Research map / Label-efficient and robust learning / OOD and anomaly detection
Multiple Class Novelty Detection Under Data Distribution Shift
Poojan Oza, Hien V. Nguyen, Vishal M. Patel
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| test on shifted USPS instead of MNIST | ALOCC · MNIST/USPS preliminary | AUROC: 0.81 -> 0.63 | Hurt |
| add gradient reversal layer domain alignment | Softmax · Digits novelty detection (average) | AUROC: 0.620 -> 0.776 | Improved |
| add gradient reversal and classification to ALOCC | ALOCC · Digits novelty detection (average) | AUROC: 0.667 -> 0.873 | Improved |
| replace ALOCC+GRL with cross-domain generator unified training | ALOCC+GRL · Digits novelty detection (average) | AUROC: 0.873 -> 0.921 | Improved |
| replace ALOCC+GRL with cross-domain generator unified training | ALOCC+GRL · Office-31 novelty detection (average) | AUROC: 0.774 -> 0.877 | Improved |
| train on unlabeled target data with ALOCC instead of source-only softmax | Softmax · Office-31 novelty detection (average) | AUROC: 0.737 -> 0.758 | Improved |
| add gradient reversal layer to ALOCC | ALOCC · Office-31 novelty detection (average) | AUROC: 0.758 -> 0.774 | Improved |
| add domain alignment to Softmax with GRL | Softmax · Office-31 novelty detection (average) | AUROC: 0.737 -> 0.762 | Improved |
| replace ALOCC+GRL with proposed method on W->D | ALOCC+GRL · Office-31 W->D | AUROC: 0.987 -> 0.938 | Hurt |
| replace ALOCC+GRL with proposed on SVHN->USPS | ALOCC+GRL · Digits SVHN->USPS | AUROC: 0.845 -> 0.895 | Improved |
| use target-only ALOCC instead of source-only softmax | Softmax · Digits novelty detection (average) | AUROC: 0.620 -> 0.667 | Improved |
Closest papers by meaning
- Adversarially Learned One-Class Classifier for Novelty Detection (CVPR 2018 · similarity 0.89)
- Adversarially Robust One-class Novelty Detection (TPAMI 2022 · similarity 0.89)
- Universal Novelty Detection Through Adaptive Contrastive Learning (CVPR 2024 · similarity 0.87)
- Contamination-Resilient Anomaly Detection via Adversarial Learning on Partially-Observed Normal and Anomalous Data (ICML 2024 · similarity 0.84)
- ResAD: A Simple Framework for Class Generalizable Anomaly Detection (NeurIPS 2024 · similarity 0.84)
- Hierarchical Novelty Detection via Fine-Grained Evidence Allocation (ICML 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
