Robust Reward Alignment via Hypothesis Space Batch Cutting

• URL: http://arxiv.org/abs/2502.02921v2
• Date: Thu, 06 Feb 2025 18:08:43 GMT
• Title: Robust Reward Alignment via Hypothesis Space Batch Cutting
• Abstract summary: We propose a robust and efficient reward alignment method based on a novel and geometrically interpretable perspective.<n>Our method iteratively refines the reward hypothesis space through "cuts" based on batches of human preferences.<n>To handle unknown erroneous preferences, we introduce a conservative cutting method within each batch.
• Score: 6.4274405516321655
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reward design for reinforcement learning and optimal control agents is challenging. Preference-based alignment addresses this by enabling agents to learn rewards from ranked trajectory pairs provided by humans. However, existing methods often struggle from poor robustness to unknown false human preferences. In this work, we propose a robust and efficient reward alignment method based on a novel and geometrically interpretable perspective: hypothesis space batched cutting. Our method iteratively refines the reward hypothesis space through "cuts" based on batches of human preferences. Within each batch, human preferences, queried based on disagreement, are grouped using a voting function to determine the appropriate cut, ensuring a bounded human query complexity. To handle unknown erroneous preferences, we introduce a conservative cutting method within each batch, preventing erroneous human preferences from making overly aggressive cuts to the hypothesis space. This guarantees provable robustness against false preferences. We evaluate our method in a model predictive control setting across diverse tasks, including DM-Control, dexterous in-hand manipulation, and locomotion. The results demonstrate that our framework achieves comparable or superior performance to state-of-the-art methods in error-free settings while significantly outperforming existing method when handling high percentage of erroneous human preferences.

Related papers

• Aligning Deep Implicit Preferences by Learning to Reason Defensively [22.548051297731416]
  We propose Critique-Driven Reasoning Alignment (CDRA) to bridge the preference inference gap.<n>CDRA reframes alignment from a scalar reward-matching task into a structured reasoning process.<n> Experiments demonstrate that CDRA excels at discovering and aligning with users' true preferences while executing robust reasoning.
  arXiv  Detail & Related papers  (2025-10-13T09:26:47Z)
• A Principled Approach to Randomized Selection under Uncertainty: Applications to Peer Review and Grant Funding [68.43987626137512]
  We propose a principled framework for randomized decision-making based on interval estimates of the quality of each item.<n>We introduce MERIT, an optimization-based method that maximizes the worst-case expected number of top candidates selected.<n>We prove that MERIT satisfies desirable axiomatic properties not guaranteed by existing approaches.
  arXiv  Detail & Related papers  (2025-06-23T19:59:30Z)
• PB$^2$: Preference Space Exploration via Population-Based Methods in Preference-Based Reinforcement Learning [2.0373030742807545]
  We identify and address this preference exploration problem through population-based methods.<n>We demonstrate that maintaining a diverse population of agents enables more comprehensive exploration of the preference landscape.<n>This diversity improves reward model learning by generating preference queries with clearly distinguishable behaviors.
  arXiv  Detail & Related papers  (2025-06-16T17:51:33Z)
• Direct Alignment with Heterogeneous Preferences [11.693372619696683]
  We show that aligning to heterogeneous preferences with a single policy is best achieved using the average reward across user types.<n>We find that minimal information can yield first-order improvements, while full feedback from each user type leads to consistent learning of the optimal policy.
  arXiv  Detail & Related papers  (2025-02-22T18:46:33Z)
• Adaptive Dense Reward: Understanding the Gap Between Action and Reward Space in Alignment [33.5805074836187]
  Reinforcement Learning from Human Feedback (RLHF) has proven highly effective in aligning Large Language Models (LLMs) with human preferences.<n>This limitation stems from RLHF's lack of awareness regarding which specific tokens should be reinforced or suppressed.<n>We propose the Adaptive Message-wise RLHF'' method, which robustly applies to various tasks.
  arXiv  Detail & Related papers  (2024-10-23T16:16:15Z)
• Optimal Design for Reward Modeling in RLHF [83.3614658277817]
  We formalize the reward training model in Reinforcement Learning from Human Feedback. We frame the selection of an effective dataset as a simple regret minimization task. We derive bounds on the simple regret under appropriate assumptions.
  arXiv  Detail & Related papers  (2024-10-22T14:36:44Z)
• Sequential Manipulation Against Rank Aggregation: Theory and Algorithm [119.57122943187086]
  We leverage an online attack on the vulnerable data collection process. From the game-theoretic perspective, the confrontation scenario is formulated as a distributionally robust game. The proposed method manipulates the results of rank aggregation methods in a sequential manner.
  arXiv  Detail & Related papers  (2024-07-02T03:31:21Z)
• A Minimaximalist Approach to Reinforcement Learning from Human Feedback [49.45285664482369]
  We present Self-Play Preference Optimization (SPO), an algorithm for reinforcement learning from human feedback. Our approach is minimalist in that it does not require training a reward model nor unstable adversarial training. We demonstrate that on a suite of continuous control tasks, we are able to learn significantly more efficiently than reward-model based approaches.
  arXiv  Detail & Related papers  (2024-01-08T17:55:02Z)
• 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)
• Fair Classifiers that Abstain without Harm [24.90899074869189]
  In critical applications, it is vital for classifiers to defer decision-making to humans. We propose a post-hoc method that makes existing classifiers selectively abstain from predicting certain samples. Our framework outperforms existing methods in terms of fairness disparity without sacrificing accuracy at similar abstention rates.
  arXiv  Detail & Related papers  (2023-10-09T23:07:28Z)
• Provable Benefits of Policy Learning from Human Preferences in Contextual Bandit Problems [82.92678837778358]
  preference-based methods have demonstrated substantial success in empirical applications such as InstructGPT. We show how human bias and uncertainty in feedback modelings can affect the theoretical guarantees of these approaches.
  arXiv  Detail & Related papers  (2023-07-24T17:50:24Z)
• Reinforcement Learning from Diverse Human Preferences [68.4294547285359]
  This paper develops a method for crowd-sourcing preference labels and learning from diverse human preferences. The proposed method is tested on a variety of tasks in DMcontrol and Meta-world. It has shown consistent and significant improvements over existing preference-based RL algorithms when learning from diverse feedback.
  arXiv  Detail & Related papers  (2023-01-27T15:18:54Z)
• Holistic Approach to Measure Sample-level Adversarial Vulnerability and its Utility in Building Trustworthy Systems [17.707594255626216]
  Adversarial attack perturbs an image with an imperceptible noise, leading to incorrect model prediction. We propose a holistic approach for quantifying adversarial vulnerability of a sample by combining different perspectives. We demonstrate that by reliably estimating adversarial vulnerability at the sample level, it is possible to develop a trustworthy system.
  arXiv  Detail & Related papers  (2022-05-05T12:36:17Z)

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.