Scholar
Randall Balestriero
Google Scholar ID: S1x_xqcAAAAJ
AI Researcher
Citations & Impact
Citations
3,148
H-index
28
i10-index
49
Publications
20
Co-authors
13
list available
Publications
Resume (English only)
Academic Achievements
- Published multiple papers and participated in several conference tutorials. For example, Meta AI released the Self Supervised Learning Cookbook, AI discovered new events in geophysical data for earthquake detection, and NYU CDS at NeurIPS 2023 discussions.
Research Experience
- Has been conducting research in learnable signal processing since 2013, particularly with learnable parametrized wavelets. Expanded research scope at Rice University in 2016, exploring deep networks from a theoretical perspective using affine spline operators, improving state-of-the-art methods like batch-normalization and generative networks. Further broadened research interests at Meta AI Research (FAIR) in 2021, including self-supervised learning and biases emerging from data-augmentation and regularization. Joined GQS, Citadel in 2023 to work on highly noisy and nonstationary financial time-series, providing AI solutions for prediction and representation learning.
Education
- Joined Rice University in 2016 for a PhD under Prof. Richard Baraniuk; joined Meta AI Research (FAIR) in 2021 for a postdoc under Prof. Yann LeCun.
Background
- Research interests include self-supervised learning, learnable signal processing, spline theory, generative AI, and large language models. Focuses on areas such as computer vision, NLP, bioacoustics, geophysics, medical data, and quantitative finance. Committed to developing novel theoretical solutions to guide practitioners, safeguard users, and pave the way towards truly autonomous AI solutions.
Miscellany
- Personal interests include using tools from signal processing to better understand neural networks.
