Dynamics of Algorithms in Disordered Systems and Learning Problems
Event details
| Date | 25.08.2023 |
| Hour | 14:00 › 16:00 |
| Speaker | Matteo Vilucchio |
| Location | |
| Category | Conferences - Seminars |
EDIC candidacy exam
Exam president: Prof. Nicolas Macris
Thesis advisor: Prof. Florent Krzakala
Co-examiner: Prof. Lenka Zdeborová
Abstract
The purpose is to study ways of sampling equilibrium measures of hamiltonians coming from spin glasses and to look at the dynamics of algorithms theoretically. I will also consider learning problems arising from machine learning.
Background papers
- Surprises in High-Dimensional Ridgeless Least Squares Interpolation by Trevor Hastie, Andrea Montanari, Saharon Rosset, Ryan J. Tibshirani; arXiv:1903.08560 [math.ST]
- On the Optimal Weighted ℓ2 Regularization in Overparameterized Linear Regression by Denny Wu, Ji Xu; NeurIPS 2020
- Asymptotics of Ridge (less) Regression under General Source Condition by Dominic Richards, Jaouad Mourtada, Lorenzo Rosasco; AISTATS 2021: 3889-3897
Exam president: Prof. Nicolas Macris
Thesis advisor: Prof. Florent Krzakala
Co-examiner: Prof. Lenka Zdeborová
Abstract
The purpose is to study ways of sampling equilibrium measures of hamiltonians coming from spin glasses and to look at the dynamics of algorithms theoretically. I will also consider learning problems arising from machine learning.
Background papers
- Surprises in High-Dimensional Ridgeless Least Squares Interpolation by Trevor Hastie, Andrea Montanari, Saharon Rosset, Ryan J. Tibshirani; arXiv:1903.08560 [math.ST]
- On the Optimal Weighted ℓ2 Regularization in Overparameterized Linear Regression by Denny Wu, Ji Xu; NeurIPS 2020
- Asymptotics of Ridge (less) Regression under General Source Condition by Dominic Richards, Jaouad Mourtada, Lorenzo Rosasco; AISTATS 2021: 3889-3897
Practical information
- General public
- Free