I am a PhD student at the University of Toronto, advised by Jimmy Ba and Geoffrey Hinton. I am also a student researcher at Google Brain and the Vector Institute. I finished my MRes in CSML at University College London, advised by Tim Lillicrap and Karl Friston.

My goal is to build intelligent machines based on concepts of the human brain, and evaluate them in complex simulations. Currently, I’m focusing on agents that learn without rewards, reinforcement learning with world models, and unsupervised learning.

If you have any questions, feel free to contact me at [email protected].

News

2019-03-01 Accepting student researchers at the Vector Institute.

2018-11-28 We organize the ICLR 2019 workshop on task-agnostic RL.

2018-02-16 Guest lecture at Stanford CS 20 on Variational Inference in TF.

Selected research

Please see my Scholar profile for a full list including article links.

• 
  Dec 10, 2018  (NeurIPS 2019, 21%)

  Bayesian Layers: A Module for Neural Network Uncertainty

  D Tran, M Dusenberry, M v d Wilk, D Hafner
• 
  Nov 30, 2018  (NeurIPS 2018 DRL Workshop)

  Modulated Policy Hierarchies

  A Pashevich, D Hafner, J Davidson, R Sukthankar, C Schmid
• 
  Jul 24, 2018  (UAI 2019, 26%)

  Noise Contrastive Priors for Functional Uncertainty

  D Hafner, D Tran, A Irpan, T Lillicrap, J Davidson
• 
  Jul 14, 2018  (NeurIPS 2018, oral 0.6%)

  Sample-Efficient Reinforcement Learning with Stochastic Ensemble Value Expansion

  J Buckman, D Hafner, G Tucker, E Brevdo, H Lee
• 
  Jul 4, 2018  (ICML 2019, oral 23%)

  Learning Latent Dynamics for Planning from Pixels

  D Hafner, T Lillicrap, I Fischer, R Villegas, D Ha, H Lee, J Davidson
• 
  Apr 27, 2018  (RSS 2018, oral 31%)

  Sim-to-Real: Learning Agile Locomotion For Quadruped Robots

  J Tan, T Zhang, E Coumans, A Iscen, Y Bai, D Hafner, S Bohez, V Vanhoucke
• 
  Nov 8, 2017  (Tech report)

  TensorFlow Agents: Efficient Batched Reinforcement Learning in TensorFlow

  D Hafner, J Davidson, V Vanhoucke
• 
  Jul 18, 2017  (NIPS 2017, 21%)

  Learning Hierarchical Information Flow with Recurrent Neural Modules

  D Hafner, A Irpan, J Davidson, N Heess
• 
  Aug 22, 2016  (ESA 2016, 27%)

  Probabilistic Routing for On-Street Parking Search

  T Arndt, D Hafner, T Kellermeier, S Krogmann, A Razmjou, M Krejca, R Rothenberger, T Friedrich
• 
  Jul 15, 2016  (BSc thesis)

  Deep Reinforcement Learning From Raw Pixels in Doom

  D Hafner

Short biography

Danijar Hafner is a PhD student at the University of Toronto advised by Jimmy Ba and Geoffrey Hinton. He is also a student researcher at Google Brain and the Vector Institute. His research focuses on artificial intelligence, specifically reinforcement learning with world models and intrinsic objectives for agents that learn without rewards. Danijar obtained his MRes at the University College London and the Gatsby Unit under the supervision of Tim Lillicrap and Karl Friston. Danijar co-authored the book “TensorFlow for Machine Intelligence” and advises Stanford’s course “TensorFlow for Deep Learning Research”. He completed his bachelor’s thesis on deep reinforcement learning for the video game Doom at the Hasso Plattner Institute, Germany.

Links

Twitter · Github · Quora · LinkedIn · Google AI · ORCID