Semi-supervised anomaly detection – towards model-independent searches of new physics

Abstract

Most classification algorithms used in high energy physics fall under the category of supervised machine learning. Such methods require a training set containing both signal and background events and are prone to classification errors should this training data be systematically inaccurate for example due to the assumed MC model. To complement such model-dependent searches, we propose an algorithm based on semi-supervised anomaly detection techniques, which does not require a MC training sample for the signal data. We first model the background using a multivariate Gaussian mixture model. We then search for deviations from this model by fitting to the observations a mixture of the background model and a number of additional Gaussians. This allows us to perform pattern recognition of any anomalous excess over the background. We show by a comparison to neural network classifiers that such an approach is a lot more robust against misspecification of the signal MC than supervised classification. In cases where there is an unexpected signal, a neural network might fail to correctly identify it, while anomaly detection does not suffer from such a limitation. On the other hand, when there are no systematic errors in the training data, both methods perform comparably. arXiv:1112.3329v3 [physics.data-an]

Learning new physics from a machine

“…Most applications of neural networks to new physics searches aim at enhancing the sensitivity to prespecified models of the resonant or nonresonant type. Using machine learning techniques for model-independent new physics searches has been proposed in [64]; however Gaussian mixture models are employed rather than neural networks and the overall strategy is quite different from ours. Reference [47] uses neural networks, but with the purpose of enhancing the sensitivity to resonant bumps that emerge in a prespecified kinematical variable.…”

Novelty detection meets collider physics

“…Some preliminary and at least partially related efforts have been made at jet [19,20] and event [21][22][23][24][25][26] level. For novelty detection with given feature representation, its sensitivity depends crucially on the performance of novelty evaluators.…”

Guiding new physics searches with unsupervised learning

“…Refs. [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27] and in particular the recent work of Ref. [8] with which we share some ideas, although with a very different implementation).…”

“…[8] with which we share some ideas, although with a very different implementation). On the other hand, applications of unsupervised learning have been relatively unexplored [10,28,29]. In unsupervised learning the data are not labeled, so the presence and the characteristics of new phenomena in the data are not known a priori.…”