arXiv AI

Reinforcing the World's Edge: A Continual Learning Problem in the Multi-Agent-World Boundary

The paper studies a stationary decentralized Markov game where a focal agent experiences drifting rewards and dynamics due to learning peers, framing this as an agent‑centric continual reinforcement‑learning problem. It introduces the concept of an invariant core—maximal abstract patterns common to many successful trajectories—and proves a worst‑case conditioning theorem linking trajectory‑law drift to success coverage. The authors provide theoretical guarantees for survival horizon, first‑exit law, and regret, and validate their predictions with solvable models and empirical studies in continual control, cue‑MNIST, and Level‑Based Foraging.

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
arXiv AI
Aug 18

UC-PSRO: Utility-Conditioned Policy-Space Response Oracles with a Communication-Dropout Curriculum for Game-Theoretic Course-of-Action Generation in Adversarial Swarms

arXiv:2608. 15372v1 Announce Type: new Abstract: We study generating game-theoretically optimized Courses of Action (COAs) for a Blue UAS swarm against an adaptive Red adversary in a communication-degraded environment, motivated by (but not derived from) a public U.

By Phillip Jiang
arXiv Machine Learning
Jul 27

Embodiment-Induced Coordination Regimes in Tabular Multi-Agent Q-Learning

arXiv:2601. 17454v2 Announce Type: replace-cross Abstract: Centralized value learning underlies a broad class of multi-agent reinforcement learning methods, but its claimed advantage is typically evaluated in settings that confound coordination structure with function approximation and partial observability.

By Muhammad Ahmed Atif, Nehal Naeem Haji, Mohammad Shahid Shaikh, Muhammad Ebad Atif
arXiv AI
4d ago

Explore More, Drift Less: Outcome-Only Reinforcement Learning Can Suffice for Long-Horizon Interactive Agents

The paper proposes CANOPY, a minimalist reinforcement learning protocol that addresses two common pitfalls—signal starvation and policy drift—in outcome‑only RL for long‑horizon interactive tasks. By scaling same‑task exploration, keeping updates on‑policy, and anchoring updates with KL divergence, CANOPY enables a Qwen3‑14B agent to achieve top leaderboard results on the AppWorld coding benchmark without auxiliary supervision or elaborate scaffolding. The approach also improves performance on SWE‑bench for a Qwen3.5‑9B model.

By Liming Pu, Xiaoxia Li, Yifu Liu, Teng Cao, Bin Yang
arXiv AI
Jun 16

Direction-Conditioned Policies via Compositional Subgoal Scoring for Online Goal-Conditioned Reinforcement Learning

arXiv:2606. 16515v1 Announce Type: cross Abstract: Hamilton-Jacobi-Bellman theory implies that the optimal goal-conditioned action depends on the goal only through the gradient of the goal-reaching distance at the current state, yet standard online GCRL still conditions the actor on the raw goal -- a signal that is geometrically uninformative when the goal is far from the data distribution.

By Swaminathan S K, Damiya Gondha, Theyanesh Eswaramoorthy Rajahkrishnan, Aritra Hazra
arXiv Machine Learning
Aug 28

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.

By Zhenya Liu, Yang Meng, Zhuokai Zhao, Xuefeng Liu, Yuxin Chen
arXiv Machine Learning
Aug 17

What preferences can - and cannot - predict in multi-agent online learning

arXiv:2608. 13810v1 Announce Type: cross Abstract: We examine the interplay between ordinal, preference-based solution concepts in games and the long-run behavior of game dynamics, asking in particular to what extent the combinatorial data of a game -- its preference graph -- determine the outcomes of no-regret learning dynamics -- such as follow-the-regularized-leader (FTRL).

By Omar Abbadi, Rida Laraki, Panayotis Mertikopoulos
arXiv AI
Jun 30

Pessimism's Paradox: Conservative Offline Training Amplifies Reward Hacking During Online Adaptation in Reasoning Models

arXiv:2606. 30627v1 Announce Type: cross Abstract: Conservative offline training is widely advocated as a safe foundation for subsequent online adaptation: if a policy stays close to well-supported behaviour, the argument goes, it is less likely to exploit imperfections in a learned reward model.

By Subramanyam Sahoo, Aman Chadha, Vinija Jain, Divya Chaudhary