Zobrazeno 1 - 10
of 19
pro vyhledávání: '"Arous, G. Ben"'
Autor:
Baity-Jesi, M., Sagun, L., Geiger, M., Spigler, S., Arous, G. Ben, Cammarota, C., LeCun, Y., Wyart, M., Biroli, G.
Publikováno v:
PMLR 80:324-333, 2018; Republication with DOI (cite this one): J. Stat. Mech. (2019) 124013
We analyze numerically the training dynamics of deep neural networks (DNN) by using methods developed in statistical physics of glassy systems. The two main issues we address are (1) the complexity of the loss landscape and of the dynamics within it,
Externí odkaz:
http://arxiv.org/abs/1803.06969
Autor:
Arous, G. Ben, Gun, O.
We consider Random Hopping Time (RHT) dynamics of the Sherrington - Kirkpatrick (SK) model and p-spin models of spin glasses. For any of these models and for any inverse temperature we prove that, on time scales that are sub-exponential in the dimens
Externí odkaz:
http://arxiv.org/abs/1010.5190
We give an asymptotic evaluation of the complexity of spherical p-spin spin-glass models via random matrix theory. This study enables us to obtain detailed information about the bottom of the energy landscape, including the absolute minimum (the grou
Externí odkaz:
http://arxiv.org/abs/1003.1129
Autor:
Arous, G. Ben, Kargin, V.
Publikováno v:
Probability Theory and Related Fields (2010), 147, 161-183
We continue here the study of free extreme values begun in Ben Arous and Voiculescu (2006). We study the convergence of the free point processes associated with free extreme values to a free Poisson random measure (Voiculescu (1998), Barndorff-Nielse
Externí odkaz:
http://arxiv.org/abs/0903.2672
Publikováno v:
J. Stat. Mech. (2008) L04003
Aging has become the paradigm to describe dynamical behavior of glassy systems, and in particular spin glasses. Trap models have been introduced as simple caricatures of effective dynamics of such systems. In this Letter we show that in a wide class
Externí odkaz:
http://arxiv.org/abs/0712.1914
In this letter we announce rigorous results on the phenomenon of aging in the Glauber dynamics of the random energy model and their relation to Bouchaud's 'REM-like' trap model. We show that, below the critical temperature, if we consider a time-scal
Externí odkaz:
http://arxiv.org/abs/cond-mat/0110223
Publikováno v:
Probability Theory & Related Fields. 2001, Vol. 120 Issue 1, p1. 67p.
Autor:
Arous, G. Ben, Guionnet, A.
Publikováno v:
Probability Theory & Related Fields. 1997, Vol. 108 Issue 4, p517. 26p.
Autor:
AROUS, G. BEN1 benarous@cims.nyu.edu
Publikováno v:
Xviith International Congress on Mathematical Physics. 2013, p403-403. 1p.
Autor:
Arous, G. Ben, Léandre, R.
Publikováno v:
Probability Theory and Related Fields; September 1991, Vol. 90 Issue: 3 p377-402, 26p