MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Space
arXiv:2105.01879
Abstract
Detecting out-of-distribution (OOD) inputs is a central challenge for safely deploying machine learning models in the real world. Existing solutions are mainly driven by small datasets, with low resolution and very few class labels (e.g., CIFAR). As a result, OOD detection for large-scale image classification tasks remains largely unexplored. In this paper, we bridge this critical gap by proposing a group-based OOD detection framework, along with a novel OOD scoring function termed MOS. Our key idea is to decompose the large semantic space into smaller groups with similar concepts, which allows simplifying the decision boundaries between in- vs. out-of-distribution data for effective OOD detection. Our method scales substantially better for high-dimensional class space than previous approaches. We evaluate models trained on ImageNet against four carefully curated OOD datasets, spanning diverse semantics. MOS establishes state-of-the-art performance, reducing the average FPR95 by 14.33% while achieving 6x speedup in inference compared to the previous best method.
arXiv commentPaper accepted as an oral presentation in CVPR'21
References in corpus (12)
- Auto-Encoding Variational Bayes
- Distilling the Knowledge in a Neural Network
- Conditional Image Generation with PixelCNN Decoders
- A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
- A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
- Deep Anomaly Detection with Outlier Exposure
- Outlier Exposure with Confidence Control for Out-of-Distribution Detection
- SelectiveNet: A Deep Neural Network with an Integrated Reject Option
- Discriminative out-of-distribution detection for semantic segmentation
- HD-CNN: Hierarchical Deep Convolutional Neural Network for Large Scale Visual Recognition
- Confidence estimation in Deep Neural networks via density modelling
- Self-informed neural network structure learning