e-COP : Episodic Constrained Optimization of Policies

• URL: http://arxiv.org/abs/2406.09563v1
• Date: Thu, 13 Jun 2024 20:12:09 GMT
• Title: e-COP : Episodic Constrained Optimization of Policies
• Authors: Akhil Agnihotri, Rahul Jain, Deepak Ramachandran, Sahil Singla,
• Abstract summary: We present the first policy optimization algorithm for constrained Reinforcement Learning (RL) in episodic (finite horizon) settings. We show that our algorithm has similar or better performance than SoTA (non-episodic) algorithms adapted for the episodic setting.
• Score: 12.854752753529151
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: In this paper, we present the $\texttt{e-COP}$ algorithm, the first policy optimization algorithm for constrained Reinforcement Learning (RL) in episodic (finite horizon) settings. Such formulations are applicable when there are separate sets of optimization criteria and constraints on a system's behavior. We approach this problem by first establishing a policy difference lemma for the episodic setting, which provides the theoretical foundation for the algorithm. Then, we propose to combine a set of established and novel solution ideas to yield the $\texttt{e-COP}$ algorithm that is easy to implement and numerically stable, and provide a theoretical guarantee on optimality under certain scaling assumptions. Through extensive empirical analysis using benchmarks in the Safety Gym suite, we show that our algorithm has similar or better performance than SoTA (non-episodic) algorithms adapted for the episodic setting. The scalability of the algorithm opens the door to its application in safety-constrained Reinforcement Learning from Human Feedback for Large Language or Diffusion Models.

Related papers

• Provably Mitigating Overoptimization in RLHF: Your SFT Loss is Implicitly an Adversarial Regularizer [52.09480867526656]
  We identify the source of misalignment as a form of distributional shift and uncertainty in learning human preferences. To mitigate overoptimization, we first propose a theoretical algorithm that chooses the best policy for an adversarially chosen reward model. Using the equivalence between reward models and the corresponding optimal policy, the algorithm features a simple objective that combines a preference optimization loss and a supervised learning loss.
  arXiv  Detail & Related papers  (2024-05-26T05:38:50Z)
• Stepwise Alignment for Constrained Language Model Policy Optimization [12.986006070964772]
  Safety and trustworthiness are indispensable requirements for real-world applications of AI systems using large language models (LLMs) This paper formulates human value alignment as an optimization problem of the language model policy to maximize reward under a safety constraint. One key idea behind SACPO, supported by theory, is that the optimal policy incorporating reward and safety can be directly obtained from a reward-aligned policy.
  arXiv  Detail & Related papers  (2024-04-17T03:44:58Z)
• From Optimization to Control: Quasi Policy Iteration [3.4376560669160394]
  We introduce a novel control algorithm coined as quasi-policy iteration (QPI) QPI is based on a novel approximation of the "Hessian" matrix in the policy iteration algorithm by exploiting two linear structural constraints specific to MDPs. It exhibits an empirical convergence behavior similar to policy iteration with a very low sensitivity to the discount factor.
  arXiv  Detail & Related papers  (2023-11-18T21:00:14Z)
• PARL: A Unified Framework for Policy Alignment in Reinforcement Learning from Human Feedback [106.63518036538163]
  We present a novel unified bilevel optimization-based framework, textsfPARL, formulated to address the recently highlighted critical issue of policy alignment in reinforcement learning. Our framework addressed these concerns by explicitly parameterizing the distribution of the upper alignment objective (reward design) by the lower optimal variable. Our empirical results substantiate that the proposed textsfPARL can address the alignment concerns in RL by showing significant improvements.
  arXiv  Detail & Related papers  (2023-08-03T18:03:44Z)
• Maximum-Likelihood Inverse Reinforcement Learning with Finite-Time Guarantees [56.848265937921354]
  Inverse reinforcement learning (IRL) aims to recover the reward function and the associated optimal policy. Many algorithms for IRL have an inherently nested structure. We develop a novel single-loop algorithm for IRL that does not compromise reward estimation accuracy.
  arXiv  Detail & Related papers  (2022-10-04T17:13:45Z)
• Policy Optimization for Stochastic Shortest Path [43.2288319750466]
  We study policy optimization for the shortest path (SSP) problem. We propose a goal-oriented reinforcement learning model that strictly generalizes the finite-horizon model. For most settings, our algorithm is shown to achieve a near-optimal regret bound.
  arXiv  Detail & Related papers  (2022-02-07T16:25:14Z)
• Instance-Dependent Confidence and Early Stopping for Reinforcement Learning [99.57168572237421]
  Various algorithms for reinforcement learning (RL) exhibit dramatic variation in their convergence rates as a function of problem structure. This research provides guarantees that explain textitex post the performance differences observed. A natural next step is to convert these theoretical guarantees into guidelines that are useful in practice.
  arXiv  Detail & Related papers  (2022-01-21T04:25:35Z)
• On the Optimality of Batch Policy Optimization Algorithms [106.89498352537682]
  Batch policy optimization considers leveraging existing data for policy construction before interacting with an environment. We show that any confidence-adjusted index algorithm is minimax optimal, whether it be optimistic, pessimistic or neutral. We introduce a new weighted-minimax criterion that considers the inherent difficulty of optimal value prediction.
  arXiv  Detail & Related papers  (2021-04-06T05:23:20Z)
• An Asymptotically Optimal Primal-Dual Incremental Algorithm for Contextual Linear Bandits [129.1029690825929]
  We introduce a novel algorithm improving over the state-of-the-art along multiple dimensions. We establish minimax optimality for any learning horizon in the special case of non-contextual linear bandits.
  arXiv  Detail & Related papers  (2020-10-23T09:12:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.