Detecting Underspecification with Local Ensembles
Abstract
We present local ensembles, a method for detecting underspecification -- when many possible predictors are consistent with the training data and model class -- at test time in a pre-trained model. Our method uses local second-order information to approximate the variance of predictions across an ensemble of models from the same class. We compute this approximation by estimating the norm of the component of a test point's gradient that aligns with the low-curvature directions of the Hessian, and provide a tractable method for estimating this quantity. Experimentally, we show that our method is capable of detecting when a pre-trained model is underspecified on test data, with applications to out-of-distribution detection, detecting spurious correlates, and active learning.
Cite
@article{arxiv.1910.09573,
title = {Detecting Underspecification with Local Ensembles},
author = {David Madras and James Atwood and Alex D'Amour},
journal= {arXiv preprint arXiv:1910.09573},
year = {2021}
}
Comments
Published as a conference paper at ICLR 2020 under the title "Detecting Extrapolation with Local Ensembles"