REVIEW 16 cited by
Classification without labels: Learning from mixed samples in high energy physics
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
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.
Forward citations
Cited by 16 Pith papers
-
Simplex Demixing: Disentangling Multiple Light-Flavor Jets at Colliders
Simplex demixing recovers T mutually irreducible jet-flavor topics from M mixed samples via the (T−1)-simplex geometry of a multi-category classifier, demonstrated on Pythia dijets.
-
Learning Standard Model structure from LHC data with Riemannian flow matching
ShellFlow, a Riemannian flow-matching transformer fed only on-shell and invariant-mass priors and ~8×10^8 recorded ATLAS events, reproduces the SM's dilepton resonances, Weinberg angle, and top/W mass peaks in a singl...
-
Model-agnostic search for dijet resonances with anomalous jet substructure in proton-proton collisions at $\sqrt{s}$ = 13 TeV
A model-agnostic CMS search for dijet resonances with anomalous jet substructure finds no excess and reports first exclusion limits on several benchmark signals, with ML anomaly detection improving sensitivity over in...
-
Multiclass Classification without Labels via Posterior Simplex Geometry
Bayes-optimal mixture posteriors lie on a latent (K−1)-simplex whose vertices identify classes and mixing weights, enabling prior-free multiclass recovery from mixture identity alone.
-
Towards anomaly detection searches for new physics signatures including Higgs bosons with weakly supervised machine learning
HAXAD, a weakly supervised anomaly-detection search for Higgs-plus-X new physics, is extended with new embeddings and limit-setting, and on 470 fb^-1 of pseudo-data it matches or exceeds the best single cut-based limi...
-
Explicit or Implicit? Encoding Physics at the Precision Frontier
On three precision classification tasks — reweighting-based unfolding, likelihood-ratio estimation, and weakly supervised anomaly detection — a Lorentz-equivariant transformer and a pretrained foundation model perform...
-
Look everywhere effects in anomaly detection
Weakly supervised anomaly detectors that train and test on the same data produce badly miscalibrated p-values; independent test sets are calibrated but insensitive, while k-fold cross-validation is a workable middle ground.
-
Graph theory inspired anomaly detection at the LHC
Sparse globally rigid graph representations of jets, combined with roughly 30 reclustered subjets, improve graph autoencoder anomaly detection on the LHC Olympics benchmark.
-
How much joint resummation do we need?
Joint resummation of two angularities, rather than one or many, yields the largest gain in predicting other angularities in e+ e- dijet events.
-
Mono-X Signal Characterization from Two-component Dark Matter Using a Convolutional Neural Network
A 1D multi-channel CNN trained on normalized histograms of simulated mono-jet and mono-Z events can partially classify one- versus two-component dark matter and regress masses, but only in a background-free, model-spe...
-
Weakly supervised machine learning for model-agnostic searches of new phenomena in the $\gamma$-ray sky
Weakly supervised classifiers trained on background-versus-mixture samples can identify anomalous gamma-ray sources without labeled signal templates, approaching supervised performance in controlled benchmarks.
-
Optimal Transport Event Representation for Anomaly Detection
Adding a few optimal-transport-based features to standard jet observables nearly doubles anomaly-detection significance at 0.5% signal injection on LHC Olympics benchmarks.
-
Robust resonant anomaly detection with NPLM
NPLM-based classifiers and end-to-end NPLM outperform BDT-based anomaly detection at low signal injection on the LHCO and RODEM benchmarks, with lower variance across hyperparameters.
-
Quantum similarity learning for anomaly detection
A hybrid Transformer-quantum circuit similarity-learning network reaches AUC 96.1% on simulated di-Higgs anomaly detection, slightly above a classical baseline, with clustering mitigating shot noise.
-
Improving the performance of weak supervision searches using data augmentation
Physics-inspired data augmentation halves the signal data requirement for CWoLa weak supervision searches, cutting the practical sensitivity threshold from roughly 6 sigma to roughly 3 sigma.
-
Machine Learning is Good for Physics - and Vice Versa
A perspective essay arguing that AI should be integrated into fundamental physics while preserving the field's statistical and theory-based standards, and that physics can enrich machine learning.
Discussion (0). Continue with ORCID to comment.