Kenji Doya

About

Kenji Doya is a pioneering computational neuroscientist and robotics researcher whose work sits at the intersection of reinforcement learning, biological neural mechanisms, and adaptive robotic systems. Best known for bridging computational theory with biological understanding, Doya has made foundational contributions to explaining how animals and artificial agents learn goal-directed behaviors through reward-based feedback. His highly cited work on the computational theory and biological mechanisms of reinforcement learning — accumulating over 240 citations across related publications — has become essential reading for researchers seeking to understand how the brain implements learning algorithms analogous to those in machine learning. Beyond theoretical contributions, Doya has demonstrated remarkable range in applied research. His work on CPG-based biped locomotion, hierarchical reinforcement learning for robotic stand-up behaviors, and the ambitious Cyber Rodent project — which explored self-preservation and self-reproduction in autonomous robots — reflects a sustained commitment to grounding computational ideas in physically embodied systems. He has also advanced policy search methods, intrinsic motivation frameworks, and multi-agent learning architectures. Collectively, his publications have garnered hundreds of citations, establishing him as an influential voice shaping how both neuroscience and robotics communities think about adaptive, reward-driven learning systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Reinforcement learning: Computational theory and biological mechanisms

   124 citations · 2007
2. 2
   Reinforcement learning: Computational theory and biological mechanisms

   117 citations · 2007
3. 3
   Competitive-Cooperative-Concurrent Reinforcement Learning with Importance Sampling

   98 citations · 2004
4. 4
   Learning CPG-based biped locomotion with a policy gradient method

   98 citations · 2006
5. 5
   The Cyber Rodent Project: Exploration of Adaptive Mechanisms for Self-Preservation and Self-Reproduction

   71 citations · 2005
6. 6
   Reinforcement learning of dynamic motor sequence: learning to stand up

   68 citations · 2002
7. 7
   Constrained reinforcement learning from intrinsic and extrinsic rewards

   49 citations · 2007
8. 8
   Acquisition of stand-up behavior by a real robot using hierarchical reinforcement learning

   49 citations · 2001
9. 9
   Adaptive Baseline Enhances EM-Based Policy Search: Validation in a View-Based Positioning Task of a Smartphone Balancer

   41 citations · 2017
10. 10
    Reinforcement learning with via-point representation

    37 citations · 2004

Key Collaborators

Contact & Links