About

Abbas Abdolmaleki is a prominent researcher at the intersection of reinforcement learning (RL) and robotics, with a particular focus on making RL algorithms practical, robust, and deployable in real-world settings. His most influential contributions center on offline reinforcement learning, where his "Keep Doing What Worked" framework — cited over 100 times across its publications — introduced behavioral modeling priors to constrain off-policy learning from fixed datasets, a critical advance for applying RL to physical systems like robot control. Abdolmaleki has also made significant strides in robust RL, developing frameworks that handle model misspecification and sim-to-real transfer challenges that frequently undermine real-world deployment. His work on constrained and multi-objective policy optimization addresses the practical need for safe, smooth control signals in continuous robotic tasks, moving beyond naive bang-bang solutions. From humanoid soccer locomotion to complex robotic stacking of diverse geometric shapes, his research portfolio spans both foundational algorithmic development and compelling applied demonstrations. With a growing citation record across multiple high-impact publications and contributions spanning hybrid discrete-continuous action spaces, Abdolmaleki has established himself as a versatile and rigorous contributor to the advancement of intelligent, deployable robotic systems.

Research Focus

Key Achievements

13
H-Index
35
Papers
456
Total Citations
13
Avg Citations/Paper
🏆 Most Cited Paper
Keep Doing What Worked: Behavior Modelling Priors for Offline Reinforcement Learning
55 citations · 2020
📈 Most Prolific Year: 2019 (9 Papers)
🤝 Key Collaborators: 92
🏛 Institutions: Google (United States), University of Aveiro, Universidade do Porto, Google (United Kingdom), Google DeepMind (United Kingdom), University of Minho

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago