Do Neural Networks Generalize from Self-Averaging Sub-classifiers in the Same Way As Adaptive Boosting?

TL;DR

This paper investigates whether neural networks generalize in the same way as adaptive boosting by analyzing their behavior as ensembles of interpolating sub-classifiers, supported by theoretical and experimental evidence.

Contribution

It establishes a novel connection between the generalization mechanisms of deep neural networks and boosted classifiers, highlighting the role of self-averaging over sub-classifiers.

Findings

Neural networks learn a series of boosted classifiers.

Dropout-trained NNs exhibit self-averaging behavior similar to boosting.

Theoretical analysis links NNs' generalization to ensemble averaging.

Abstract

In recent years, neural networks (NNs) have made giant leaps in a wide variety of domains. NNs are often referred to as black box algorithms due to how little we can explain their empirical success. Our foundational research seeks to explain why neural networks generalize. A recent advancement derived a mutual information measure for explaining the performance of deep NNs through a sequence of increasingly complex functions. We show deep NNs learn a series of boosted classifiers whose generalization is popularly attributed to self-averaging over an increasing number of interpolating sub-classifiers. To our knowledge, we are the first authors to establish the connection between generalization in boosted classifiers and generalization in deep NNs. Our experimental evidence and theoretical analysis suggest NNs trained with dropout exhibit similar self-averaging behavior over interpolating sub-classifiers as cited in popular explanations for the post-interpolation generalization phenomenon in boosting.