Learning Policy with Compositional Generalizability using Self-Supervised Object Proposals [NeurIPS 2020]

NeurIPS 2020

Learning Policy with Compositional Generalizability using Self-Supervised Object Proposals [NeurIPS 2020]

Nov 08, 2020
|
47 views
|
Details
Paper Abstract: We study how to learn a policy with compositional generalizability. We propose a two-stage framework, which refactorizes a high-reward teacher policy into a generalizable student policy with strong inductive bias. Particularly, we implement an object-centric GNN-based student policy, whose input objects are learned from images through self-supervised learning. Empirically, we evaluate our approach on four difficult tasks that require compositional generalizability, and achieve superior performance compared to baselines. Authors: Tongzhou Mu, Jiayuan Gu, Zhiwei Jia, Hao Tang, Hao Su

Comments
loading...