Multitask Soft Option Learning

Conference paper (2020)

Authors

Maximilian Igl University of Oxford
Andrew Gambardella University of Oxford
J. He Interactive Intelligence - EEMCS
Nantas Nardelli University of Oxford
N Siddharth University of Oxford
J.W. Böhmer University of Oxford
Shimon Whiteson University of Oxford
Research Group
Interactive Intelligence (EEMCS) (TU Delft)
Copyright: © 2020 Maximilian Igl, Andrew Gambardella, J. He, Nantas Nardelli, N Siddharth, J.W. Böhmer, Shimon Whiteson
More Info
expand_more

Published Date

2020

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Electrical Engineering, Mathematics and Computer Science

Department

Intelligent Systems

Research Group

Interactive Intelligence

Abstract

We present Multitask Soft Option Learning (MSOL), a hierarchical multitask framework based on Planning as Inference. MSOL extends the concept of options, using separate variational posteriors for each task, regularized by a shared prior. This “soft” version of options avoids several instabilities during training in a multitask setting, and provides a natural way to learn both intra-option policies and their terminations. Furthermore, it allows fine-tuning of options for new tasks without forgetting their learned policies, leading to faster training without reducing the expressiveness of the hierarchical policy. We demonstrate empirically that MSOL significantly outperforms both hierarchical and flat transfer-learning baselines.

Files

Igl20a_1.pdf
(.pdf | 1.32 Mb)