Unsupervised Anomaly Detection with Rejection
Lists
Only you see this
Abstract
Anomaly detection aims at detecting unexpected behaviours in the data. Because anomaly detection is usually an unsupervised task, traditional anomaly detectors learn a decision boundary by employing heuristics based on intuitions, which are hard to verify in practice. This introduces some uncertainty, especially close to the decision boundary, that may reduce the user trust in the detector's predictions. A way to combat this is by allowing the detector to reject examples with high uncertainty (Learning to Reject). This requires employing a confidence metric that captures the distance to the decision boundary and setting a rejection threshold to reject low-confidence predictions. However, selecting a proper metric and setting the rejection threshold without labels are challenging tasks. In this paper, we solve these challenges by setting a constant rejection threshold on the stability metric computed by ExCeeD. Our insight relies on a theoretical analysis of such a metric. Moreover, setting a constant threshold results in strong guarantees: we estimate the test rejection rate, and derive a theoretical upper bound for both the rejection rate and the expected prediction cost. Experimentally, we show that our method outperforms some metric-based methods.
BibTeX
@inproceedings{perini2023unsupervised,
author = {Lorenzo Perini and Jesse Davis},
title = {{Unsupervised Anomaly Detection with Rejection}},
booktitle = {Advances in Neural Information Processing Systems},
year = {2023},
url = {https://openreview.net/forum?id=WK8LQzzHwW},
}| Classification with Rejection Based on Cost-sensitive Classification | arXiv · 2020 |
| Rejection via Learning Density Ratios | NeurIPS · 2024 |
| Machine learning with a reject option: a survey | Machine Learning · 2024 |
| CADET: Calibrated Anomaly Detection for Mitigating Hardness Bias | IJCAI · 2022 |
| A novel reject option applied to sleep stage scoring | SDM · 2023 |
| Deep Active Learning for Anomaly Detection | arXiv · 2018 |