arXiv:2606. 14029v1 Announce Type: new Abstract: Constrained MDPs (CMDPs) are a widely adopted framework for incorporating safety into RL agents; however, the framework does not support risk-sensitive constraints.
arXiv:2601. 22993v4 Announce Type: replace Abstract: We introduce Canary, a risk-averse method designed to optimize Value-at-Risk (VaR) constrained reinforcement learning (RL) problems.
arXiv:2509. 25582v4 Announce Type: replace Abstract: In-context reinforcement learning (ICRL) is an emerging RL paradigm where an agent, after pretraining, can adapt to out-of-distribution test tasks without any parameter updates, instead relying on an expanding context of interaction history.
arXiv:2606. 14415v1 Announce Type: new Abstract: Safe reinforcement learning (Safe RL) aims to maximize expected return while satisfying safety constraints, typically modeled as Constrained Markov Decision Processes (CMDPs).
arXiv:2607. 15457v1 Announce Type: new Abstract: We study robust peak-cost constrained reinforcement learning (RP-CRL), where the objective is to maximize expected reward while controlling the maximum cost encountered along a trajectory.
The paper introduces Canary, a risk‑averse reinforcement learning method that optimizes Value‑at‑Risk (VaR) constraints. By applying Cantelli’s inequality, Canary derives a tractable, conservative, and smooth bound on the VaR constraint using only the first two moments of the cost return, yielding a stable constraint estimator even with tight violation thresholds. Extending the trust‑region framework of Constrained Policy Optimization (CPO), the authors provide worst‑case bounds for policy improvement and constraint violation, and empirically demonstrate that Canary reliably satisfies the VaR constraint in every tested environment.
The paper introduces new evaluation metrics for safe reinforcement learning that go beyond average safety guarantees by examining how often and how severely safety bounds are violated, consistency across tasks and bounds, and the relationship between training-time and final policy behavior. It also proposes a safety tier system for categorizing algorithms and presents empirical safety evaluations on multiple navigation tasks. The authors recommend reporting aggregate metrics, distributional data, and task‑specific results together, and provide an open‑source suite, SafeRLEval, to facilitate reliable safety assessment.
The paper introduces a chance-constrained approach to fine‑tune large language models (LLMs) that limits the proportion of safety examples whose performance degrades beyond a set threshold relative to a reference model. By replacing the discontinuous violation indicator with a differentiable majorization, the authors derive a tractable, conservative constraint and a closed‑form, constraint‑aware gradient update that focuses on examples near or above the degradation threshold. Experiments on harmful fine‑tuning across three tasks and models show that this tail‑aware method consistently outperforms existing safety‑preserving baselines, suggesting that safety preservation should be treated as a reliability‑constrained optimization problem rather than average‑risk regularization.
arXiv:2506. 02255v2 Announce Type: replace Abstract: Most existing safe reinforcement learning (RL) benchmarks focus on robotics and control tasks, offering limited relevance to high-stakes domains that involve structured constraints, mixed-integer decisions, and industrial complexity.
arXiv:2607. 12784v1 Announce Type: cross Abstract: Reinforcement Learning has revolutionized the landscape of robotic research, allowing robust learning of complex robotic skills in simulation.
arXiv:2609.34426v2 Announce Type: replace Abstract: This paper addresses the problem of safe offline reinforcement learning, which involves training a policy to satisfy safety constraints using an of...
Reinforcement Learning has revolutionized the landscape of robotic research, allowing robust learning of complex robotic skills in simulation. However, real-world deployment in open-ended environments requires strong safety guarantees to prevent dangerous or harmful behaviors.
arXiv:2608. 16068v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed as agents that rely on system prompts to use tools and complete tasks.