← Back to all news
arXiv Machine Learning

Shared Actors Need Not Share Critics: Effects of Value Mismatch in Parallel Reinforcement Learning

The paper investigates the problem of sharing a single critic across multiple parallel environments in reinforcement learning. It shows that when environments assign different expected returns to the same state, a shared critic must reconcile conflicting value targets, which can distort advantage estimates and misguide policy updates. The authors propose a simple fix—providing the critic with the environment index—demonstrating through bandit models and experiments on CartPole, MuJoCo, BipedalWalker, and 16 Procgen games that this conditional critic stabilizes learning and boosts returns, achieving a 40.8% improvement in aggregate normalized return on unseen levels.

Related stories

arXiv AI
Oct 9

Constrained Command-Conditioned Reinforcement Learning with Bandit Strategy Selection in Real-Time Strategy Games

The paper presents a constrained command‑conditioned Proximal Policy Optimization (PPO) executor for the MicroRTS real‑time strategy environment, enabling a single execution policy to follow diverse strategic commands. A Thompson‑sampling bandit strategist selects command tuples based on in‑game observations of the opponent’s strategy, rather than relying on opponent identity. In controlled experiments, the strategist‑executor system outperforms a flat PPO baseline against three of the four strongest opponents on a training map, including two held‑out opponents, while showing no significant difference against the remaining opponents.

By Nick Leenders, Roy Lindelauf, Joost van Oijen, Boris Cule
agentsreinforcement-learning
More like this →
arXiv Machine Learning
Jul 13

SafeExplorer: An Unbiased Policy Gradient for Reinforcement Learning with Recovery Interventions

arXiv:2607. 08925v1 Announce Type: new Abstract: Training reinforcement-learning agents directly on physical robots makes every fall costly, since a fall can damage the platform and cannot be undone like a simulator reset; the goal is therefore to minimize falls during training rather than trade them off against return, as constrained Markov decision process (MDP) formulations do.

By Elham Daneshmand, Majid Khadiv, Glen Berseth, Hsiu-Chin Lin
agentsreinforcement-learningroboticsbenchmarkssafety
More like this →
arXiv Machine Learning
Oct 8

An Informational Curse of Horizon in Goal-Conditioned Policy Learning

The paper identifies an informational curse of horizon in goal-conditioned policy learning, showing that longer goal relabeling horizons during training can hurt policy generalization and performance. Controlled experiments with oracle planners separate training horizons from test horizons, revealing that behavioral cloning policies degrade with longer horizons while reinforcement learning mitigates this effect. The authors attribute this to reduced conditional mutual information between actions and hindsight-relabeled goals, and demonstrate that distilling short-horizon policy Jacobians into long-horizon policies improves performance, especially in combinatorial manipulation tasks.

By John L. Zhou, Yuxuan Dong, Jonathan C. Kao
reinforcement-learningroboticssafety
More like this →
arXiv AI
Sep 1

BCPPO: Bachelier-Inspired Constrained Proximal Policy Optimization for Tail-Risk-Aware Safe Reinforcement Learning

BCPPO is a new variant of Proximal Policy Optimization that uses Bachelier-inspired cost‑prediction networks to generate a smooth penalty based on disagreement among critics. The method keeps temporal‑difference learning unchanged, applies a saturation‑aware controller to manage cost penalties, and deploys only the policy network. Across extensive experiments, BCPPO outperforms comparators in achieving higher mean returns while maintaining lower or comparable CVaR in all tested tasks.

By Dongsheng Hou, Yanqiao Chen, Yuhan Rui
reinforcement-learningsafety
More like this →
arXiv Machine Learning
Aug 13

When Offline Evaluation Misleads: A Diagnostic Protocol for Reward and Policy Selection in Delayed-Feedback Contextual Bandits

arXiv:2608. 11560v1 Announce Type: new Abstract: Personalizing marketing messages with contextual multi-armed bandits (CMABs) drives real business value, yet the objective that ultimately matters - a downstream conversion - is observed only weeks later, too late to drive online learning.

By Sang Su Lee, Vineeth Loganathan, Shishir Dash, Vijay Raghavan
reinforcement-learningbenchmarkssafety
More like this →
arXiv AI
Sep 24

Reinforcement Learning with Decomposed Subtasks

The paper introduces Reinforcement Learning with Decomposed Subtasks (RLDS), a method that splits trajectory rewards into per‑subtask shares before policy updates, replacing the scalar advantage used in Group Relative Policy Optimization (GRPO). RLDS employs Subtask‑Decomposed Advantage Estimation (SDAE) to compute group‑relative advantages and distribute credit to tokens based on subtask importance, focusing on steps where a reflection marks a subtask as consequential. Experiments on four benchmarks—FrozenLake, HotpotQA, ScienceWorld, and DeepResearch—show that RLDS improves performance on high‑heterogeneity tasks (ScienceWorld and FrozenLake) and is more compute‑efficient than scalar GRPO for long rollouts.

By Mattie Terzolo, Mikolaj Sacha, Ayan Sinha, Andrew Rabinovich
llmsagentsreinforcement-learningroboticsbenchmarks
More like this →