Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone

• URL: http://arxiv.org/abs/2412.06685v1
• Date: Mon, 09 Dec 2024 17:28:03 GMT
• Title: Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone
• Authors: Max Sobol Mark, Tian Gao, Georgia Gabriela Sampaio, Mohan Kumar Srirama, Archit Sharma, Chelsea Finn, Aviral Kumar,
• Abstract summary: We develop an offline and online fine-tuning approach called policy-agnostic RL (PA-RL) We show the first result that successfully fine-tunes OpenVLA, a 7B generalist robot policy, autonomously with Cal-QL, an online RL fine-tuning algorithm.
• Score: 72.17534881026995
• License:
• Abstract: Recent advances in learning decision-making policies can largely be attributed to training expressive policy models, largely via imitation learning. While imitation learning discards non-expert data, reinforcement learning (RL) can still learn from suboptimal data. However, instantiating RL training of a new policy class often presents a different challenge: most deep RL machinery is co-developed with assumptions on the policy class and backbone, resulting in poor performance when the policy class changes. For instance, SAC utilizes a low-variance reparameterization policy gradient for Gaussian policies, but this is unstable for diffusion policies and intractable for autoregressive categorical policies. To address this issue, we develop an offline RL and online fine-tuning approach called policy-agnostic RL (PA-RL) that can effectively train multiple policy classes, with varying architectures and sizes. We build off the basic idea that a universal supervised learning loss can replace the policy improvement step in RL, as long as it is applied on "optimized" actions. To obtain these optimized actions, we first sample multiple actions from a base policy, and run global optimization (i.e., re-ranking multiple action samples using the Q-function) and local optimization (i.e., running gradient steps on an action sample) to maximize the critic on these candidates. PA-RL enables fine-tuning diffusion and transformer policies with either autoregressive tokens or continuous action outputs, at different sizes, entirely via actor-critic RL. Moreover, PA-RL improves the performance and sample-efficiency by up to 2 times compared to existing offline RL and online fine-tuning methods. We show the first result that successfully fine-tunes OpenVLA, a 7B generalist robot policy, autonomously with Cal-QL, an online RL fine-tuning algorithm, improving from 40% to 70% in the real world in 40 minutes.

Related papers

• Is Value Learning Really the Main Bottleneck in Offline RL? [70.54708989409409]
  We show that the choice of a policy extraction algorithm significantly affects the performance and scalability of offline RL. We propose two simple test-time policy improvement methods and show that these methods lead to better performance.
  arXiv  Detail & Related papers  (2024-06-13T17:07:49Z)
• Offline-Boosted Actor-Critic: Adaptively Blending Optimal Historical Behaviors in Deep Off-Policy RL [42.57662196581823]
  Off-policy reinforcement learning (RL) has achieved notable success in tackling many complex real-world tasks. Most existing off-policy RL algorithms fail to maximally exploit the information in the replay buffer. We present Offline-Boosted Actor-Critic (OBAC), a model-free online RL framework that elegantly identifies the outperforming offline policy.
  arXiv  Detail & Related papers  (2024-05-28T18:38:46Z)
• Behavior Proximal Policy Optimization [14.701955559885615]
  offline reinforcement learning (RL) is a challenging setting where existing off-policy actor-critic methods perform poorly. Online on-policy algorithms are naturally able to solve offline RL. We propose Behavior Proximal Policy Optimization (BPPO), which solves offline RL without any extra constraint or regularization.
  arXiv  Detail & Related papers  (2023-02-22T11:49:12Z)
• Offline RL With Realistic Datasets: Heteroskedasticity and Support Constraints [82.43359506154117]
  We show that typical offline reinforcement learning methods fail to learn from data with non-uniform variability. Our method is simple, theoretically motivated, and improves performance across a wide range of offline RL problems in Atari games, navigation, and pixel-based manipulation.
  arXiv  Detail & Related papers  (2022-11-02T11:36:06Z)
• Jump-Start Reinforcement Learning [68.82380421479675]
  We present a meta algorithm that can use offline data, demonstrations, or a pre-existing policy to initialize an RL policy. In particular, we propose Jump-Start Reinforcement Learning (JSRL), an algorithm that employs two policies to solve tasks. We show via experiments that JSRL is able to significantly outperform existing imitation and reinforcement learning algorithms.
  arXiv  Detail & Related papers  (2022-04-05T17:25:22Z)
• POPO: Pessimistic Offline Policy Optimization [6.122342691982727]
  We study why off-policy RL methods fail to learn in offline setting from the value function view. We propose Pessimistic Offline Policy Optimization (POPO), which learns a pessimistic value function to get a strong policy. We find that POPO performs surprisingly well and scales to tasks with high-dimensional state and action space.
  arXiv  Detail & Related papers  (2020-12-26T06:24:34Z)
• MOPO: Model-based Offline Policy Optimization [183.6449600580806]
  offline reinforcement learning (RL) refers to the problem of learning policies entirely from a large batch of previously collected data. We show that an existing model-based RL algorithm already produces significant gains in the offline setting. We propose to modify the existing model-based RL methods by applying them with rewards artificially penalized by the uncertainty of the dynamics.
  arXiv  Detail & Related papers  (2020-05-27T08:46:41Z)
• BRPO: Batch Residual Policy Optimization [79.53696635382592]
  In batch reinforcement learning, one often constrains a learned policy to be close to the behavior (data-generating) policy. We propose residual policies, where the allowable deviation of the learned policy is state-action-dependent. We derive a new for RL method, BRPO, which learns both the policy and allowable deviation that jointly maximize a lower bound on policy performance.
  arXiv  Detail & Related papers  (2020-02-08T01:59:33Z)

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.