作者: Patrick L. Combettes , Jean-Christophe Pesquet
DOI:
关键词:
摘要: Motivated by structures that appear in deep neural networks, we investigate nonlinear composite models alternating proximity and affine operators defined on different spaces. We first show a wide range of activation used networks are actually operators. then establish conditions for the averagedness proposed constructs their asymptotic properties. It is shown limit resulting process solves variational inequality which, general, does not derive from minimization problem.