Randall Balestriero
Randall Balestriero is an assistant professor of computer science at Brown University. He is known for his research in practical deep learning theory, self-supervised learning, and learnable signal processing, including work on parametrized wavelets deployed in NASA's Mars SEIS mission. He previously served as a postdoctoral researcher at Meta AI Research (FAIR) with Yann LeCun and as a quantitative researcher at Citadel's GQS, focusing on financial time-series prediction and representation learning. He co-authored Meta's Cookbook of Self-Supervised Learning and has developed theoretical frameworks such as the affine spline operator view of deep networks.
“The anti-collapse mechanism 'introduces only two practical hyperparameters... with λ_sig being the main parameter to tune'”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.