SSD: A Unified Framework for Self-Supervised Outlier Detection

Abstract

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.

ICLR 2021

References in corpus (10)

• Auto-Encoding Variational Bayes
• A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
• Unsupervised Representation Learning by Predicting Image Rotations
• A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
• Deep Anomaly Detection with Outlier Exposure
• Toward Supervised Anomaly Detection
• Learning Confidence for Out-of-Distribution Detection in Neural Networks
• Anomaly Detection using One-Class Neural Networks
• Classification-Based Anomaly Detection for General Data
• Reducing Network Agnostophobia

Cited by in corpus (12)

• A Unified Survey on Anomaly, Novelty, Open-Set, and Out-of-Distribution Detection: Solutions and Future Challenges
• Do We Really Need to Learn Representations from In-domain Data for Outlier Detection?
• SLA${}^2$P: Self-supervised Anomaly Detection with Adversarial Perturbation
• Deep Variational Semi-Supervised Novelty Detection
• NGC: A Unified Framework for Learning with Open-World Noisy Data
• OMASGAN: Out-of-Distribution Minimum Anomaly Score GAN for Sample Generation on the Boundary
• Test Sample Accuracy Scales with Training Sample Density in Neural Networks
• Robust Out-of-Distribution Detection on Deep Probabilistic Generative Models
• Unsupervised Approaches for Out-Of-Distribution Dermoscopic Lesion Detection
• Why Out-of-distribution Detection in CNNs Does Not Like Mahalanobis -- and What to Use Instead
• FROB: Few-shot ROBust Model for Classification and Out-of-Distribution Detection
• Understanding the Role of Self-Supervised Learning in Out-of-Distribution Detection Task