Anomaly preserving contrastive neural embeddings for end-to-end model-independent searches at the LHC

TL;DR

This paper introduces contrastive neural embeddings for end-to-end, model-independent anomaly detection at the LHC, improving discovery power for rare signals by preserving anomalies in learned low-dimensional representations.

Contribution

It develops a contrastive learning framework for feature extraction that enhances anomaly detection in high-energy physics data, combining supervised and self-supervised methods with domain knowledge.

Findings

Significant improvement in detection sensitivity for rare physics signals.

Optimal embeddings balance background structure preservation and anomaly enhancement.

Combining compression with domain knowledge yields the best anomaly detection performance.

Abstract

Anomaly detection - identifying deviations from Standard Model predictions - is a key challenge at the Large Hadron Collider due to the size and complexity of its datasets. This is typically addressed by transforming high-dimensional detector data into lower-dimensional, physically meaningful features. We tackle feature extraction for anomaly detection by learning powerful low-dimensional representations via contrastive neural embeddings. This approach preserves potential anomalies indicative of new physics and enables rare signal extraction using novel machine learning-based statistical methods for signal-independent hypothesis testing. We compare supervised and self-supervised contrastive learning methods, for both MLP- and Transformer-based neural embeddings, trained on the kinematic observables of physics objects in LHC collision events. The learned embeddings serve as input representations for signal-agnostic statistical detection methods in inclusive final states. We achieve significant improvement in discovery power for both rare new physics signals and rare Standard Model processes across diverse final states, demonstrating its applicability for efficiently searching for diverse signals simultaneously. We study the impact of architectural choices, contrastive loss formulations, supervision levels, and embedding dimensionality on anomaly detection performance. We show that the optimal representation for background classification does not always maximize sensitivity to new physics signals, revealing an inherent trade-off between background structure preservation and anomaly enhancement. We demonstrate that combining compression with domain knowledge for label encoding produces the most effective data representation for statistical discovery of anomalies.