Glen Berseth

Glen Berseth

I am a Postdoctoral Researcher at the Berkeley Artificial Intelligence Research (BAIR) working in the Robotic AI & Learning (RAIL) lab with Sergey Levine. My research combines deep learning and reinforcement learning on high-dimensional control problems.

I received my PhD at the Department of Computer Science at the University of British Columbia in 2019 where I worked on reinforcement learning, machine learning and motion planning with Michiel van de Panne. I received my BSc degree in Computer Science from York University in 2012 and my MSc from York University under the supervision of Petros Faloutsos in 2014 for optimization and authoring crowd simulations. I have published in a wide range of areas including computer animation, machine learning and robotics and am a NSERC scholarship award winner. You can find a list of projects and publications here.

While learning agents hold the promise of being general problem-solvers, they often struggle with environmental and task diversity. My work aims to move away from training agents to maximize reward on narrow tasks and instead looks into developing agents that learn general-purpose skills that contribute to solving a wide variety of tasks. Specifically, my work has focused on combining deep learning and reinforcement learning methods to solve high-dimensional perception and planning problems.

News

  • Apr 2021: Our research paper that will be presented at ICRA2021 on RL for bipedal robots was featured in MIT Technology Review
  • Mar 2021: Associate Editor for IROS 2021
  • Feb 2021: Two papers accepted to ICRA2021!
  • Jan 2021: SMiRL: Surprise Minimizing RL in Unstable Environments receives oral presentation at ICLR 2021 (top 1.8% of submitted papers)
  • Jan 2021: Two papers accepted to ICLR 2021
  • Jan 2021: Invited talk at IJCAI workshop Neuro-Cognitive Modeling of Humans and Environments
  • Aug 2020: Deep Integration of Physical Humanoid Control and Crowd Navigation receives best paper runner up at MIG2020
  • Sep 2019: Visual Imitation work featured in MIT CSAIL podcast
  • Feb 2019: Defended Ph.D. Thesis at the University of British Columbia under the supervision of Michiel van de Panne
  • Aug 2017: SIGGRAPH paper DeepLoco: Dynamic Locomotion Skills Using Hierarchical Deep Reinforcement Learning covered in the Popular Mechanics.
  • Mar 2017: Awarded NSERC PhD scholarship
  • May 2016: Modelling Dynamic Brachiation receives best poster award at Graphics Interfaces 2016
  • Mar 2016: Robust Space-Time Footsteps for Agent-Based Steering receives best short paper award at CASA 2016

Ravjot Singh Articles

Hierarchical Reinforcement Learning

Unsupervised Reinforcement Learning

Planning

Imitation

Life Long Learning

EnvironmentDesign