Papers

47

Total Citations

2,529

H-Index

21

About

Nicolas Heess is a prominent researcher at the intersection of deep reinforcement learning, robotics, and motor control, whose work has fundamentally advanced how autonomous agents acquire complex physical skills. Based primarily at DeepMind, Heess has made seminal contributions to sparse-reward reinforcement learning, most notably through frameworks that integrate demonstration data with policy gradient methods — work that has collectively garnered hundreds of citations and shaped modern robot learning pipelines. His landmark contributions include leveraging demonstrations for sparse-reward robotics (510 citations), the Scheduled Auxiliary Control paradigm for learning from scratch (155 citations), and the widely adopted dm_control simulation suite (186 citations), which has become a standard benchmark platform for the continuous control community. Heess has also pioneered sim-to-real transfer techniques and skill embedding approaches that enable robots to generalize across diverse manipulation and locomotion tasks. His more recent work training bipedal humanoid robots to play soccer using deep RL (147 citations) exemplifies his ambition to bridge fundamental research and physically embodied intelligence. Spanning visuomotor control, imitation learning, and whole-body animation, his research continues to define the frontier of what autonomous physical agents can achieve.

Research Focus

Key Achievements

21
H-Index
47
Papers
2,529
Total Citations
54
Avg Citations/Paper
🏆 Most Cited Paper
Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards
510 citations · 2017
📈 Most Prolific Year: 2020 (12 Papers)
🤝 Key Collaborators: 261
🏛 Institutions: Google DeepMind (United Kingdom), Google (United States), University College London

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
    Catch & Carry
    98 citations · 2020

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago