Itay Lavie
pronounced ee-TIE lah-VEE
Hi! I’m Itay, an applied mathematics PhD student at Harvard in the Pehlevan Group.
I work at the intersection of physics, computer science, and mathematics. I’m fascinated by how simple, well-understood units interact to produce rich emergent behavior, especially intelligence. That means I mostly do ML theory using tools from statistical physics.
Outside research, I enjoy hiking 🥾, trekking 🏔️, scuba diving 🤿, and paragliding 🪂.
Feel free to reach out.
News
| Oct 06, 2026 | New preprint with Clarissa Lauditi and Cengiz Pehlevan — A Solvable Model of Adaptive Learning Rate Rescaling, where we use DMFT to study how normalized SGD accelerates training, where that acceleration breaks down at the edge of stability, and what it means for trading off width vs. batch size. |
|---|---|
| Sep 27, 2026 | I will present Phase Transitions in Attention in NeurIPS, see you in Sydney! |
| Jun 16, 2026 | New preprint out — Phase Transitions in Attention, where we work out a Bayesian theory for how copy heads emerge, and find a an interesting phase diagram. |
| Sep 01, 2025 | Started my PhD in applied mathematics at Harvard, joining the Pehlevan Group. |
Selected publications
- NeurIPSAccepted to NeurIPS 2026, 2026
- ICMLIn Proceedings of the 41st International Conference on Machine Learning, 2024