Why Does Unsupervised Pre-training Help Deep Learning?

TLDR

Recent work on deep architectures such as Deep Belief Networks and auto‑encoders has achieved impressive results, especially when an unsupervised pre‑training phase is used, yet the underlying reasons for this success remain unclear. This study seeks to explain how unsupervised pre‑training improves deep learning performance and why it matters for future advances. The authors test multiple hypotheses about pre‑training by running extensive simulations. Experiments show that pre‑training benefits deeper models, larger capacity, and more data, and that it steers learning toward minima that generalize better, supporting a regularization interpretation.

Abstract

Much recent research has been devoted to learning algorithms for deep architectures such as Deep Belief Networks and stacks of auto-encoder variants, with impressive results obtained in several areas, mostly on vision and language data sets. The best results obtained on supervised learning tasks involve an unsupervised learning component, usually in an unsupervised pre-training phase. Even though these new algorithms have enabled training deep models, many questions remain as to the nature of this difficult learning problem. The main question investigated here is the following: how does unsupervised pre-training work? Answering this questions is important if learning in deep architectures is to be further improved. We propose several explanatory hypotheses and test them through extensive simulations. We empirically show the influence of pre-training with respect to architecture depth, model capacity, and number of training examples. The experiments confirm and clarify the advantage of unsupervised pre-training. The results suggest that unsupervised pre-training guides the learning towards basins of attraction of minima that support better generalization from the training data set; the evidence from these results supports a regularization explanation for the effect of pre-training.