RA-PbRL: Provably Efficient Risk-Aware Preference-Based Reinforcement Learning
- URL: http://arxiv.org/abs/2410.23569v1
- Date: Thu, 31 Oct 2024 02:25:43 GMT
- Title: RA-PbRL: Provably Efficient Risk-Aware Preference-Based Reinforcement Learning
- Authors: Yujie Zhao, Jose Efraim Aguilar Escamill, Weyl Lu, Huazheng Wang,
- Abstract summary: Preference-based Reinforcement Learning (PbRL) studies the problem where agents receive only preferences over pairs of trajectories in each episode.
Traditional risk-aware objectives and algorithms are not applicable in such one-episode-reward settings.
We introduce Risk-Aware- PbRL, an algorithm designed to optimize both nested and static objectives.
- Score: 7.407106653769627
- License:
- Abstract: Preference-based Reinforcement Learning (PbRL) studies the problem where agents receive only preferences over pairs of trajectories in each episode. Traditional approaches in this field have predominantly focused on the mean reward or utility criterion. However, in PbRL scenarios demanding heightened risk awareness, such as in AI systems, healthcare, and agriculture, risk-aware measures are requisite. Traditional risk-aware objectives and algorithms are not applicable in such one-episode-reward settings. To address this, we explore and prove the applicability of two risk-aware objectives to PbRL: nested and static quantile risk objectives. We also introduce Risk-Aware- PbRL (RA-PbRL), an algorithm designed to optimize both nested and static objectives. Additionally, we provide a theoretical analysis of the regret upper bounds, demonstrating that they are sublinear with respect to the number of episodes, and present empirical results to support our findings. Our code is available in https://github.com/aguilarjose11/PbRLNeurips.
Related papers
- Evaluating the Paperclip Maximizer: Are RL-Based Language Models More Likely to Pursue Instrumental Goals? [33.11148546999906]
Key concern is textitinstrumental convergence, where an AI system develops unintended intermediate goals that override the ultimate objective and deviate from human-intended goals.
This issue is particularly relevant in reinforcement learning (RL)-trained models, which can generate creative but unintended strategies to maximize rewards.
We show that RL-driven models exhibit a stronger tendency for instrumental convergence due to their optimization of goal-directed behavior in ways that may misalign with human intentions.
arXiv Detail & Related papers (2025-02-16T16:29:20Z) - Iterative Nash Policy Optimization: Aligning LLMs with General Preferences via No-Regret Learning [55.65738319966385]
We propose a novel online algorithm, iterative Nash policy optimization (INPO)
Unlike previous methods, INPO bypasses the need for estimating the expected win rate for individual responses.
With an LLaMA-3-8B-based SFT model, INPO achieves a 42.6% length-controlled win rate on AlpacaEval 2.0 and a 37.8% win rate on Arena-Hard.
arXiv Detail & Related papers (2024-06-30T08:00:34Z) - REBEL: Reward Regularization-Based Approach for Robotic Reinforcement Learning from Human Feedback [61.54791065013767]
A misalignment between the reward function and human preferences can lead to catastrophic outcomes in the real world.
Recent methods aim to mitigate misalignment by learning reward functions from human preferences.
We propose a novel concept of reward regularization within the robotic RLHF framework.
arXiv Detail & Related papers (2023-12-22T04:56:37Z) - Contrastive Preference Learning: Learning from Human Feedback without RL [71.77024922527642]
We introduce Contrastive Preference Learning (CPL), an algorithm for learning optimal policies from preferences without learning reward functions.
CPL is fully off-policy, uses only a simple contrastive objective, and can be applied to arbitrary MDPs.
arXiv Detail & Related papers (2023-10-20T16:37:56Z) - A Multiplicative Value Function for Safe and Efficient Reinforcement
Learning [131.96501469927733]
We propose a safe model-free RL algorithm with a novel multiplicative value function consisting of a safety critic and a reward critic.
The safety critic predicts the probability of constraint violation and discounts the reward critic that only estimates constraint-free returns.
We evaluate our method in four safety-focused environments, including classical RL benchmarks augmented with safety constraints and robot navigation tasks with images and raw Lidar scans as observations.
arXiv Detail & Related papers (2023-03-07T18:29:15Z) - B-Pref: Benchmarking Preference-Based Reinforcement Learning [84.41494283081326]
We introduce B-Pref, a benchmark specially designed for preference-based RL.
A key challenge with such a benchmark is providing the ability to evaluate candidate algorithms quickly.
B-Pref alleviates this by simulating teachers with a wide array of irrationalities.
arXiv Detail & Related papers (2021-11-04T17:32:06Z) - Preference-based Reinforcement Learning with Finite-Time Guarantees [76.88632321436472]
Preference-based Reinforcement Learning (PbRL) replaces reward values in traditional reinforcement learning to better elicit human opinion on the target objective.
Despite promising results in applications, the theoretical understanding of PbRL is still in its infancy.
We present the first finite-time analysis for general PbRL problems.
arXiv Detail & Related papers (2020-06-16T03:52:41Z)
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.