English

Multi-task Reinforcement Learning with a Planning Quasi-Metric

Machine Learning 2020-12-08 v3 Machine Learning

Abstract

We introduce a new reinforcement learning approach combining a planning quasi-metric (PQM) that estimates the number of steps required to go from any state to another, with task-specific "aimers" that compute a target state to reach a given goal. This decomposition allows the sharing across tasks of a task-agnostic model of the quasi-metric that captures the environment's dynamics and can be learned in a dense and unsupervised manner. We achieve multiple-fold training speed-up compared to recently published methods on the standard bit-flip problem and in the MuJoCo robotic arm simulator.

Keywords

Cite

@article{arxiv.2002.03240,
  title  = {Multi-task Reinforcement Learning with a Planning Quasi-Metric},
  author = {Vincent Micheli and Karthigan Sinnathamby and François Fleuret},
  journal= {arXiv preprint arXiv:2002.03240},
  year   = {2020}
}

Comments

Deep RL Workshop, NeurIPS 2020

R2 v1 2026-06-23T13:35:24.477Z