Séminaire de Probabilités et Statistique :

Le 30 mai 2022 à 16:00 - UM - Bât 09 - Salle de conférence (1er étage)

Présentée par Sardy Sylvain - Université de Genève

A phase transition for finding needles in nonlinear haystacks with LASSO artificial neural networks (Séminaire supplémentaire - 16h)

To fit sparse linear associations, a LASSO sparsity inducing penalty with a single hyperparameter provably allows to recover the important features (needles) with high probability in certain regimes even if the sample size is smaller than the dimension of the input vector (haystack).

We investigate whether a phase transition also exists to fit sparse nonlinear associations known as artificial neural networks. Using certain activation functions, proper selection of the penalty parameter $\lambda$ and a thresholding inducing optimization, we observe good results on simulated and real data.

Séminaire supplémentaire à 16h en salle 109 (IMAG, bâtiment 9).
Également retransmis sur zoom : https://umontpellier-fr.zoom.us/j/94087408185