Home

Melrose Roderick

I completed my Ph.D. at Carnegie Mellon University advised by Zico Kolter. My research focused on improving safety and robustness of reinforcement learning algorithms, allowing such methods to be applied to real-world problems, with a particular interest in environmental sustainability. My thesis committee included Ruslan Salakhutdinov, Jeff Schneider, and Felix Berkenkamp. My work was supported by an NSF Graduate Research Fellowship. I received my B.Sc. at Brown University in Applied Math and Computer Science.

Currently, I am working on something in stealth.

Publications

Provably Safe PAC-MDP Exploration Using Analogies

Melrose Roderick, Vaishnavh Nagarajan, and Zico Kolter
Proceedings of the 24th International Conference on Artificial Intelligence and Statistics (AISTATS), 2021
[video] [poster] [code]

Enforcing Robust Control Guarantees Within Neural Network Policies

Priya L. Donti, Melrose Roderick, Mahyar Fazlyab, and Zico Kolter
9th International Conference on Learning Representations (ICLR), 2021
[poster] [code]

Deep Abstract Q-Networks

Melrose Roderick, Christopher Grimm, and Stefanie Tellex
Proceedings of the 17th International Conference on Autonomous Agents and Multi Agent Systems (AAMAS), 2018, (Nominated for best student paper)
[poster] [code]

Mean Actor Critic

Kavosh Asadi, Cameron Allen, Melrose Roderick, Abdel-rahman Mohamed, George Konidaris, Michael Littman
arXiv preprint arXiv:1709.00503, 2017

Dex-Net 1.0: A Cloud-Based Network of 3D Objects for Robust Grasp Planning Using a Multi-Armed Bandit Model with Correlated Rewards

Jeffrey Mahler, Florian T. Pokorny, Brian Hou, Melrose Roderick, Michael Laskey, Mathieu Aubry, Kai Kohlhoff, Torsten Kröger, James Kuffner, Ken Goldberg
IEEE International Conference on Robotics and Automation (ICRA), 2016 (Nominated for best manipulation paper)