Fast Structure Learning for Deep Feedforward Networks via Tree Skeleton Expansion

Zhourong Chen, Xiaopeng Li, Zhiliang Tian, Nevin L. Zhang

Abstract: Despite the popularity of deep learning, structure learning for deep models remains a relatively under-explored area. In contrast, structure learning has been studied extensively for probabilistic graphical models (PGMs). In particular, an efficient algorithm has been developed for learning a class of tree-structured PGMs called hierarchical latent tree models (HLTMs), where there is a layer of observed variables at the bottom and multiple layers of latent variables on top. In this paper, we propose a simple unsupervised method for learning the structures of feedforward neural networks (FNNs) based on HLTMs. The idea is to expand the connections in the tree skeletons from HLTMs and to use the resulting structures for FNNs. Our method is very fast and it yields deep structures of virtually the same quality as those produced by the very time-consuming grid search method.