Alex Lambert
 |
Researcher |
About me
I'm taking a leave from research to work as a data engineer at Dataminded.
Before that, I was a researcher associated with the KU Leuven university, where I worked with Johan Suykens on kernel methods and duality.
I have a PhD in Machine Learning, prepared at Télécom Paris under the supervision of Florence d'Alché-Buc and Zoltan Szabo.
My PhD aimed at efficiently predicting functional outputs using kernels, the slides and manuscript are available.
I graduated from Télécom Paris and M2 Data Science from Institut Polytechnique de Paris.
More information about my background can be found on my resume.
Research Interests
I'm interested in machine learning, and my research topics include:
Operator-valued kernels, integral operators, random features for large scale learning
Convex optimization, shape constraints, differences of convex functions
Multi-task learning, functional output regression, quantile regression
Kernel PCA, kernel SVD
My list of publication is available here.
News
I just started a new position as data engineer @Dataminded.
New ICML paper accepted ! “Learning in Feature Spaces via Coupled Covariances: Asymmetric Kernel SVD and Nyström method” -> link.
I was part of the organizing team of the DEEPK workhop on deep learning and kernel methods at Leuven.
I gave a talk at the MIND/SODA team seminar on “Robustness and sparsity through Moreau envelopes in kernel-based settings” slides.
Our paper “Extending Kernel PCA through Dualization: Sparsity, Robustness, and Fast Algorithms” has been accepted at ICML !
Our paper on robust and sparse functional output regression has been accepted at ICML !