Out-of-distribution Detection in Classifiers via Generation

Abstract

By design, discriminatively trained neural network classifiers produce reliable predictions only for in-distribution samples. For their real-world deployments, detecting out-of-distribution (OOD) samples is essential. Assuming OOD to be outside the closed boundary of in-distribution, typical neural classifiers do not contain the knowledge of this boundary for OOD detection during inference. There have been recent approaches to instill this knowledge in classifiers by explicitly training the classifier with OOD samples close to the in-distribution boundary. However, these generated samples fail to cover the entire in-distribution boundary effectively, thereby resulting in a sub-optimal OOD detector. In this paper, we analyze the feasibility of such approaches by investigating the complexity of producing such "effective" OOD samples. We also propose a novel algorithm to generate such samples using a manifold learning network (e.g., variational autoencoder) and then train an n+1 classifier for OOD detection, where the $n+1^{th}$ class represents the OOD samples. We compare our approach against several recent classifier-based OOD detectors on MNIST and Fashion-MNIST datasets. Overall the proposed approach consistently performs better than the others.

arXiv commentNeurIPS 2019, Safety and Robustness in Decision Making Workshop

References in corpus (12)

• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• ADADELTA: An Adaptive Learning Rate Method
• A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
• Deep Bayesian Active Learning with Image Data
• EMNIST: an extension of MNIST to handwritten letters
• Deep Anomaly Detection with Outlier Exposure
• Learning Confidence for Out-of-Distribution Detection in Neural Networks
• Natural Adversarial Examples
• Generative Probabilistic Novelty Detection with Adversarial Autoencoders
• Safer Classification by Synthesis
• Building robust classifiers through generation of confident out of distribution examples
• Analysis of Confident-Classifiers for Out-of-distribution Detection

Cited by in corpus (11)

• Uncertainty-Aware Reliable Text Classification
• Detecting Out-of-Distribution Examples with In-distribution Examples and Gram Matrices
• Embedded out-of-distribution detection on an autonomous robot platform
• Detecting Backdoors in Neural Networks Using Novel Feature-Based Anomaly Detection
• A Unified Benchmark for the Unknown Detection Capability of Deep Neural Networks
• Hybrid Models for Open Set Recognition
• Distribution-aware Fairness Test Generation
• BaCOUn: Bayesian Classifers with Out-of-Distribution Uncertainty
• Bayesian OOD detection with aleatoric uncertainty and outlier exposure
• Rainproof: An Umbrella To Shield Text Generators From Out-Of-Distribution Data
• FROB: Few-shot ROBust Model for Classification and Out-of-Distribution Detection