Research


Past and Current Research

My current research focusses on solar information processing. The goal is to exploit a maximum of physical information from the large amounts of space solar data. This calls for the development of automated feature recognition tools and more largely for a data science that is grounded in statistics, image processing, observational expertise, and solar physics in order to access physical information in the data that is often hidden due to their dynamical range, noise, or complexity and size of the dataset.

At Rice University, I worked on distributed estimation in sensor network. The aim is to find distributed estimation algorithms that are local, scalable, energy-consumption efficient, and fault-tolerant.

My doctoral research was concerned with the construction of wavelet transforms that automatically adapt to to the design at-hand, be it irregular, stochastic or auto-regressive. This allows to have a fast and efficient denoising method for nonparametric regression. Univariate and bivariate irregular designs were considered.

Publications

Solar physics

Sensor Networks

Wavelet methods

Conference Proceedings

PhD Thesis

Nonparametric stochastic regression using design-adapted wavelets [pdf] (3.885Kb)

Master Thesis


Last update: May 2017