From learning to optimize to learning optimization algorithms (2024)
C. Castera, P. Ochs
Preprint
Welcome
I am a researcher in optimization and machine learning. Since February 2022, I am a postdoc at the University of Tübingen, working on the TRINOM-DS project led by Peter Ochs and Jalal Fadili.
Prior to that, I wrote my PhD thesis (readable here) at the IRIT laboratory in Toulouse (supervised by Cédric Févotte, Edouard Pauwels and Jérôme Bolte). I was a member of the FACTORY project.
My research focuses on optimization algorithms for machine learning (mainly second-order methods with application to deep learning), and learning optimization algorithms.
Research interests
- Continuous optimization (convex & non-convex, non-smooth, stochastic)
- Learning to optimize
- Optimization for deep learning & automatic hyper-parameter tuning
Selected publications
Near-optimal closed-loop method via Lyapunov damping for convex optimization (2023)
S. Maier, C. Castera, P. Ochs
Preprint
Continuous Newton-like methods featuring inertia and variable mass (2023)
C. Castera, H. Attouch, J. Fadili, P. Ochs
SIAM Journal on Optimization (SIOPT)
An inertial Newton algorithm for deep learning (2021)
C. Castera, J. Bolte, C. Févotte, E. Pauwels
Journal of Machine Learning Research (JMLR)
Teaching
From 2018 to 2021 I gave tutorials and practicals to master students at INP-ENSEEIHT where I taught the following:
- Tutorials & practicals on Probability (random variables, conditional expectation, etc.)
- Tutorials & practicals on Statistics (estimation, tests)
- Tutorials on Integration theory and Fourier transform
Contact
[firstname].[lastname]@protonmail.com