Daphne Cornelisse

Hello! I am a second-year Ph.D. student at NYU, supervised by Professor Eugene Vinitsky. My research centers on developing effective and human-compatible agents. More concretely, I currently focus on two areas: 1) Controllable behavior generation, where I combine data-driven and mechanistic approaches to model human behavior and design reliable evaluation protocols; and 2) Adaptation, where I aim to understand the ingredients that enable agents to flexibly and safely adjust to unseen multi-agent settings.


Outside the lab, I enjoy boxing, going for a run along the East River, reading, and sketching.


Google Scholar / GitHub / Twitter / Goodreads

News

Papers 

GPUDrive: Data-driven, multi-agent driving simulation at 1 million FPS

Saman Kazemkhani*, Aarav Pandya*, Daphne Cornelisse*, Brennan Shacklett, Eugene Vinitsky

*Equal Contribution

In submission


Paper | Tweet | Code 

Human-compatible driving partners through data-regularized self-play reinforcement learning

Daphne Cornelisse, Eugene Vinitsky

Reinforcement Learning Conference (RLC), 2024 


Paper | Tweet | Project page | Code | Slides | Talk

Neural payoff machines: predicting fair and stable payoff allocations among team members 

Daphne Cornelisse, Thomas Rood, Mateusz Malinowski, Yoram Bachrach, Tal Kachman

NeurIPS 2022

Paper | Poster | Tweet | Master thesis