LGJun 6, 2021

Exploring the Limits of Out-of-Distribution Detection

arXiv:2106.03004v3415 citations
Originality Highly original
AI Analysis

It addresses a critical challenge in AI safety and reliability for deploying models in real-world scenarios, with incremental advancements in performance across various data modalities.

The paper tackles the problem of near out-of-distribution detection in deep neural networks, achieving significant improvements in AUROC, such as from 85% to over 96% on CIFAR-100 vs CIFAR-10 using pre-trained transformers, and up to 99.46% with few-shot outlier exposure.

Near out-of-distribution detection (OOD) is a major challenge for deep neural networks. We demonstrate that large-scale pre-trained transformers can significantly improve the state-of-the-art (SOTA) on a range of near OOD tasks across different data modalities. For instance, on CIFAR-100 vs CIFAR-10 OOD detection, we improve the AUROC from 85% (current SOTA) to more than 96% using Vision Transformers pre-trained on ImageNet-21k. On a challenging genomics OOD detection benchmark, we improve the AUROC from 66% to 77% using transformers and unsupervised pre-training. To further improve performance, we explore the few-shot outlier exposure setting where a few examples from outlier classes may be available; we show that pre-trained transformers are particularly well-suited for outlier exposure, and that the AUROC of OOD detection on CIFAR-100 vs CIFAR-10 can be improved to 98.7% with just 1 image per OOD class, and 99.46% with 10 images per OOD class. For multi-modal image-text pre-trained transformers such as CLIP, we explore a new way of using just the names of outlier classes as a sole source of information without any accompanying images, and show that this outperforms previous SOTA on standard vision OOD benchmark tasks.

Code Implementations1 repo
Foundations

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

Your Notes