RL Stability

RL: Reinforcement Learning

Latest papers

42 papers9 papers in the last four weeks, up 200% on the four weeks before. 0.1% of all new papers.

Learning When to Act: Communication-Efficient Reinforcement Learning via Run-Time Assurance
May 11, 2026Adam Haroon, Erick J. Rodríguez-Seda, Cody Fleming +1RL Control Lyapunov-Based Control
Department of Mechanical Engineering, Iowa State University, Ames, IA, USA · Navy Center for Applied Research in Artificial Intelligence (NCARAI), U.S. Naval Research Laboratory, Washington, D.C., USA · Department of Weapons, Robotics, and Control Engineering, United States Naval Academy, Annapolis, MD, USA
T $^2$ PO: Uncertainty-Guided Exploration Control for Stable Multi-Turn Agentic Reinforcement Learning
May 4, 2026Haixin Wang, Hejie Cui, Chenwei Zhang +7RL for LLM Agents Reinforcement Learning
University of California, Los Angeles · 2Amazon.com Inc.
Stability of Control Lyapunov Function Guided Reinforcement Learning
May 3, 2026Zachary Olkin, William D. Compton, Aaron D. AmesRL Control Lyapunov Stability
The authors are with the Department of Control and Dynamical Systems at the California Institute of Technology.
AdamO: A Collapse-Suppressed Optimizer for Offline RL
May 3, 2026Nan Qiao, Sheng Yue, Shuning Wang +1Reinforcement Learning Orthogonal Optimization
1Tsinghua University, China · 2Central South University, China · 3Sun Yat-sen University, China
GIFT: Global stabilisation via Intrinsic Fine Tuning
Apr 25, 2026Rory Young, Nicolas PugeaultRL Control Reinforcement Learning
School of Computing Science University of Glasgow
Target-Aligned Reinforcement Learning
Mar 31, 2026Leonard S. Pleiss, James Harrison, Maximilian SchifferReinforcement Learning RL Control
Google DeepMind
Enhancing Reinforcement Learning Fine-Tuning with an Online Refiner
Mar 18, 2026Hao Ma, Zhiqiang Pu, Yang Liu +1RL Fine-Tuning RL for Language Models
School of Artificial Intelligence, University of Chinese Academy of Sciences · Institute of Automation, Chinese Academy of Sciences
GAC: Stabilizing Asynchronous RL Training for LLMs via Gradient Alignment Control
Mar 2, 2026Haofeng Xu, Junwei Su, Yukun Tian +3RL Stability Asynchronous Advantage Actor-Critic
School of Computing and Data Science, The University of Hong Kong · Alibaba Group · Southeast University
ARLArena: A Unified Framework for Stable Agentic Reinforcement Learning
Feb 25, 2026Xiaoxuan Wang, Han Zhang, Haixin Wang +11Agentic RL Reinforcement Learning
Department of Computer Science, University of California, Los Angeles, California, USA · University of Wisconsin–Madison, Madison, Wisconsin, USA
Stable Deep Reinforcement Learning via Isotropic Gaussian Representations
Feb 22, 2026Ali Saheb Pasand, Johan Obando-Ceron, Aaron Courville +2RL Stability Representation Learning for RL
Mila – Quebec AI Institute · McGill · Université de Montréal +1
QUATRO: Query-Adaptive Trust Region Policy Optimization for LLM Fine-tuning
Feb 4, 2026Doyeon Lee, Eunyi Lyou, Hyunsoo Cho +3Trust-Region Optimization Policy Learning
Seoul National University · Department of Artificial Intelligence / Institute for Multiscale Matter and Systems, Ewha Woman University · Geogia Institute of Technology
BNPO: Beta Normalization Policy Optimization
Jun 3, 2025Changyi Xiao, Mengdi Zhang, Yixin CaoRL for Language Model Reasoning Policy Gradient Methods
School of Computer Science, Fudan University · Meituan Group

May 11, 2026J/K move · Enter open · S save

Learning When to Act: Communication-Efficient Reinforcement Learning via Run-Time Assurance

Adam Haroon, Erick J. Rodríguez-Seda, Cody Fleming, Tristan Schuler

Department of Mechanical Engineering, Iowa State University, Ames, IA, USA · Navy Center for Applied Research in Artificial Intelligence (NCARAI), U.S. Naval Research Laboratory, Washington, D.C., USA · Department of Weapons, Robotics, and Control Engineering, United States Naval Academy, Annapolis, MD, USA

RL Control Lyapunov-Based Control Safe RL Policy Learning Robot Safety Feedback Control+3

Open paper AskPDF arXiv

Safe reinforcement learning (RL) typically asks

\textit{what}

an agent should do. We ask

\textit{when}

it needs to act, and show that a single policy can jointly learn control inputs and communication-efficient timing decisions under a pointwise Lyapunov safety shield. We focus on stabilization around a known equilibrium, where CARE-based LQR backups, Lyapunov certificates, and classical Lyapunov-STC are well defined, enabling clean comparison against analytical baselines. A run-time assurance (RTA) layer overrides the policy via a one-step-ahead Lyapunov prediction and a precomputed LQR backup, providing a strictly stronger guarantee than constrained MDP methods that enforce safety only in expectation. On an inverted pendulum, cart--pole, and planar quadrotor, the learned policy achieves

1.91\times

1.45\times

, and

3.51\times

higher mean inter-sample interval (MSI) than a Lyapunov-triggered baseline; a fixed LQR controller at the same average rate is unstable on all three plants, showing that adaptive timing, not a lower average rate, makes sparsity safe. A CARE-derived Lyapunov reward transfers across environments without redesign, with a single weight

w_c

controlling the stability--communication tradeoff; ablations confirm the RTA shield is essential, with its removal reducing MSI by

1.27

1.84\times

and degrading state norms. A preference-conditioned extension recovers the full tradeoff frontier from one model at

\tfrac{2}{11}

of training compute, and SAC experiments show the results are algorithm-agnostic across discrete and continuous domains. A 12-state 3D quadrotor case study extends the framework to higher-dimensional systems where classical STC is intractable, and robustness to

\pm30\%

mass variation and disturbances shows graceful degradation, with the RTA absorbing what the learned policy cannot.