aiqbench analysis
This paper introduces ATLAS, a novel reinforcement learning framework designed to overcome the trade-off between sample efficiency and robustness in non-stationary environments. By decoupling transition dynamics from reward signals using successor features, it achieves near-instantaneous adaptation while mitigating catastrophic forgetting—a critical advancement for continual learning systems facing environmental shifts.
ATLAS: Adaptive Topological Learning with Abstract Successors for Continual Learning
arXiv:2608.04334v1 Announce Type: new Abstract: Contemporary model-free reinforcement learning algorithms can achieve very high performance, but have low sample efficiency and are not robust to changes in the environment. Model-based algorithms have much higher sample efficiency, but still fail when the environment shifts. This paper introduces Adaptive Topological Learning with Abstract Successors (ATLAS) to combat these challenges. ATLAS uses a Grow When Required network with Successor Features in order to achieve high sample efficiency while also robustly tackling catastrophic forgetting. We evaluate ATLAS in spatial navigation tasks, benchmarking its performance against common on-policy and off-policy algorithms. Our empirical results demonstrate that by structurally decoupling transition dynamics from the reward signal, ATLAS achieves near-instantaneous adaptation to new goals and can exhibit positive backward transfer, significantly outperforming baseline methods in non-stationary environments.
Abstract:Contemporary model-free reinforcement learning algorithms can achieve very high performance, but have low sample efficiency and are not robust to changes in the environment. Model-based algorithms have much higher sample efficiency, but still fail when the environment shifts. This paper introduces Adaptive Topological Learning with Abstract Successors (ATLAS) to combat these challenges. ATLAS uses a Grow When Required network with Successor Features in order to achieve high sample efficiency while also robustly tackling catastrophic forgetting. We evaluate ATLAS in spatial navigation tasks, benchmarking its performance against common on-policy and off-policy algorithms. Our empirical results demonstrate that by structurally decoupling transition dynamics from the reward signal, ATLAS achieves near-instantaneous adaptation to new goals and can exhibit positive backward transfer, significantly outperforming baseline methods in non-stationary environments.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2608.04334 [cs.LG]
(or arXiv:2608.04334v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2608.04334 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 222 words
Related coverage
A2TTA: Anchored-and-Agile Test-Time Adaptation for Evolving Traffic Sensor Networks
arXiv · Jul 29, 2026
Online Versatile Incremental Learning: Towards Class and Domain-Agnostic Adaptation at Any Time
arXiv · Sep 30, 2026
Building Blocks for Foundation Model Training and Inference on AWS
Hugging Face · May 11, 2026
ATLAS-AL: Adaptive Trust-Region for Latent Adversarial Searches via Active Learning
arXiv · Oct 7, 2026