[2603.02935] Contextual Latent World Models for Offline Meta Reinforcement Learning

Computer Science > Machine Learning arXiv:2603.02935 (cs) [Submitted on 3 Mar 2026] Title:Contextual Latent World Models for Offline Meta Reinforcement Learning Authors:Mohammadreza Nakheai, Aidan Scannell, Kevin Luck, Joni Pajarinen View a PDF of the paper titled Contextual Latent World Models for Offline Meta Reinforcement Learning, by Mohammadreza Nakheai and 3 other authors View PDF HTML (experimental) Abstract:Offline meta-reinforcement learning seeks to learn policies that generalize across related tasks from fixed datasets. Context-based methods infer a task representation from transition histories, but learning effective task representations without supervision remains a challenge. In parallel, latent world models have demonstrated strong self-supervised representation learning through temporal consistency. We introduce contextual latent world models, which condition latent world models on inferred task representations and train them jointly with the context encoder. This enforces task-conditioned temporal consistency, yielding task representations that capture task-dependent dynamics rather than merely discriminating between tasks. Our method learns more expressive task representations and significantly improves generalization to unseen tasks across MuJoCo, Contextual-DeepMind Control, and Meta-World benchmarks. Subjects: Machine Learning (cs.LG) Cite as: arXiv:2603.02935 [cs.LG]   (or arXiv:2603.02935v1 [cs.LG] for this version)   https://doi.org/10.48550/arXiv.2...