Probabilistic Modeling of Deep Features for Out-of-Distribution and Adversarial Detection

Abstract

We present a principled approach for detecting out-of-distribution (OOD) and adversarial samples in deep neural networks. Our approach consists in modeling the outputs of the various layers (deep features) with parametric probability distributions once training is completed. At inference, the likelihoods of the deep features w.r.t the previously learnt distributions are calculated and used to derive uncertainty estimates that can discriminate in-distribution samples from OOD samples. We explore the use of two classes of multivariate distributions for modeling the deep features - Gaussian and Gaussian mixture - and study the trade-off between accuracy and computational complexity. We demonstrate benefits of our approach on image features by detecting OOD images and adversarially-generated images, using popular DNN architectures on MNIST and CIFAR10 datasets. We show that more precise modeling of the feature distributions result in significantly improved detection of OOD and adversarial samples; up to 12 percentage points in AUPR and AUROC metrics. We further show that our approach remains extremely effective when applied to video data and associated spatio-temporal features by detecting adversarial samples on activity classification tasks using UCF101 dataset, and the C3D network. To our knowledge, our methodology is the first one reported for reliably detecting white-box adversarial framing, a state-of-the-art adversarial attack for video classifiers.

References in corpus (2)

• UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
• EMNIST: an extension of MNIST to handwritten letters

Cited by in corpus (8)

• Why is the Mahalanobis Distance Effective for Anomaly Detection?
• Learning image representations for anomaly detection: application to discovery of histological alterations in drug development
• Do We Really Need to Learn Representations from In-domain Data for Outlier Detection?
• Deep Probabilistic Models to Detect Data Poisoning Attacks
• Multi-Loss Sub-Ensembles for Accurate Classification with Uncertainty Estimation
• Energy-Based Anomaly Detection and Localization
• Out-Of-Distribution Detection With Subspace Techniques And Probabilistic Modeling Of Features
• Mitigating Sampling Bias and Improving Robustness in Active Learning