CURTAINs for your sliding window: Constructing unobserved regions by transforming adjacent intervals

Abstract

We propose a new model independent technique for constructing background data templates for use in searches for new physics processes at the LHC. This method, called Curtains, uses invertible neural networks to parameterise the distribution of side band data as a function of the resonant observable. The network learns a transformation to map any data point from its value of the resonant observable to another chosen value. Using Curtains, a template for the background data in the signal window is constructed by mapping the data from the side-bands into the signal region. We perform anomaly detection using the Curtains background template to enhance the sensitivity to new physics in a bump hunt. We demonstrate its performance in a sliding window search across a wide range of mass values. Using the LHC Olympics dataset, we demonstrate that Curtains matches the performance of other leading approaches which aim to improve the sensitivity of bump hunts, can be trained on a much smaller range of the invariant mass, and is fully data driven.

$\nu$-flows: Conditional neutrino regression

“…Our method for #» p ν likelihood estimation, called ν-Flows, is built using cINNs. These types of networks have already been used in collider physics, with notable applications including event generation [24], anomaly detection [25][26][27], density estimation [28], detector unfolding [29], and detector simulation [30,31].…”

Anomaly detection in the presence of irrelevant features

“…where m ≡ m JJ ; m L and m R are the lower and upper boundaries of the signal region in m JJ ; and the vector ⃗ x includes both auxiliary (relevant) and irrelevant features. While more elaborate methods of interpolation exist [8,11,32], this simple form is chosen here for the following reasons:…”

“…Even if it were possible in principle, the dependence of collider analyses on complex simulations to interpret measured signals would make such an approach extremely JHEP02(2024)220 sensitive to mismodelling errors at all stages of the simulation chain. Instead, recent work focuses on a simpler task of anomaly detection when localized with respect to a particular variable [3][4][5][6][7][8][9][10][11][12][13]. A well-studied benchmark example, starting with the work of [3,6], is a search for a dijet resonance, in which the signal jets are produced by a boosted resonance decay which is imprinted in non-trivial jet substructure.…”

How to understand limitations of generative networks

“…These networks should be trained on first-principle simulations, easy to handle, efficient to ship, powerful in amplifying the training samples [49,50], and -most importantly -precise. Going beyond forward generation, conditional generative networks can also be applied to probabilistic unfolding [51][52][53][54][55][56], inference [57,58], or anomaly detection [59][60][61][62][63][64], reinforcing the precision requirements.…”