(Machine) Learning to Do More with Less

Abstract

Determining the best method for training a machine learning algorithm is critical to maximizing its ability to classify data. In this paper, we compare the standard "fully supervised" approach (that relies on knowledge of event-by-event truth-level labels) with a recent proposal that instead utilizes class ratios as the only discriminating information provided during training. This so-called "weakly supervised" technique has access to less information than the fully supervised method and yet is still able to yield impressive discriminating power. In addition, weak supervision seems particularly well suited to particle physics since quantum mechanics is incompatible with the notion of mapping an individual event onto any single Feynman diagram. We examine the technique in detail -- both analytically and numerically -- with a focus on the robustness to issues of mischaracterizing the training samples. Weakly supervised networks turn out to be remarkably insensitive to systematic mismodeling. Furthermore, we demonstrate that the event level outputs for weakly versus fully supervised networks are probing different kinematics, even though the numerical quality metrics are essentially identical. This implies that it should be possible to improve the overall classification ability by combining the output from the two types of networks. For concreteness, we apply this technology to a signature of beyond the Standard Model physics to demonstrate that all these impressive features continue to hold in a scenario of relevance to the LHC.

arXiv comment32 pages, 12 figures. Example code is provided at https://github.com/bostdiek/PublicWeaklySupervised . v3: Version published in JHEP, discussion added

References in corpus (25)

• The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations
• The anti-k_t jet clustering algorithm
• FastJet user manual
• Popular Ensemble Methods: An Empirical Study
• Searching for Exotic Particles in High-Energy Physics with Deep Learning
• Echoes of a Hidden Valley at Hadron Colliders
• Performance of $b$-Jet Identification in the ATLAS Experiment
• Classification without labels: Learning from mixed samples in high energy physics
• Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics Synthesis
• Deep learning in color: towards automated quark/gluon jet discrimination
• Jet-Images -- Deep Learning Edition
• Deep-learning Top Taggers or The End of QCD?
• Parameterized Machine Learning for High-Energy Physics
• Macroscopic Strings and "Quirks" at Colliders
• Jet-Images: Computer Vision Inspired Techniques for Jet Tagging
• Weakly Supervised Classification in High Energy Physics
• How Much Information is in a Jet?
• Triggering Soft Bombs at the LHC
• The BSM-AI project: SUSY-AI - Generalizing LHC limits on Supersymmetry with Machine Learning
• Jets and Photons
• PhysicsGP: A Genetic Programming Approach to Event Selection
• Fast supersymmetry phenomenology at the Large Hadron Collider using machine learning techniques
• Dissecting Jets and Missing Energy Searches Using $n$-body Extended Simplified Models
• SCYNet: Testing supersymmetric models at the LHC with neural networks
• Tagging new physics with charm

Cited by in corpus (53)

• Jet Substructure at the Large Hadron Collider: A Review of Recent Advances in Theory and Machine Learning
• Searching for New Physics with Deep Autoencoders
• Searching for long-lived particles beyond the Standard Model at the Large Hadron Collider
• Energy Flow Networks: Deep Sets for Particle Jets
• The Machine Learning Landscape of Top Taggers
• Learning New Physics from a Machine
• Extending the Bump Hunt with Machine Learning
• QCD or What?
• Novelty Detection Meets Collider Physics
• Adversarially-trained autoencoders for robust unsupervised new physics searches
• Dijet resonance search with weak supervision using $\sqrt{s}=13$ TeV $pp$ collisions in the ATLAS detector
• Pulling Out All the Tops with Computer Vision and Deep Learning
• Guiding New Physics Searches with Unsupervised Learning
• The Dark Machines Anomaly Score Challenge: Benchmark Data and Model Independent Event Classification for the Large Hadron Collider
• JUNIPR: a Framework for Unsupervised Machine Learning in Particle Physics
• Uncovering latent jet substructure
• Learning to Classify from Impure Samples with High-Dimensional Data
• Autoencoders for unsupervised anomaly detection in high energy physics
• On the Topic of Jets: Disentangling Quarks and Gluons at Colliders
• Gravitational wave signal recognition of O1 data by deep learning
• Novel Jet Observables from Machine Learning
• Machine Learning Uncertainties with Adversarial Neural Networks
• Interaction networks for the identification of boosted $H\to b\overline{b}$ decays
• Jet Charge and Machine Learning
• Rare and Different: Anomaly Scores from a combination of likelihood and out-of-distribution models to detect new physics at the LHC
• Machine learning Calabi-Yau metrics
• An operational definition of quark and gluon jets
• A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty
• Deep-Learning Jets with Uncertainties and More
• Challenges for Unsupervised Anomaly Detection in Particle Physics
• Mapping Machine-Learned Physics into a Human-Readable Space
• Interpretable Deep Learning for Two-Prong Jet Classification with Jet Spectra
• INFERNO: Inference-Aware Neural Optimisation
• A Theory of Quark vs. Gluon Discrimination
• Per-Object Systematics using Deep-Learned Calibration
• Automating the Construction of Jet Observables with Machine Learning
• Mass Agnostic Jet Taggers
• Deep learning for the R-parity violating supersymmetry searches at the LHC
• CapsNets Continuing the Convolutional Quest
• Learning to Identify Electrons
• Identifying the relevant dependencies of the neural network response on characteristics of the input space
• A Triumvirate of AI Driven Theoretical Discovery
• Learning to Inflate
• Quark jet rates and quark/gluon discrimination in multi-jet final states
• Mixture-of-theories Training: Can We Find New Physics and Anomalies Better by Mixing Physical Theories?
• Class Imbalance Techniques for High Energy Physics
• Searching for dark jets with displaced vertices using weakly supervised machine learning
• Unraveling the couplings of a Drell-Yan produced $Z^{\prime }$ with heavy-flavor tagging
• Searching for periodic signals in kinematic distributions using continuous wavelet transforms
• How much joint resummation do we need?
• Search for periodic signals in the dielectron and diphoton invariant mass spectra using 139 fb${}^{-1}$ of $pp$ collisions at $\sqrt{s}=$ 13 TeV with the ATLAS detector
• Machine learning approach for the search of resonances with topological features at the Large Hadron Collider
• Trials Factor for Semi-Supervised NN Classifiers in Searches for Narrow Resonances at the LHC