Markov chain stochastic DCA and applications in deep learning with PDEs regularization

Hoang Phuc Hau Luu, Hoai Minh Le, Hoai An Le Thi

Abstract: Highlights•We study stochastic DC programs where i.i.d. samples are unavailable.•An algorithm named Markov chain stochastic DCA is proposed and analyzed.•Two realizations of the algorithm are designed in the context of deep learning.•We conduct numerical experiments on time series prediction and image classification.