Weight decay induced phase transitions in multilayer neural networks

Дата и время публикации : 1999-01-19T14:22:47Z

Авторы публикации и институты :
M. Ahr
M. Biehl
E. Schloesser

Ссылка на журнал-издание: Ссылка на журнал-издание не найдена
Коментарии к cтатье: 7 pages, 7 figures
Первичная категория: cond-mat.dis-nn

Все категории : cond-mat.dis-nn

Краткий обзор статьи: We investigate layered neural networks with differentiable activation function and student vectors without normalization constraint by means of equilibrium statistical physics. We consider the learning of perfectly realizable rules and find that the length of student vectors becomes infinite, unless a proper weight decay term is added to the energy. Then, the system undergoes a first order phase transition between states with very long student vectors and states where the lengths are comparable to those of the teacher vectors. Additionally in both configurations there is a phase transition between a specialized and an unspecialized phase. An anti-specialized phase with long student vectors exists in networks with a small number of hidden units.

Category: Physics