About

Hi there! Working from the CERN laboratory near the Swiss Alps, I am currently a Postdoctoral Associate at New York University with the Physics Department and the Center for Data Science. I do research in machine learning in close collaboration with Kyle Cranmer and the ATLAS experiment. Before that, I was a Doctoral Student at the University of Liège in Belgium.

As a researcher, my far ambition is to unlock discoveries of a new kind by making machine learning and artificial intelligence a cornerstone of the modern scientific method. Using particle physics as a test bed, my present research interests circle around how to use or design new machine learning algorithms to approach data-driven scientific problems in new and transformative ways. With this goal in mind, my topics of research include methods for simulator-based likelihood-free inference, algorithms to handle systematic uncertainties in inference models, and developments towards the automatization of sience.

Curriculum vitae.

Publications

See also Google Scholar or ORCID.

2017

2016

2015

  • Pitfalls of evaluating a classifier’s performance in high energy physics applications
  • Ethnicity sensitive author disambiguation using semi-supervised learning
  • Collaborative analysis of gigapixel images using Cytomine
  • Scikit-learn: Machine Learning Without Learning the Machinery
  • Solar Energy Prediction: An International Contest to Initiate Interdisciplinary Research on Compelling Meteorological Problems

2014

2013

2012

2011

  • Learning to rank with extremely randomized trees

2010

  • A zealous parallel gradient descent algorithm
  • Collaborative filtering: Scalable approaches using restricted Boltzmann machines

Talks

2016

2015

2014

2013

2012

2011

  • Large-scale machine learning for collaborative filtering

2010

  • A zealous parallel gradient descent algorithm

Software

Scikit-Learn

Scikit-Learn is an easy-to-use and general-purpose machine learning library written in Python. It integrates classic machine learning algorithms within the scientific Python ecosystem (numpy, scipy, matplotlib). I am an active developper since 2011.

Carl

Carl is a toolbox for likelihood-free inference in Python.