QCD or what?

Abstract

Autoencoder networks, trained only on QCD jets, can be used to search for anomalies in jet-substructure. We show how, based either on images or on 4-vectors, they identify jets from decays of arbitrary heavy resonances. To control the backgrounds and the underlying systematics we can de-correlate the jet mass using an adversarial network. Such an adversarial autoencoder allows for a general and at the same time easily controllable search for new physics. Ideally, it can be trained and applied to data in the same phase space region, allowing us to efficiently search for new physics using un-supervised learning.

Mass agnostic jet taggers

“…However, there have been many studies of jet taggers using other representations, such as images, sequences, or graphs. Mass decorrelation has been done in images with Planing [4] and Adversarial training [32], but it would be interesting to see how all of the techniques studied here could be applied to the different representations, and if any additional advantage is offered. Additionally, decorrelating in both the jet mass and the transverse momentum could make for a stable jet tagger (See Ref.…”

Exploring the space of jets with CMS open data

“…In terms of quantitative measures, we showed how to use the EMD to characterize the impact of detector effects and to calculate the intrinsic dimension of a jet ensemble. For qualitative studies, we showed how to use the EMD to identify the most representative jets in a histogram bin and the least representative jets in the ensemble as a whole, where the latter analysis is particularly interesting in the context of anomaly detection for new physics searches [171][172][173][174][175][176][177].…”

Anomaly detection with density estimation

“…• More recently, a variety of approaches have been proposed, often relying on sophisticated deep learning techniques, that attempt to be both signal and background model agnostic, to varying degrees. These include approaches based on autoencoders [26][27][28][29][30][31], weak supervision [32,33], nearest neighbor algorithms [34][35][36], probabilistic modeling [37], and others [38]. These are indicated in the upper-right corner of Fig.…”

“…1(a), we have also attempted to illustrate in finer detail the differences between some recent model-agnostic approaches. For example, the autoencoder is in the farthest corner since it assumes almost nothing about the signal or the background but can be run directly on the data, as long as the signal is sufficiently rare [26,27]. The tradeoff is that there is no optimality guarantee for the autoencoder -any signals that it does find will be found in a rather uncontrolled manner.…”

“…Yet, they have varying degrees of both signal-model and background-model independence, as there is often a tradeoff between the broadness of a search and how sensitive it is to particular classes of signal scenarios. Existing and proposed model-agnostic searches range from fully-signal-model independent but fully-background model dependent [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25] (because they compare data to SM simulation); to varying degrees of partial signal-model and background-model independence [26][27][28][29][30][31][32][33][34][35][36][37][38]. A comprehensive overview of existing model-agnostic approaches and how they are classified in terms of signal and background model independence will be given in Section 2.…”