paper-with-me

홈 › Papers

Behavior-Consistent Deep Reinforcement Learning

2026-05-20 · Marcel Hussing, Liv G. d'Aliberti, Claas Voelcker, Benjamin Eysenbach, Eric Eaton arxiv

Reinforcement learning (RL) often exhibits high variance across training runs, leading to unreliable performance and posing a major challenge to deployment in real-world domains. In this work, we address the challenge of cross-run policy divergence by formalizing the problem of behavior-consistent RL, where the objective is to obtain policies that are both high-performing and distributionally similar across training runs. Our key observation is that maximum-entropy RL provides a direct mechanism for controlling behavioral divergence by anchoring runs to a common (uniform) prior. We prove that, for Boltzmann policies, choosing the temperature proportional to $Q$-function disagreement bounds the pairwise KL divergence between the induced policies. However, we also show that naïvely increasing entropy might impair policy optimization while amplifying off-policy error. Building upon these observations, we propose $Q$-value Expectile Disagreement (QED), a state-dependent temperature schedule that uses double-critic disagreement as a single-run proxy for cross-run disagreement. Empirically, we demonstrate that across 18 continuous-control tasks, QED reduces across-run divergence by two orders of magnitude without sacrificing performance, resulting in a considerable reduction in return variance at modest sample-efficiency costs.

📄 PDF Abstract BibTeX arXiv:2605.21214

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

Behavior-Constrained Reinforcement Learning with Receding-Horizon Credit Assignment for High-Performance Control

2026-04-03 · Siwei Ju, Jan Tauberschmidt, Oleg Arenz, Peter van Vliet 외 arxiv

Learning high-performance control policies that remain consistent with expert behavior is a fundamental challenge in robotics. Reinforcement learning can discover high-performing strategies but often departs from desirab…

Reinforcement Learning

DCIR: Dynamic Consistency Intrinsic Reward for Multi-Agent Reinforcement Learning

2023-12-10 · Kunyang Lin, Yufeng Wang, Peihao Chen, Runhao Zeng 외

Learning optimal behavior policy for each agent in multi-agent systems is an essential yet difficult problem. Despite fruitful progress in multi-agent reinforcement learning, the challenge of addressing the dynamics of w…

Multi-agent Reinforcement Learningreinforcement-learningReinforcement LearningStarcraft+1

NADPEx: An on-policy temporally consistent exploration method for deep reinforcement learning

2018-12-21 · ICLR 2019 5 · Sirui Xie, Junning Huang, Lanxin Lei, Chunxiao Liu 외

Reinforcement learning agents need exploratory behaviors to escape from local optima. These behaviors may include both immediate dithering perturbation and temporally consistent exploration. To achieve these, a stochasti…

continuous-controlContinuous ControlDeep Reinforcement LearningMuJoCo+3

BehaviorGuard: Online Backdoor Defense for Deep Reinforcement Learning

2026-05-07 · Yinbo Yu, Xueyu Yin, Jiadai Wang, Chunwei Tian 외 arxiv

Backdoor attacks pose a serious threat to deep reinforcement learning (DRL). Current defenses typically rely on reward anomalies to reverse-engineer triggers and model finetuning to remove backdoors. However, complex tri…

Reinforcement Learning

The LoCA Regret: A Consistent Metric to Evaluate Model-Based Behavior in Reinforcement Learning

2020-07-07 · NeurIPS 2020 12 · Harm van Seijen, Hadi Nekoei, Evan Racah, Sarath Chandar

Deep model-based Reinforcement Learning (RL) has the potential to substantially improve the sample-efficiency of deep RL. While various challenges have long held it back, a number of papers have recently come out reporti…

General Reinforcement LearningModel-based Reinforcement LearningReinforcement Learning (RL)Representation Learning