Papers
14
Total Citations
322
H-Index
8
About
Haitham Bou Ammar is a prominent researcher whose work spans reinforcement learning, multi-task learning, robotics control, and safe autonomous systems. His most influential contribution, "Online Multi-Task Learning for Policy Gradient Methods" (2014, 149 citations), addressed a critical challenge in robotics: improving sample efficiency by enabling agents to transfer knowledge across related tasks using policy gradient algorithms — a breakthrough that significantly advanced the field of sequential decision-making. His early work on nonlinear control of quadrotor UAVs (2010) and landing controllers demonstrated strong foundations in dynamic systems and autonomous aerial robotics. Bou Ammar has consistently pushed the boundaries of robot learning in high-stakes, high-speed environments, with notable contributions to air hockey robotics, kinodynamic planning using deep neural networks, and safe reinforcement learning via Control Lyapunov Barrier Functions — addressing the critical gap between impressive RL performance and real-world safety guarantees. His use of probabilistic models such as Boltzmann Machines for activity recognition and user tracking further reflects his versatility across machine learning methodologies. Notably, his pedagogical work on teaching reinforcement learning through physical robots highlights a commitment to accessible AI education. Across his career, Bou Ammar has accumulated over 300 citations, establishing himself as an impactful voice in intelligent robotics and autonomous learning systems.
Research Focus
Key Achievements
Top Papers
- 1Online Multi-Task Learning for Policy Gradient Methods149 citations · 2014
- 2
- 3Nonlinear tracking and landing controller for quadrotor aerial robots44 citations · 2010
- 4
- 5Efficient and Reactive Planning for High Speed Robot Air Hockey14 citations · 2021
- 6
- 7Teaching Reinforcement Learning using a Physical Robot10 citations · 2012
- 8Inexpensive user tracking using Boltzmann Machines9 citations · 2014
- 9Robot Reinforcement Learning on the Constraint Manifold6 citations · 2021
- 10A Nonparametric Evaluation of SysML-based Mechatronic Conceptual Design5 citations · 2012