English

TD-MPC-Opt: Distilling Model-Based Multi-Task Reinforcement Learning Agents

Machine Learning 2025-07-03 v1 Robotics

Abstract

We present a novel approach to knowledge transfer in model-based reinforcement learning, addressing the critical challenge of deploying large world models in resource-constrained environments. Our method efficiently distills a high-capacity multi-task agent (317M parameters) into a compact model (1M parameters) on the MT30 benchmark, significantly improving performance across diverse tasks. Our distilled model achieves a state-of-the-art normalized score of 28.45, surpassing the original 1M parameter model score of 18.93. This improvement demonstrates the ability of our distillation technique to capture and consolidate complex multi-task knowledge. We further optimize the distilled model through FP16 post-training quantization, reducing its size by \sim50\%. Our approach addresses practical deployment limitations and offers insights into knowledge representation in large world models, paving the way for more efficient and accessible multi-task reinforcement learning systems in robotics and other resource-constrained applications. Code available at https://github.com/dmytro-kuzmenko/td-mpc-opt.

Keywords

Cite

@article{arxiv.2507.01823,
  title  = {TD-MPC-Opt: Distilling Model-Based Multi-Task Reinforcement Learning Agents},
  author = {Dmytro Kuzmenko and Nadiya Shvai},
  journal= {arXiv preprint arXiv:2507.01823},
  year   = {2025}
}

Comments

Preprint of a manuscript submitted for peer review

R2 v1 2026-07-01T03:43:27.017Z