Margin Maximization as Lossless Maximal Compression

TL;DR

This paper interprets margin maximization in classification as achieving lossless maximal compression of training data, providing new insights into generalization and explaining the success of algorithms like gradient boosting.

Contribution

It offers an information-theoretic perspective on margin maximization, connecting it to lossless data compression and broadening understanding of generalization in supervised learning.

Findings

Margin maximization corresponds to lossless data compression.

The interpretation explains the success of gradient boosting.

The approach provides theoretical and empirical support for the connection.

Abstract

The ultimate goal of a supervised learning algorithm is to produce models constructed on the training data that can generalize well to new examples. In classification, functional margin maximization -- correctly classifying as many training examples as possible with maximal confidence --has been known to construct models with good generalization guarantees. This work gives an information-theoretic interpretation of a margin maximizing model on a noiseless training dataset as one that achieves lossless maximal compression of said dataset -- i.e. extracts from the features all the useful information for predicting the label and no more. The connection offers new insights on generalization in supervised machine learning, showing margin maximization as a special case (that of classification) of a more general principle and explains the success and potential limitations of popular learning algorithms like gradient boosting. We support our observations with theoretical arguments and empirical evidence and identify interesting directions for future work.