David Millard

Presented Federated IRL Work at L4DC 2026

2026-06-17T13:00:00+00:00

I attended the 8th Annual Learning for Dynamics & Control Conference (L4DC 2026) at the University of Southern California, where I presented our work, Can Optimal Transport Improve Federated Inverse Reinforcement Learning?. The paper introduces an optimal transport-based approach for combining locally learned reward functions across heterogeneous agents.

Joined the Human-Centered Computing Lab at the University of Rochester

2026-06-01T13:00:00+00:00

I’m excited to share that I joined the Human-Centered Computing Lab at the University of Rochester under the advisement of Distinguished Professor Dr. Mark Bocko. I look forward to contributing to the lab’s research and beginning this new collaboration.

Paper Accepted to the ICLR 2026 AI&PDE Workshop

2026-03-01T14:00:00+00:00

I’m excited to share that our paper was accepted to the ICLR 2026 Workshop on AI and Partial Differential Equations (AI&PDE).

Presented PEARL at the 2026 Joint Mathematics Meetings

2026-01-06T14:00:00+00:00

I attended the 2026 Joint Mathematics Meetings in Washington, D.C., where I gave a talk in the SIAM Minisymposium on Recent Advances in Numerical Linear Algebra. I presented our work, PEARL: Preconditioner Enhancement through Actor-critic Reinforcement Learning, which develops a reinforcement learning approach for learning matrix preconditioners.

Selected as an AI-PROWIL IRES Scholar

2025-05-26T13:00:00+00:00

I’m honored to have been selected as an AI-PROWIL IRES Scholar for the Fall of 2025! Through this program, I’ll conduct collaborative research with University West in Trollhättan Sweden, focusing on real-time deviation control for additive manufacturing processes.

Accepted to SURF Research Fellowship

2025-05-16T13:00:00+00:00

I’m excited to share that I’ve been accepted into the SURF Research Fellowship! As part of this opportunity, I’ll continue my work on applying conformal prediction in function space, a direction that bridges rigorous uncertainty quantification with the expressive power of scientific machine learning models.