Felipe Nuti

I am a 4th year Mathematics and Computer Science undergraduate at the University of Oxford, doing research at the Visual Geometry Group under Tim Franzmeyer and João Henriques.

My main research interest is generative modeling; in particular diffusion models and their properties relating to steerability and interpretability. Recently, we proposed the first method for extracting reward functions from diffusion models.

I am also interested in learning more about theoretical perspectives on foundation models.

This past summer, I interned at Citadel Securities as a Quantitative Researched in Systematic FX. Before that, I did an internship on Causal Inference and Machine Learning at QuantCo Zürich.

Email  /  Scholar  /  LinkedIn  /  Twitter  /  Github

Research

A method for quantifying differences in preferences between any two diffusion models.

Based on Jon Barron's website source code.