Jeremy Bernstein
Professor · MIT
Jeremy Bernstein, an MIT professor, studies stochastic gradient optimization and neural networks.
- Approximation Theory

- Representation Learning: Theory

- Scaling Rules for Optimization

- Metrized Deep Learning

Playlists
1About
Jeremy Bernstein, an MIT professor, studies stochastic gradient optimization and neural networks. His graduate lectures for MIT 6.7960: Deep Learning prove universal approximation for ReLU networks, link wide networks to Gaussian processes, and recast gradient descent under chosen norms.
Taught with
Same subject, AI & Machine Learning
Connections
7Their places, courses, books and lectures.


