Systematics aware learning: a case study in High Energy Physics

Victor Estrade (1) , Cécile Germain (1) , Isabelle Guyon (1) , David Rousseau (2)

1 TAU - TAckling the Underspecified
2 LAL - Laboratoire de l'Accélérateur Linéaire

Experimental science often has to cope with systematic errors that coherently bias data. We analyze this issue on the analysis of data produced by experiments of the Large Hadron Collider at CERN as a case of supervised domain adaptation. Systematics-aware learning should create an efficient representation that is insensitive to perturbations induced by the systematic effects. We present an experimental comparison of the adversarial knowledge-free approach and a less data-intensive alternative.

Domains: Machine Learning [stat.ML]

Victor Estrade, Cécile Germain, Isabelle Guyon, David Rousseau. Systematics aware learning: a case study in High Energy Physics. ESANN 2018 - 26th European Symposium on Artificial Neural Networks, Apr 2018, Bruges, Belgium. ⟨hal-01715155⟩