El Mahdi Chayti

Joined

September 2021

Names

El Mahdi Chayti (Preferred)

,

EL MAHDI CHAYTI

Emails

****@epfl.ch (Confirmed)

Personal Links

Semantic Scholar

Career & Education History

PhD student

EDIC, Swiss Federal Institute of Technology Lausanne (epfl.ch)

2020 – 2025

MS student

École Polytechnique (polytechnique.edu)

2018 – 2019

MS student

Ecole Nationale Supérieure de Techniques Avancées (ensta-paris.fr)

2018 – 2019

Undergrad student

École Polytechnique (polytechnique.edu)

2015 – 2018

Advisors, Relations & Conflicts

PhD Advisor

2020 – 2025

Expertise

No areas of expertise listed

Publications

A Split-Client Approach to Second-Order Optimization
El Mahdi Chayti, Martin Jaggi
- Submitted to ICLR 2026
- Readers: Everyone
$\alpha$-LoRA: Effective Fine-Tuning via Base Model Rescaling
Aymane El Firdoussi, El Mahdi Chayti, Mohamed El Amine Seddik, Martin Jaggi
- Submitted to ICLR 2026
- Readers: Everyone
Faster Gradient Methods for Highly-smooth Stochastic Bilevel Optimization
Lesi Chen, Junru Li, El Mahdi Chayti, Jingzhao Zhang
- ICLR 2026 Poster
- Readers: Everyone
Faster Gradient Methods for Highly-smooth Stochastic Bilevel Optimization
Lesi Chen, Junru Li, El Mahdi Chayti, Jingzhao Zhang
- NeurIPS 2025 Workshop
- Readers: Everyone
Improving Stochastic Cubic Newton with Momentum
El Mahdi Chayti, Nikita Doikov, Martin Jaggi
- AISTATS 2025 Poster
- Readers: Everyone
Unified Convergence Theory of Stochastic and Variance-Reduced Cubic Newton Methods
El Mahdi Chayti, Martin Jaggi, Nikita Doikov
- Accepted by TMLR
- Readers: Everyone
Optimization with Access to Auxiliary Information
El Mahdi Chayti, Sai Praneeth Karimireddy
- Accepted by TMLR
- Readers: Everyone
Unified Convergence Theory of Stochastic and Variance-Reduced Cubic Newton Methods
El Mahdi Chayti, Martin Jaggi, Nikita Doikov
- Submitted to NeurIPS 2023
- Readers: Everyone
Second-Order Optimization with Lazy Hessians
Nikita Doikov, El Mahdi Chayti, Martin Jaggi
- ICML 2023 OralPoster
- Readers: Everyone

Co-Authors