LGNov 1, 2023

Mahalanobis-Aware Training for Out-of-Distribution Detection

arXiv:2311.00808v1h-index: 5
Originality Incremental advance
AI Analysis

This addresses the critical need for reliable anomaly detection in open-world settings, though it appears incremental as it builds upon existing Mahalanobis distance methods.

The paper tackled the problem of detecting out-of-distribution samples for safe deep learning deployment by introducing a novel loss function and training recipe, resulting in a reduction of the false-positive rate by over 50% on CIFAR-10 far-OOD tasks.

While deep learning models have seen widespread success in controlled environments, there are still barriers to their adoption in open-world settings. One critical task for safe deployment is the detection of anomalous or out-of-distribution samples that may require human intervention. In this work, we present a novel loss function and recipe for training networks with improved density-based out-of-distribution sensitivity. We demonstrate the effectiveness of our method on CIFAR-10, notably reducing the false-positive rate of the relative Mahalanobis distance method on far-OOD tasks by over 50%.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes