Manfred Diaz
Papers
3
Total Citations
37
H-Index
2
About
Manfred Diaz is a researcher advancing the frontiers of robot learning and domain transfer. His most influential work, "Active Domain Randomization" (2019, 33 citations), critically examines how domain randomization techniques can improve an agent's ability to generalize across different environments. This research is pivotal for enabling robots to transfer skills from simulation to the real world without extensive retraining, addressing a fundamental challenge in sim-to-real transfer. Diaz also contributes to safe and interactive robot teaching through his work on "Uncertainty-Aware Policy Sampling and Mixing for Safe Interactive Imitation Learning" (2021), which tackles the distributional shift problem in imitation learning—a key barrier to deploying robots that learn from human demonstrations. His participation in the AI Driving Olympics at NeurIPS 2018 further demonstrates his commitment to benchmarking and advancing autonomous driving systems. With a focus on making robots more adaptable and safer to teach, Diaz’s research is shaping how machines learn from both simulated and real-world interactions.
Research Focus
Key Achievements
Top Papers
- 1Active Domain Randomization33 citations · 2019
- 2The AI Driving Olympics at NeurIPS 20182 citations · 2019
- 3