The Extreme Value Machine

Abstract

It is often desirable to be able to recognize when inputs to a recognition function learned in a supervised manner correspond to classes unseen at training time. With this ability, new class labels could be assigned to these inputs by a human operator, allowing them to be incorporated into the recognition function --- ideally under an efficient incremental update mechanism. While good algorithms that assume inputs from a fixed set of classes exist, e.g., artificial neural networks and kernel machines, it is not immediately obvious how to extend them to perform incremental learning in the presence of unknown query classes. Existing algorithms take little to no distributional information into account when learning recognition functions and lack a strong theoretical foundation. We address this gap by formulating a novel, theoretically sound classifier --- the Extreme Value Machine (EVM). The EVM has a well-grounded interpretation derived from statistical Extreme Value Theory (EVT), and is the first classifier to be able to perform nonlinear kernel-free variable bandwidth incremental learning. Compared to other classifiers in the same deep network derived feature space, the EVM is accurate and efficient on an established benchmark partition of the ImageNet dataset.

Pre-print of a manuscript accepted to the IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI) journal

References in corpus (1)

• Towards Open World Recognition

Cited by in corpus (50)

• Recent Advances in Open Set Recognition: A Survey
• Adversarial Reciprocal Points Learning for Open Set Recognition
• Visual Identification of Individual Holstein-Friesian Cattle via Deep Metric Learning
• Data-Driven Fault Diagnosis Analysis and Open-Set Classification of Time-Series Data
• Appropriateness of Performance Indices for Imbalanced Data Classification: An Analysis
• Collective Decision of One-vs-Rest Networks for Open Set Recognition
• Collective decision for open set recognition
• A Survey on Open Set Recognition
• Memory Based Online Learning of Deep Representations from Video Streams
• Open-Set Support Vector Machines
• Entropic Out-of-Distribution Detection: Seamless Detection of Unknown Examples
• Open World Object Detection: A Survey
• Incremental Deep Neural Network Learning using Classification Confidence Thresholding
• Entropic Out-of-Distribution Detection
• Vocabulary-informed Zero-shot and Open-set Learning
• A modeler's guide to extreme value software
• Deep CNN-based Multi-task Learning for Open-Set Recognition
• Automatic Open-World Reliability Assessment
• EMG Signal Classification Using Reflection Coefficients and Extreme Value Machine
• SMART-Vision: Survey of Modern Action Recognition Techniques in Vision
• Detecting Backdoors in Neural Networks Using Novel Feature-Based Anomaly Detection
• One-vs-Rest Network-based Deep Probability Model for Open Set Recognition
• Extreme Value Theory for Open Set Classification -- GPD and GEV Classifiers
• Self-Supervised Features Improve Open-World Learning
• Unsupervised Out-of-Distribution Detection by Maximum Classifier Discrepancy
• An Unbiased Risk Estimator for Learning with Augmented Classes
• Classification-Reconstruction Learning for Open-Set Recognition
• uPredict: A User-Level Profiler-Based Predictive Framework for Single VM Applications in Multi-Tenant Clouds
• A Target-Agnostic Attack on Deep Models: Exploiting Security Vulnerabilities of Transfer Learning
• Open-set Face Recognition using Ensembles trained on Clustered Data
• Continual egocentric object recognition
• A Novel Open Set Energy-based Flow Classifier for Network Intrusion Detection
• Deep Learning and Traffic Classification: Lessons learned from a commercial-grade dataset with hundreds of encrypted and zero-day applications
• AI, Meet Human: Learning Paradigms for Hybrid Decision Making Systems
• Vocabulary-informed Extreme Value Learning
• Evaluation of Various Open-Set Medical Imaging Tasks with Deep Neural Networks
• Thinkback: Task-SpecificOut-of-Distribution Detection
• Handwriting Recognition with Novelty
• Learning without Seeing nor Knowing: Towards Open Zero-Shot Learning
• torchosr -- a PyTorch extension package for Open Set Recognition models evaluation in Python
• Unsupervised detection and open-set classification of fast-ramped flexibility activation events
• Open-Set Face Identification on Few-Shot Gallery by Fine-Tuning
• TFW, DamnGina, Juvie, and Hotsie-Totsie: On the Linguistic and Social Aspects of Internet Slang
• One-Class Meta-Learning: Towards Generalizable Few-Shot Open-Set Classification
• Integrating Propositional and Relational Label Side Information for Hierarchical Zero-Shot Image Classification
• Incrementally Zero-Shot Detection by an Extreme Value Analyzer
• Recognition Awareness: An Application of Latent Cognizance to Open-Set Recognition
• Model Rectification via Unknown Unknowns Extraction from Deployment Samples
• Why Out-of-distribution Detection in CNNs Does Not Like Mahalanobis -- and What to Use Instead
• A novel network training approach for open set image recognition