Research

My current research revolves around AI x Physics.

Weighing Black Holes with Deep/Machine Learning

• As an undergraduate at the University of Illinois at Urbana-Champaign (UIUC), my work involved developing machine learning and deep learning algorithms to weigh supermassive black holes employing time series data from the Sloan Digital Sky Survey.
  • Some Press
  • Conference Video
  • Poster
  • Code

Neural Scaling Laws

• Neural networks seem to obey peculiar scaling laws as a function of modifying their 1) model size and 2) training data size – particularly, predictable, power law-scaling.

Neural Networks Through the Lens of Statistical Field Theory

• Under the guidance of Dr. Yonatahn Kahn at UIUC in the high-energy theory group, I studied deep neural networks through the lens of statistical field theory and verified analytical results with numerical simulations. I worked to understand the nearly-Gaussian corrections on the distribution of weights of particular neurons in a deep linear network after passing data.