IRC-safe Graph Autoencoder for unsupervised anomaly detection

Abstract

Anomaly detection through employing machine learning techniques has emerged as a novel powerful tool in the search for new physics beyond the Standard Model. Historically similar to the development of jet observables, theoretical consistency has not always assumed a central role in the fast development of algorithms and neural network architectures. In this work, we construct an infrared and collinear safe autoencoder based on graph neural networks by employing energy-weighted message passing. We demonstrate that whilst this approach has theoretically favourable properties, it also exhibits formidable sensitivity to non-QCD structures.

arXiv comment16 pages, 5 figures, Matched with the published version

References in corpus (27)

• The anti-k_t jet clustering algorithm
• An Introduction to PYTHIA 8.2
• A Brief Introduction to PYTHIA 8.1
• DELPHES 3, A modular framework for fast simulation of a generic collider experiment
• Anomaly Detection with Density Estimation
• Searching for New Physics with Deep Autoencoders
• Anomaly Detection for Resonant New Physics with Machine Learning
• Deep-learned Top Tagging with a Lorentz Layer
• QCD or What?
• Novelty Detection Meets Collider Physics
• Finding physics signals with shower deconstruction
• Guiding New Physics Searches with Unsupervised Learning
• Better Latent Spaces for Better Autoencoders
• Finding physics signals with event deconstruction
• A robust anomaly finder based on autoencoders
• Next-to-Leading Order Gluonic Three Jet Production at Hadron Colliders
• A strategy for a general search for new phenomena using data-derived signal regions and its application within the ATLAS experiment
• Understanding Jet Scaling and Jet Vetos in Higgs Searches
• Equivariant Energy Flow Networks for Jet Tagging
• Jets plus Missing Energy with an Autofocus
• Approaching robust EFT limits for CP-violation in the Higgs sector
• ATLAS Violating CP Effectively
• Matching Matters!
• Landscaping CP-violating BSM scenarios
• Effective field theory and scalar extensions of the top quark sector
• Universal Function Approximation on Graphs
• Anomaly Detection for Physics Analysis and Less than Supervised Learning

Cited by in corpus (4)

• Lorentz group equivariant autoencoders
• Jet tagging algorithm of graph network with HaarPooling message passing
• Anomalies, Representations, and Self-Supervision
• Comparing Point Cloud Strategies for Collider Event Classification