Foundation policies with hilbert representations

S Park, T Kreiman, S Levine - ar** an agent capable of adapting to unseen environments remains a difficult
challenge in imitation learning. This work presents Adaptive Return-conditioned Policy …

The power of resets in online reinforcement learning

Z Mhammedi, DJ Foster, A Rakhlin - arxiv preprint arxiv:2404.15417, 2024 - arxiv.org
Simulators are a pervasive tool in reinforcement learning, but most existing algorithms
cannot efficiently exploit simulator access--particularly in high-dimensional domains that …

Reinforcement Learning under Latent Dynamics: Toward Statistical and Algorithmic Modularity

P Amortila, DJ Foster, N Jiang… - Advances in …, 2025 - proceedings.neurips.cc
Real-world applications of reinforcement learning often involve environments where agents
operate on complex, high-dimensional observations, but the underlying (``latent'') dynamics …

Learning latent dynamic robust representations for world models

R Sun, H Zang, X Li, R Islam - arxiv preprint arxiv:2405.06263, 2024 - arxiv.org
Visual Model-Based Reinforcement Learning (MBRL) promises to encapsulate agent's
knowledge about the underlying dynamics of the environment, enabling learning a world …

Leveraging Separated World Model for Exploration in Visually Distracted Environments

K Huang, S Wan, M Shao, HH Sun… - Advances in …, 2025 - proceedings.neurips.cc
Abstract Model-based unsupervised reinforcement learning (URL) has gained prominence
for reducing environment interactions and learning general skills using intrinsic rewards …

Masked and Inverse Dynamics Modeling for Data-Efficient Reinforcement Learning

YJ Lee, J Kim, YJ Park, M Kwak… - IEEE Transactions on …, 2024 - ieeexplore.ieee.org
In pixel-based deep reinforcement learning (DRL), learning representations of states that
change because of an agent's action or interaction with the environment poses a critical …

Policy-shaped prediction: avoiding distractions in model-based reinforcement learning

M Hutson, I Kauvar, N Haber - Advances in Neural …, 2025 - proceedings.neurips.cc
Abstract Model-based reinforcement learning (MBRL) is a promising route to sample-
efficient policy optimization. However, a known vulnerability of reconstruction-based MBRL …

Learning Abstract World Model for Value-preserving Planning with Options

R Rodriguez-Sanchez, G Konidaris - arxiv preprint arxiv:2406.15850, 2024 - arxiv.org
General-purpose agents require fine-grained controls and rich sensory inputs to perform a
wide range of tasks. However, this complexity often leads to intractable decision-making …

Mastering robot manipulation with multimodal prompts through pretraining and multi-task fine-tuning

J Li, Q Gao, M Johnston, X Gao, X He… - arxiv preprint arxiv …, 2023 - arxiv.org
Prompt-based learning has been demonstrated as a compelling paradigm contributing to
large language models' tremendous success (LLMs). Inspired by their success in language …