Papers
31
Total Citations
3,154
H-Index
17
About
Yuval Tassa is a leading researcher at the intersection of robotics, optimal control, and reinforcement learning, whose work has fundamentally shaped how complex physical systems are controlled and simulated. Best known for advancing Differential Dynamic Programming (DDP) and online trajectory optimization, Tassa has demonstrated that real-time model predictive control is achievable even for high-dimensional humanoid robots — a long-standing challenge in the field. His 2012 paper on synthesizing complex behaviors through trajectory optimization (760 citations) remains a landmark contribution, showing humanoid robots recovering from falls and performing acrobatic maneuvers autonomously. Tassa's influence extends to the broader research infrastructure that underpins modern robotics and reinforcement learning. As a key contributor to MuJoCo and the dm_control software suite (collectively hundreds of citations), he helped establish the simulation environments that countless researchers now rely upon daily. His work on safe exploration in reinforcement learning addresses the critical challenge of deploying learning agents in real-world systems without catastrophic constraint violations. More recently, his research on bipedal robots learning soccer skills through deep RL (147 citations) showcases his continued push toward agile, learned motor behaviors, cementing his reputation as a uniquely versatile figure bridging classical control theory and modern machine learning.
Research Focus
Key Achievements
Top Papers
- 1
- 2Control-limited differential dynamic programming441 citations · 2014
- 3
- 4Safe Exploration in Continuous Action Spaces275 citations · 2018
- 5Whole-body model-predictive control applied to the HRP-2 humanoid228 citations · 2015
- 6dm_control: Software and tasks for continuous control186 citations · 2020
- 7
- 8An integrated system for real-time model predictive control of humanoid robots126 citations · 2013
- 9Data-efficient Deep Reinforcement Learning for Dexterous Manipulation118 citations · 2017
- 10Receding Horizon Differential Dynamic Programming117 citations · 2007