Generative Probabilistic Novelty Detection with Adversarial Autoencoders

Abstract

Novelty detection is the problem of identifying whether a new data point is considered to be an inlier or an outlier. We assume that training data is available to describe only the inlier distribution. Recent approaches primarily leverage deep encoder-decoder network architectures to compute a reconstruction error that is used to either compute a novelty score or to train a one-class classifier. While we too leverage a novel network of that kind, we take a probabilistic approach and effectively compute how likely is that a sample was generated by the inlier distribution. We achieve this with two main contributions. First, we make the computation of the novelty probability feasible because we linearize the parameterized manifold capturing the underlying structure of the inlier distribution, and show how the probability factorizes and can be computed with respect to local coordinates of the manifold tangent space. Second, we improved the training of the autoencoder network. An extensive set of results show that the approach achieves state-of-the-art results on several benchmark datasets.

References in corpus (15)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Auto-Encoding Variational Bayes
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
• Conditional Image Generation with PixelCNN Decoders
• Pixel Recurrent Neural Networks
• Autoencoding beyond pixels using a learned similarity metric
• One-Class Classification: Taxonomy of Study and Review of Techniques
• A geometric analysis of subspace clustering with outliers
• Learning Confidence for Out-of-Distribution Detection in Neural Networks
• TurkerGaze: Crowdsourcing Saliency with Webcam based Eye Tracking
• Robust PCA via Outlier Pursuit
• Coherence Pursuit: Fast, Simple, and Robust Principal Component Analysis
• PixelCNN++: Improving the PixelCNN with Discretized Logistic Mixture Likelihood and Other Modifications
• Learning Temporal Regularity in Video Sequences

Cited by in corpus (46)

• Generalized Out-of-Distribution Detection: A Survey
• CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances
• The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentation
• One-Class Classification: A Survey
• Out-of-distribution Detection in Classifiers via Generation
• Anomaly Detection based on Zero-Shot Outlier Synthesis and Hierarchical Feature Distillation
• Understanding Anomaly Detection with Deep Invertible Networks through Hierarchies of Distributions and Features
• Perceptual Image Anomaly Detection
• Revisiting Bayesian Autoencoders with MCMC
• Information bottleneck through variational glasses
• On Last-Layer Algorithms for Classification: Decoupling Representation from Uncertainty Estimation
• The Clever Hans Effect in Anomaly Detection
• Statistical process monitoring of artificial neural networks
• Unsupervised Out-of-Distribution Detection with Batch Normalization
• VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization
• GRAM: An Interpretable Approach for Graph Anomaly Detection using Gradient Attention Maps
• Normality-Calibrated Autoencoder for Unsupervised Anomaly Detection on Data Contamination
• Puzzle-AE: Novelty Detection in Images through Solving Puzzles
• Rethinking Assumptions in Deep Anomaly Detection
• Novelty Detection Via Blurring
• Improved Adversarial Robustness by Reducing Open Space Risk via Tent Activations
• This is not what I imagined: Error Detection for Semantic Segmentation through Visual Dissimilarity
• Open Set Recognition with Conditional Probabilistic Generative Models
• OneFlow: One-class flow for anomaly detection based on a minimal volume region
• Out-of-domain Detection for Natural Language Understanding in Dialog Systems
• Focus Your Distribution: Coarse-to-Fine Non-Contrastive Learning for Anomaly Detection and Localization
• Do We Really Need to Learn Representations from In-domain Data for Outlier Detection?
• Manifolds for Unsupervised Visual Anomaly Detection
• Unsupervised Pixel-level Road Defect Detection via Adversarial Image-to-Frequency Transform
• GRD-Net: Generative-Reconstructive-Discriminative Anomaly Detection with Region of Interest Attention Module
• Novelty Detection via Robust Variational Autoencoding
• Towards Open World Object Detection
• Deep Context-Aware Novelty Detection
• Deep Anomaly Detection by Residual Adaptation
• Fair Normalizing Flows
• Anomaly Detection via Self-organizing Map
• Discriminative Multi-level Reconstruction under Compact Latent Space for One-Class Novelty Detection
• Autoencoding Under Normalization Constraints
• MMF: A loss extension for feature learning in open set recognition
• Unsupervised Learning of Multi-level Structures for Anomaly Detection
• Multi-Class Anomaly Detection
• Flow-based SVDD for anomaly detection
• Anomaly Detection Based on Multiple-Hypothesis Autoencoder
• Active Learning for UAV-based Semantic Mapping
• Multi-class Novelty Detection Using Mix-up Technique
• OAAE: Adversarial Autoencoders for Novelty Detection in Multi-modal Normality Case via Orthogonalized Latent Space