Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

SSD: A Unified Framework for Self-Supervised Outlier Detection

About

We ask the following question: what training information is required to design an effective outlier/out-of-distribution (OOD) detector, i.e., detecting samples that lie far away from the training distribution? Since unlabeled data is easily accessible for many applications, the most compelling approach is to develop detectors based on only unlabeled in-distribution data. However, we observe that most existing detectors based on unlabeled data perform poorly, often equivalent to a random prediction. In contrast, existing state-of-the-art OOD detectors achieve impressive performance but require access to fine-grained data labels for supervised training. We propose SSD, an outlier detector based on only unlabeled in-distribution data. We use self-supervised representation learning followed by a Mahalanobis distance based detection in the feature space. We demonstrate that SSD outperforms most existing detectors based on unlabeled data by a large margin. Additionally, SSD even achieves performance on par, and sometimes even better, with supervised training based detectors. Finally, we expand our detection framework with two key extensions. First, we formulate few-shot OOD detection, in which the detector has access to only one to five samples from each class of the targeted OOD dataset. Second, we extend our framework to incorporate training data labels, if available. We find that our novel detection framework based on SSD displays enhanced performance with these extensions, and achieves state-of-the-art performance. Our code is publicly available at https://github.com/inspire-group/SSD.

Vikash Sehwag, Mung Chiang, Prateek Mittal• 2021

Related benchmarks

TaskDatasetResultRank
Image ClassificationImageNet-1K
Top-1 Acc78.8
836
Out-of-Distribution DetectionSUN OOD with ImageNet-1k In-distribution (test)
FPR@9578.23
159
Out-of-Distribution DetectionTextures
AUROC0.8618
141
Out-of-Distribution DetectionImageNet OOD Average 1k (test)
FPR@9540.64
137
Out-of-Distribution DetectionCIFAR-100
AUROC88.97
107
Out-of-Distribution DetectionCIFAR-10
AUROC97.36
105
OOD DetectionCIFAR-100 standard (test)
AUROC (%)88.91
94
Out-of-Distribution DetectionCIFAR-10 vs CIFAR-100 (test)
AUROC94.1
93
OOD DetectionCIFAR-10 (IND) SVHN (OOD)
AUROC0.9941
91
Out-of-Distribution DetectionCIFAR-100 SVHN in-distribution out-of-distribution (test)
AUROC99.1
90
Showing 10 of 75 rows
...

Other info

Code

Follow for update