Related papers: Improved jet substructure methods: Y-splitter and …
We introduce a novel approach to jet tagging and classification through the use of techniques inspired by computer vision. Drawing parallels to the problem of facial recognition in images, we define a jet-image using calorimeter towers as…
In this work, we present a new technique to measure the longitudinal and transverse polarization fractions of hadronic decays of boosted $W$ bosons. We introduce a new jet substructure observable denoted as $p_\theta$, which is a proxy…
Machine learning methods incorporating deep neural networks have been the subject of recent proposals for new hadronic resonance taggers. These methods require training on a dataset produced by an event generator where the true class labels…
In time for the first tests on LHC data we introduce a set of improvements and tests of purely kinematic top tagging algorithms. First, we show how different jet algorithms can be used for different transverse momentum regimes. Combining…
Spectral clustering and co-clustering are well-known techniques in data analysis, and recent work has extended spectral clustering to square, symmetric tensors and hypermatrices derived from a network. We develop a new tensor spectral…
We study the complexity of cutting planes and branching schemes from a theoretical point of view. We give some rigorous underpinnings to the empirically observed phenomenon that combining cutting planes and branching into a branch-and-cut…
In searches for new physics in the energy regime of the LHC, it is becoming increasingly important to distinguish single-jet objects that originate from the merging of the decay products of W bosons produced with high transverse momenta…
This paper describes an innovative way to optimize a multivariate classifier, in particular a Support Vector Machine algorithm, on a problem characterized by a biased training sample. This is possible thanks to the feedback of a…
We present three case studies at a 100 TeV proton collider for how jet analyses can be improved using new jet (sub)structure techniques. First, we use the winner-take-all recombination scheme to define a recoil-free jet axis that is robust…
In these proceedings, we report on recent results related to vector boson-tagged jet production in heavy ion collisions and the related modification of jet substructure, such as jet shapes and jet momentum sharing distributions.…
Invariant mass spectra for jets reconstructed using the anti-kt and Cambridge-Aachen algorithms are studied for different jet "grooming" techniques in data corresponding to an integrated luminosity of 5 inverse femtobarns, recorded with the…
Building on the notion of a particle physics detector as a camera and the collimated streams of high energy particles, or jets, it measures as an image, we investigate the potential of machine learning techniques based on deep learning…
Trimming is a core technique in geometric modeling. Unfortunately, the resulting objects do not take the requirements of numerical simulations into account and yield various problems. This paper outlines principal issues of trimmed models…
We present a consistent way of combining associated weak boson radiation in hard dijet events with hard QCD radiation in Drell-Yan-like scatterings. This integrates multiple tree-level calculations with vastly different cross sections, QCD-…
The latest progress in calculating electroweak gauge boson production in association with QCD jets at hadron colliders is summarized. Particular emphasis is given to the recently completed QCD one-loop calculations of W+3jets and Wb final…
Structured pruning is an effective approach for compressing large pre-trained neural networks without significantly affecting their performance. However, most current structured pruning methods do not provide any performance guarantees, and…
Automatic segmentation of an image to identify all meaningful parts is one of the most challenging as well as useful tasks in a number of application areas. This is widely studied. Selective segmentation, less studied, aims to use limited…
Gradient boosted decision trees are some of the most popular algorithms in applied machine learning. They are a flexible and powerful tool that can robustly fit to any tabular dataset in a scalable and computationally efficient way. One of…
The success of many machine learning and pattern recognition methods relies heavily upon the identification of an appropriate distance metric on the input data. It is often beneficial to learn such a metric from the input training data,…
Grooming techniques, such as soft drop, play a central role in reducing sensitivity of jets to e.g. underlying event and hadronization at current collider experiments. The momentum sharing fraction $z_g$, of the two branches in a jet that…