Latent Cluster Separation as a Regularizer for Lifelong Anomaly Detection
Paper · https://arxiv.org/abs/2611.03652
arXiv:2611.03652v1 Announce Type: new Abstract: Lifelong (continual) anomaly detection systems must learn a sequence of tasks without forgetting how to recognize anomaly types seen earlier -- a property known as backward transfer. A common way to try to improve backward transfer is to reshape a model's internal representation around learnable concept centroids, typically coupled with a new way of scoring anomalies based on distance to those centroids. Because these two design choices are usually introduced together, it is unclear which one is actually responsible for any performance gain. In this paper, we unbundle the two mechanisms within a VAE-based lifelong anomaly detector and study them separately. We find that used purely as a training-time regularizer -- while leaving anomaly scoring unchanged -- latent cluster separation can improve detection quality and backward transfer, though whether it outperforms the alternative scoring approach depends on the dataset's underlying representation. Switching to distance-based scoring is instead where most of the instability commonly associated with this class of method actually originates; we identify and explain the underlying mechanism. This benefit depends on the dataset's underlying representation, and we report this as practical guidance for choosing between the two design choices, rather than as a list of caveats.
First seen by Oriel on .
AI summary
This paper separates latent cluster regularization from distance-based anomaly scoring in a VAE-based lifelong detector, finding that regularization improves detection and backward transfer more reliably.
- The study examines lifelong anomaly detection, where models learn sequential tasks while retaining recognition of previously seen anomaly types.
- Latent cluster separation improves detection quality and backward transfer when used only during training, without changing anomaly scoring.
- Distance-based scoring causes most of the instability associated with this method, and the paper identifies its underlying mechanism.
- Whether distance-based scoring outperforms regular scoring depends on the dataset’s underlying representation.
- The findings provide practical guidance for choosing between latent regularization and alternative anomaly scoring designs.
#anomaly detection #continual learning #variational autoencoders #representation learning
Generated by gpt-5.6-sol. Check the original for details.
Aliases
| Source | External ID | URL |
|---|---|---|
| arXiv AI Papers | oai:arXiv.org:2611.03652v1 | https://arxiv.org/abs/2611.03652 Read discussion |