Universal New Physics Latent Space

Abstract

We develop a machine learning method for mapping data originating from both Standard Model processes and various theories beyond the Standard Model into a unified representation (latent) space while conserving information about the relationship between the underlying theories. We apply our method to three examples of new physics at the LHC of increasing complexity, showing that models can be clustered according to their LHC phenomenology: different models are mapped to distinct regions in latent space, while indistinguishable models are mapped to the same region. This opens interesting new avenues on several fronts, such as model discrimination, selection of representative benchmark scenarios, and identifying gaps in the coverage of model space.

arXiv comment25 pages, 17 figures

References in corpus (16)

• The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations
• DELPHES 3, A modular framework for fast simulation of a generic collider experiment
• Parton distributions with LHC data
• The automation of next-to-leading order electroweak calculations
• Supersymmetry and the LHC Inverse Problem
• The Dark Machines Anomaly Score Challenge: Benchmark Data and Model Independent Event Classification for the Large Hadron Collider
• Reinterpretation of LHC Results for New Physics: Status and Recommendations after Run 2
• The BSM-AI project: SUSY-AI - Generalizing LHC limits on Supersymmetry with Machine Learning
• Machine Learning for Anomaly Detection in Particle Physics
• Deep learning in the heterotic orbifold landscape
• The Hidden Geometry of Particle Collisions
• The Footprint of F-theory at the LHC
• Unravelling Strings at the LHC
• Neural Embedding: Learning the Embedding of the Manifold of Physics Data
• Mitigation of the LHC Inverse Problem
• Probing intractable beyond-standard-model parameter spaces armed with Machine Learning