Dark matter and microlensing

My research until 2004 focused on the analysis of astrophysical data : I help develop computer codes that scan enormous amounts of lightcurves in order to give, for each of them, a probability that the underlying object belongs to a desired family of transient objects (variable stars, microlensings, novae, and others).

For my PhD work, I worked on the data analysis in the AGAPE collaboration. We were looking for microlensing events hiding in CCD images. I wrote a small introduction to what AGAPE and then POINT-AGAPE. Apart from some existing statistical criteria that I improved, I also devised some new selection criteria which I incorporated in a data analysis pipeline.

During my postdoc in the Theoretical Physics Department of the University of Oxford, I pursued what I had initiated during my PhD, but instead of using statistical criteria that work serially, I wanted to find some statistical criteria that work in parallel, and I succeeded thanks to the use of artificial neural networks (ANN), which had never been used in this context before. I continued this work with Vasily Belokurov and Wyn Evans, who are at the Institute of Astronomy at the University of Cambridge, and some papers were published showing the power of this method.

Modeling of open source vs closed source development

Damien Challet and I developed a simple model that compares the development of software in the open source and closed source framework. What we found had quite an echo in the Internet community, and we are now working on some improvements in the model. Nature Science Update made a summary of our work.