arXiv Analytics

Sign in

arXiv:1806.05161 [stat.ML]AbstractReferencesReviewsResources

Overfitting or perfect fitting? Risk bounds for classification and regression rules that interpolate

Mikhail Belkin, Daniel Hsu, Partha Mitra

Published 2018-06-13Version 1

Many modern machine learning models are trained to achieve zero or near-zero training error in order to obtain near-optimal (but non-zero) test error. This phenomenon of strong generalization performance for "overfitted" / interpolated classifiers appears to be ubiquitous in high-dimensional data, having been observed in deep networks, kernel machines, boosting and random forests. Their performance is robust even when the data contain large amounts of label noise. Very little theory is available to explain these observations. The vast majority of theoretical analyses of generalization allows for interpolation only when there is little or no label noise. This paper takes a step toward a theoretical foundation for interpolated classifiers by analyzing local interpolating schemes, including geometric simplicial interpolation algorithm and weighted $k$-nearest neighbor schemes. Consistency or near-consistency is proved for these schemes in classification and regression problems. These schemes have an inductive bias that benefits from higher dimension, a kind of "blessing of dimensionality". Finally, connections to kernel machines, random forests, and adversarial examples in the interpolated regime are discussed.

Related articles: Most relevant | Search more
arXiv:1106.0730 [stat.ML] (Published 2011-06-03, updated 2017-05-22)
Rademacher complexity of stationary sequences
arXiv:1803.00276 [stat.ML] (Published 2018-03-01)
Model-Based Clustering and Classification of Functional Data
arXiv:1808.03064 [stat.ML] (Published 2018-08-09)
Gradient and Newton Boosting for Classification and Regression