FoMo X: Modular Explainability Signals for Outlier Detection Foundation Models

2026-03-18

AI Summary

Existing foundation models for tabular anomaly detection lack interpretability, hindering their applicability in safety-critical decision-making scenarios. To address this limitation, this work proposes the FoMo-X framework, which introduces a lightweight diagnostic head atop a frozen PFN backbone, thereby embedding modular diagnostic signals—such as anomaly severity grading and uncertainty quantification—directly into the zero-shot anomaly detection pipeline for the first time. Leveraging generative simulation priors during training, FoMo-X distills computationally expensive uncertainty estimation techniques like Monte Carlo Dropout into a single forward pass. This approach efficiently reproduces authentic diagnostic signals on benchmarks such as ADBench with negligible inference overhead, significantly enhancing the trustworthiness, real-time performance, and operational utility of zero-shot anomaly detection.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Tabular foundation models, specifically Prior-Data Fitted Networks (PFNs), have revolutionized outlier detection (OD) by enabling unsupervised zero-shot adaptation to new datasets without training. However, despite their predictive power, these models typically function as opaque black boxes, outputting scalar outlier scores that lack the operational context required for safety-critical decision-making. Existing post-hoc explanation methods are often computationally prohibitive for real-time deployment or fail to capture the epistemic uncertainty inherent in zero-shot inference. In this work, we introduce FoMo-X, a modular framework that equips OD foundation models with intrinsic, lightweight diagnostic capabilities. We leverage the insight that the frozen embeddings of a pretrained PFN backbone already encode rich, context-conditioned relational information. FoMo-X attaches auxiliary diagnostic heads to these embeddings, trained offline using the same generative simulator prior as the backbone. This allows us to distill computationally expensive properties, such as Monte Carlo dropout based epistemic uncertainty, into a deterministic, single-pass inference. We instantiate FoMo-X with two novel heads: a Severity Head that discretizes deviations into interpretable risk tiers, and an Uncertainty Head that provides calibrated confidence measures. Extensive evaluation on synthetic and real-world benchmarks (ADBench) demonstrates that FoMo-X recovers ground-truth diagnostic signals with high fidelity and negligible inference overhead. By bridging the gap between foundation model performance and operational explainability, FoMo-X offers a scalable path toward trustworthy, zero-shot outlier detection.

Problem

• outlier detection
• foundation models
• explainability
• epistemic uncertainty
• zero-shot inference

Innovation

• modular explainability
• foundation models
• zero-shot outlier detection
• epistemic uncertainty
• diagnostic heads