Gradient-Based Novelty Detection Boosted by Self-Supervised Binary Classification

Keyphrases

• Novelty Detection 100%
• Gradient-based 100%
• Out-of-distribution Data 100%
• Supervised Binary Classification 100%
• Continual Learning 50%
• Mahalanobis Distance 50%
• Behavior Analysis 25%
• Multivariate Analysis 25%
• Receiver Operating Characteristic 25%
• Automatically Identify 25%
• Binary Classification 25%
• Data Monitoring 25%
• Unsupervised Method 25%
• Multiple Datasets 25%
• Supervised Methods 25%
• Anomalous Data 25%
• Monitoring Behavior 25%
• Supervised Approach 25%
• CIFAR-10 25%
• CIFAR-100 25%
• Distribution Class 25%
• Label Selection 25%
• In-distribution 25%
• Out-of-distribution Detection 25%
• Operating Area 25%
• Area Under the Precision-Recall Curve 25%

Computer Science

• Binary Classification 100%
• Data Distribution 100%
• Mahalanobis Distance 50%
• Unsupervised Method 25%
• Supervised Method 25%
• binary classifier 25%
• Monitoring Data 25%
• Conventional Method 25%
• Behavior Analysis 25%
• Class Distribution 25%