Classification without labels: learning from mixed samples in high energy physics

By E. Metodiev, Benjamin Nachman, Jesse Thaler. Journal of High Energy Physics, 2017

234 citations recorded by the source corpus.

Modern machine learning techniques can be used to construct powerful models for difficult collider physics problems. In many applications, however, these models are trained on imperfect simulations due to a lack of truth-level information in the data, which risks the model learning artifacts of the simulation. In this paper, we introduce the paradigm of classification without labels (CWoLa) in which a classifier is trained to distinguish statistical mixtures of classes, which are common in collider physics. Crucially, neither individual labels nor class proportions are required, yet we prove that the optimal classifier in the CWoLa paradigm is also the optimal classifier in the traditional fully-supervised case where all label information is available. After demonstrating the power of this method in an analytical toy example, we consider a realistic benchmark for collider physics: distinguishing quark- versus gluon-initiated jets using mixed quark/gluon training samples. More generally, CWoLa can be applied to any classification problem where labels or class proportions are unknown or simulations are unreliable, but statistical mixtures of the classes are available.

Subfields: Nuclear and High Energy Physics

View Paper: https://www.semanticscholar.org/paper/528709d8c74e64305e83fc64349ece0d02efb3a7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Nuclear and High Energy Physics

Citations in the corpus

Jet substructure at the Large Hadron Collider (2018, background): “Interestingly, recent studies have shown how modern machine learning classifiers can be directly trained on data even though there are no per-jet labels [152, 153].”
Energy flow networks: deep sets for particle jets (2018, methodology): “It would be interesting to apply all of these methods in a fully-data driven way [88–90] to test”
JUNIPR: a framework for unsupervised machine learning in particle physics (2018, background): “Early studies of weakly supervised learning in particle physics show very promising results: performance comparable to fully supervised methods was found both with low-dimensional inputs [35,36] (a few physical observables) and with very high-dimensional inputs [37] (jet images).”
Machine learning and the physical sciences (2019, background): “Recently, weakly supervised techniques have been developed that only involve identifying regions where only the class proportions are known or assuming that the relative probabilities p(y|R) are not linearly dependent (Komiske et al., 2018a; Metodiev et al., 2017) .”
Jet substructure at the Large Hadron Collider: A review of recent advances in theory and machine learning (2017, methodology): “An alternative method that does not suffer from either of these considerations is the Classification without Labels (CWoLa) approach [570].”
The Lund jet plane (2018, background): “The top representation shows the set of particles in the jet, with the direction and length of each line segment schematically representing the direction and scalar momentum (1)Note that some approaches attempt to circumvent the use of modelling tools [44, 45].”