Tyler Westenbroek

I am currently a Postdoctoral Researcher in the Paul G. Allen School of Computer Science & Engineering at the University of Washington, working with Abhishek Gupta. My research focuses on dexterous manipulation for robotics, by scaling RL training in simulation and developing efficient algorithms for real-world adaptation.

I received my PhD in Electrical Engineering and Computer Sciences from UC Berkeley (2023), working on machine learning and control under Shankar Sastry. I received my B.S. in Systems Engineering and Computer Science from Washington University, working with Humberto Gonzalez on hybrid systems.

westenbroekt[at]gmail[dot]com  /  CV  /  Scholar

Publications

Emergent Dexterity Via Diverse Resets and Large-Scale Reinforcement Learning

Simulation Distillation: Pretraining World Models in Simulation for Rapid Real-World Adaptation

RFS: Reinforcement Learning with Residual Flow Steering for Dexterous Manipulation

Rapidly Adapting Policies to the Real World via Simulation-Guided Fine-Tuning

Dense Dynamics-Aware Reward Synthesis: Integrating Prior Experience with Demonstrations

Learning to Walk from Three Minutes of Real-World Data with Semi-Structured Dynamics Models

The power of learned locally linear models for nonlinear policy optimization

Enabling Efficient, Reliable Real-World Reinforcement Learning with Approximate Physics-Based Models

Reinforcement Learning with Simple Dynamics Models and Low-Level Feedback Controllers

Lyapunov Design for Robust and Efficient Robotic Reinforcement Learning

On the Computational Consequences of Cost Function Design for Nonlinear Optimal Control

On the Stability of Nonlinear Receding Horizon Control: A Geometric Perspective

Combining Model-Based Design and Model-Free Policy Optimization to Learn Safe, Stabilizing Controllers

Smooth approximations for hybrid optimal control problems with application to robotic walking

Feedback linearization for uncertain systems via reinforcement learning

Learning min-norm stabilizing control laws for systems with unknown dynamics

Improving Input-Output Linearizing Controllers for Bipedal Robots via Reinforcement Learning

High Confidence Sets for Trajectories of Stochastic Time-Varying Nonlinear Systems

Adaptive control for linearizable systems using on-policy reinforcement learning

Competitive Statistical Estimation with Strategic Data Sources

A New Solution Concept and Family of Relaxations for Hybrid Dynamical Systems

Optimal control of piecewise-smooth control systems via singular perturbations

Statistical Estimation with Strategic Data Sources in Competitive Settings

Forked from Leonid Keselman.