I am an incoming PhD student at the University of Alberta in reinforcement learning, supervised by Prof. Dale Schuurmans. Before this, I was a reinforcement learning researcher at Kindred AI. I completed my MSc in Applied Computing and HBSc in Computer Science at the University of Toronto.
My life-long research goal is to create autonomous agents that can adapt continually in real life. In particular, I am interested in developing an agent that can 1) understand the properties of the world from 2) minimal interactions with environment and from 3) observing other agents’ behaviours. The more ambitious agent would be 4) curious about the world and explores to answer its own questions and have the ability to 5) build a repertoire of skills such that they can compose the skills to complete long horizon tasks. To achieve my goal, my research focuses on reinforcement learning (RL), imitation learning (IL), robotics, meta-learning, and continual learning.
As a first step, I am developing this repository to implement (and reproduce) recent RL research papers. My goal with this work is to identify the problems with existing algorithms. Specifically, I am looking at these algorithms from the perspective of sample efficiency and exploration. I also investigate what “dirty tricks” we need in order to reproduce these papers. So far, this repository contains some model-free RL algorithms, IL algorithms, and hierarchical RL algorithms. I plan to add some model-based RL algorithms and offline RL algorithms.
I am interested in collaborating on research projects so feel free to contact me via firstname.lastname@example.org.
- (2022/06/17) I will be joining RLAI at the University of Alberta as a PhD student in reinforcement learning under Prof. Dale Schuurmans!