Supervised Deep Learning in High Energy Phenomenology: a Mini Review*

Abstract

Deep learning, a branch of machine learning, have been recently applied to high energy experimental and phenomenological studies. In this note we give a brief review on those applications using supervised deep learning. We first describe various learning models and then recapitulate their applications to high energy phenomenological studies. Some detailed applications are delineated in details, including the machine learning scan in the analysis of new physics parameter space, the graph neural networks in the search of top-squark production and in the CP measurement of the top-Higgs coupling at the LHC.

Testing electroweak SUSY for muon g − 2 and dark matter at the LHC and beyond

“…On the other hand, it should be mentioned that the heavy higgsinos decaying to light bino in the BHL scenario will provide 3 + / E T signature at a 100 TeV hadron collider, which can exclude the higgsino mass up to about 3 TeV at 95% C.L.. Besides conventional cut-based analysis, the machine learning methods have been recently proposed to enhance the sensitivity in the search of sparticles at the LHC [65][66][67][68][69]. We expect that our result may be improved by using those advanced analysis approaches.…”

Neural network-based top tagger with two-point energy correlations and geometry of soft emissions

“…Interest in deep learning in collider physics [1][2][3][4][5] has been growing in recent years. Many applications of deep learning have appeared in jet classification , anomaly detection [27][28][29][30][31][32][33][34][35][36][37], particle identification [38][39][40], pileup mitigation [41][42][43], event generation [44][45][46][47][48][49][50][51][52][53][54][55][56][57][58], unfolding [59,60], and parton distribution functions [61][62][63][64][65][66][67][68][69][70][71]…”

“…The convergence is defined in terms of the Hausdorff metric 5. For the rectangle shape pixels, the term proportional to 4χ (0) corresponds to the number of pixels that touch only the corner of the pixels in P (0) .…”

LFV and (g-2) in non-universal SUSY models with light higgsinos

“…Here we will analyse supersymmetric non-universal Higgs models with an additional parameter for the third generation of scalar superpartners (NUHM3) [43][44][45]. We focus on the so-called light higgsino-world scenario [46][47][48][49][50][51][52][53][54][55] in which the SUSY matter scalars are pushed into the multi-TeV scale while µ 1 TeV, as natural SUSY requires. We observe that a characteristic spectrum where the lightest neutralinos and chargino are higgsino-like is automatically selected when the correct (g −2) µ phenomenology is required through a general scan that includes solutions with the usual gaugino-higgsino hierarchy.…”