Thomas Pumir

3 papers · 2018–2025 · 3 venues · across top CS/AI venues

Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.

Achievements

Jump to papers ↓

🌍 Conference Polyglot · Bronze 🏃 Academic Marathon · Sprint 🌉 Interdisciplinary Bridge · Bronze

Venues

AISTATS (1) ICLR (1) NeurIPS (1)

Top co-authors

Nicolas Boumal (2) Stephane Rivaud (1) Samy Jelassi (1) Eugene Belilovsky (1) Lewis Liu (1) Edouard Oyallon (1) Michael Eickenberg (1) Damien Scieur (1) Louis Fournier (1)

Research topics

Deep Learning (2) Machine Learning (2) Mathematics & Optimization (2) Optimization (2) Optimization & Theory · Machine Learning (2) Architectures (1) Mathematics (1) Optimization & Theory · Deep Learning (1)

Keywords

semidefinite programming (1) matrix factorization (1) model parallelism (1) hessian approximation (1) low-rank matrix (1) distributed training (1) gradient computation (1) phase retrieval (1) quasi-newton method (1) memory efficiency (1) smoothed analysis (1) low-rank optimization (1) second-order stationary point (1) semidefinite program (1) reversible architecture (1) secant equation (1) symmetric update (1)

Papers

PETRA: Parallel End-to-end Training with Reversible Architectures ICLR 2025 Generalization of Quasi-Newton Methods: Application to Robust Symmetric Multisecant Updates AISTATS 2021 Smoothed analysis of the low-rank approach for smooth semidefinite programs NeurIPS 2018