1

A Lower Bound and a Near-Optimal Algorithm for Bilevel Empirical Risk Minimization
We provide a lower bound for AID algorithm solving bilevel ERM problems. We show that SRBA, a bilevel adaptation of SARAH, is near optimal in term of sample complexity, similarly to its single level counterpart.
Mathieu Dagréou, Thomas Moreau, Samuel Vaiter, Pierre Ablin
PDF Cite arXiv PMLR
A framework for bilevel optimization that enables stochastic and global variance reduction algorithms
We propose a new framework that enables to adapt easily any first order stochastic optimization algorithm for bilevel optimization problems. In this framework, we propose SOBA, a direct adaptation of SGD, and SABA, a direct adaptation of SAGA.
Mathieu Dagréou, Pierre Ablin, Samuel Vaiter, Thomas Moreau
PDF Cite Code Poster Slides NeurIPS arXiv
Benchopt: Reproducible, efficient and collaborative optimization benchmarks
Numerical validation is at the core of machine learning research as it allows to assess the actual impact of new methods, and to …
Thomas Moreau, Mathurin Massias, Alexandre Gramfort, Pierre Ablin, Pierre-Antoine Bannier, Benjamin Charlier, Mathieu Dagréou, Tom Dupré La Tour, Ghislain Durif, Cassio F. Dantas, Quentin Klopfenstein, Johan Larsson, En Lai, Tanguy Lefort, Benoit Malézieux, Badr Moufad, Binh T. Nguyen, Alain Rakotomamonjy, Zaccharie Ramzi, Joseph Salmon, Samuel Vaiter
PDF Cite arXiv Benchopt