Preprint. Work in progress since 2021.

Toward General-Purpose Robots that Learn

Weiji Xie^1,2

¹APEX Lab, Shanghai Jiao Tong University ²TeleAI, China Telecom

Abstract

Keywords: robot learning, reinforcement learning, embodied AI, sim-to-real.

1Introduction

The author is an M.S. student in computer science at Shanghai Jiao Tong University (APEX Lab, advised by Prof. Weinan Zhang) and a research intern at TeleAI, China Telecom, working with Dr. Chenjia Bai. Prior work spans motion tracking, locomotion, sim-to-real transfer, generative models and multi-agent learning .

Stated as an objective, the research program is

(1)

where the hard part is that 𝒯 should eventually contain every task people care about. Section 3 reports progress so far.

2Recent Updates

3Work

4Education and Experience

Table 1: Path so far. Full details in the supplementary material.
Period	Position	Institution