About
Bassel El Mabsout completed his Ph.D. in the lab in Summer 2025. His research is about turning high-level objectives into robust learned robot behavior, especially on resource-constrained platforms and across the gap from simulation to reality. In the lab this produced smooth reinforcement-learning flight controllers (CAPS), compact actor networks, and Fulfillment Priority Logic. He is also a co-founder of the Neobotics Foundation.
Code and hardware artifacts
-
FPL Neural Lyapunov Software Maintainer
Training code for Lyapunov-stable neural controllers using Fulfillment Priority Logic. -
CAPS: Conditioning for Action Policy Smoothness Software Maintainer
Regularization that makes reinforcement-learning controllers produce smooth actions. -
Little Actor Software Maintainer
Training code and extended results on shrinking actors in actor-critic reinforcement learning.
6 publications with the lab
- Accelerating Lyapunov-Stable Neural Control using Fulfillment Priority Logic
- Burning Fetch Execution: A Framework for Zero-Trust Multi-party Confidential Computing
- Closing the Intent-to-Reality Gap via Fulfillment Priority Logic
- Honey. I Shrunk The Actor: A Case Study on Preserving Performance with Smaller Actors in Actor-Critic RL
- Regularizing Action Policies for Smooth Control with Reinforcement Learning
- How to Train your Quadrotor: A Framework for Consistently Smooth and Responsive Flight Control via Reinforcement Learning