

Machine Learning · Computer Science 2021-07-13 Tsung-Yen Yang , Justinian Rosca , Karthik Narasimhan , Peter J. Ramadge

In batch reinforcement learning (RL), one often constrains a learned policy to be close to the behavior (data-generating) policy, e.g., by constraining the learned action distribution to differ from the behavior policy by some maximum…

Machine Learning · Computer Science 2020-03-31 Sungryull Sohn , Yinlam Chow , Jayden Ooi , Ofir Nachum , Honglak Lee , Ed Chi , Craig Boutilier
‹ Prev 1 2 3 10 Next ›