Mahalanobis distance (MD) is a simple and popular post-processing method for detecting out-of-distribution (OOD) inputs in neural networks. We analyze its failure modes for near-OOD detection and propose a simple fix called relative Mahalanobis distance (RMD) which improves performance and is more robust to hyperparameter choice. On a wide selection of challenging vision, language, and biology OOD benchmarks (CIFAR-100 vs CIFAR-10, CLINC OOD intent detection, Genomics OOD), we show that RMD meaningfully improves upon MD performance (by up to 15% AUROC on genomics OOD).
@article{arxiv.2106.09022,
title = {A Simple Fix to Mahalanobis Distance for Improving Near-OOD Detection},
author = {Jie Ren and Stanislav Fort and Jeremiah Liu and Abhijit Guha Roy and Shreyas Padhy and Balaji Lakshminarayanan},
journal= {arXiv preprint arXiv:2106.09022},
year = {2021}
}