The webcast of my CERN talk on machine learning in high-energy physics
Interview (in French) with me in the Lettre IN2P3 InfomatiqueThe original English version
New machine learning paper
A (missing) boosting-type convergence result for AdaBoost.MH with factorized multi-class classifiers
In Conference on Computational Learning Theory (open problems track), 2014.
In this companion paper we showed that AdaBoost.MH works very well in practice with factorized base classifiers. This note states the open problem of the exponential convergence of the algorithm.
The return of AdaBoost.MH: multi-class Hamming trees
In International Conference on Learning Representations, 2014.
We train vector-valued decision trees within the framework of AdaBoost.MH. The key element of the method is a vector-valued decision stump, factorized into an input-independent vector of length K and label-independent scalar classifier.
Correlation-based construction of neighborhood and edge features
In International Conference on Learning Representations (workshop track), 2014.
Motivated by an abstract notion of low-level edge detector filters, we propose a simple method of unsupervised feature construction based on pairwise statistics of features. In the first step, we construct neighborhoods of features by regrouping features that correlate. Then we use these subsets as filters to produce new neighborhood features. Next, we connect neighborhood features that correlate, and construct edge features by subtracting the correlated neighborhood features of each other.
Introduction to multivariate discrimination
In T. Delemontex and A. Lucotte, editors, SOS 2012 – IN2P3 School of Statistics. EPJ Web of Conferences, 2013.
For physicists and other practicioners.
New physics paper
P. Abreu et
Muons in air showers at the Pierre Auger Observatory: Measurement of atmospheric production depth
Physical Review Letters D, 90:012012, 2014
Auger Collaboration paper.