Speaker: Mihoko Nojiri

Title: Morphology for Jet physics

The Minkowski functional(MF) is commonly used technique in astrophysics. In this paper, I introduce a new application of MF to jet physics. Tagging boosted objects is important to identify new physics at HL-LHC. The techniques of deep learning have been actively applied to the problem. In this paper we introduce new NN(neural network) model using a morphological information with the Minkowski functionals (MFs). The MFs provide geometric information of the jet constituents, and they are independent of the IRC(Infrared collinear) safe observables commonly used in jet physics. We emphasize a potential relationship between the morphology and the convolutional neural network.The new classifier is computationally cheap because geometrical information of jet image is efficiently compressed in a few numbers, and the tagging performance is comparable to the existing jet image classifiers. The result suggests that the our model gives efficient decomposition of the feature space for jet classification.

トップ   編集 凍結 差分 バックアップ 添付 複製 名前変更 リロード   新規 一覧 単語検索 最終更新   ヘルプ   最終更新のRSS
Last-modified: 2020-11-19 (木) 16:29:13 (62d)