Dear Colleagues,
We would like to invite you to the following SPASS seminar, jointly organized by UniPi, SNS, UniFi and UniSi (abstract below):
*Large and moderate deviations for Gaussian neural network* by *Claudio Macci* (Università di Roma Tor Vergata)
The seminar will take place on *TUE, 29.10.2024* at *14:00 CET *in Aula Seminari, Dipartimento di Matematica, UNIPI and streamed online at this link https://meet.google.com/gji-phwo-vbg.
The organizers, A. Agazzi, G. Bet, A. Caraceni, F. Grotto, G. Zanco https://sites.google.com/unipi.it/spass https://www.google.com/url?q=https://sites.google.com/unipi.it/spass&source=gmail-imap&ust=1665669490000000&usg=AOvVaw07o0tKOUGmZjDDF2Ta7pQY
--------------------------------------------
*Title: Large and moderate deviations for Gaussian neural network*
*Abstract: We prove large and moderate deviations for the output of Gaussian fully connected neural networks. The main achievements concern deep neural networks (i.e., when the model has more than one hidden layer) and hold for bounded and continuous pre-activation functions. However, for deep neural networks fed by a single input, we have results even if the pre-activation is ReLU. When the network is shallow (i.e., there is exactly one hidden layer) the large and moderate principles hold for quite general pre-activations and in an infinite-dimensional setting.Joint work with Barbara Pacchiarotti (Università di Roma Tor Vergata) and Giovanni Luca Torrisi (IAC-CNR).*