ICLAD enables unified, in-context anomaly detection for tabular data across unsupervised, semi-supervised, and one-class regimes without weight updates.
March 23, 2026
Original Paper
ICLAD: In-Context Learning for Unified Tabular Anomaly Detection Across Supervision Regimes
arXiv · 2603.19497
The Takeaway
Most tabular anomaly detection requires retraining for every new dataset and specific supervision level. This provides a 'foundation model' approach for tabular data that practitioners can apply zero-shot to diverse anomaly detection tasks.
From the abstract
Anomaly detection on tabular data is commonly studied under three supervision regimes, including one-class settings that assume access to anomaly-free training samples, fully unsupervised settings with unlabeled and potentially contaminated training data, and semi-supervised settings with limited anomaly labels. Existing deep learning approaches typically train dataset-specific models under the assumption of a single supervision regime, which limits their ability to leverage shared structures ac