Problem
Research questions and friction points this paper is trying to address.
tabular anomaly detection
prior-data fitted networks
unsupervised learning
representation learning
Adapting prior-data fitted networks for tabular anomaly detection
🤖 AI Summary
This study addresses the lack of robust deep representations and the challenge of unsupervised fine-tuning in tabular anomaly detection by proposing the ZEN and FOCUS methods. The approach leverages Prior-Data Fitted Networks (PFNs) to extract tabular features, optimizing representational separability through nearest-neighbor distance scoring and reference-set fine-tuning, thereby enabling feature adaptation without anomaly labels. Experimental results demonstrate that the proposed methods achieve the highest average AUROC on the ADBench benchmark, outperforming all baselines while exhibiting strong generalization capabilities. This work establishes an efficient new paradigm for unsupervised representation learning in tabular anomaly detection.
📝 Abstract
While deep features have transformed anomaly detection in images and video, their impact on tabular data has been less substantial, partly due to the limited availability of strong deep representations. Recently, prior-data fitted networks (PFNs) have emerged as a promising source of such representations for tabular data. In this work, we investigate how PFN representations can be adapted and leveraged for anomaly detection. The question is harder than it looks. No anomalies are available before deploy- ment, so model parameters cannot be tuned with supervision, and the reference set that defines normal behavior may itself contain the very anomalies it is supposed to reveal. We begin our study using frozen TabPFN features. Scoring each sam- ple by its distance to its nearest neighbors in feature space already gives strong results. We identify which layers to use and a feature-extraction procedure suited to the task. Next, to further improve performance, we use the reference set to fine- tune the model, so that the resulting features better separate normal samples from anomalies. On the ADBench benchmark, our fine-tuning free approach (ZEN) reaches a higher mean AUROC than every baseline, and our fine-tuned method (FOCUS) improves on it further. Our approach also generalizes across PFN models.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Prior-data Fitted Networks
Tabular Anomaly Detection
Representation Learning
Unsupervised Fine-tuning
Feature Extraction
🔎 Similar Papers
No similar papers found.
