Symmetries, safety, and self-supervision

Dillon, Barry M.; Kasieczka, G.; Olischlager, Hans; Plehn, Tilman; Sorrenson, Peter; Vogel, Lorenz

doi:10.21468/scipostphys.12.6.188

Cited by 24 publications

(17 citation statements)

References 67 publications

(82 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, we apply power counting to study energy flow polynomials (EFPs) [28], which are an overcomplete linear basis for infrared-and-collinear safe jet substructure. EFPs have been used in a variety of machine learning tasks [29][30][31][32][33][34][35][36], so it is a natural context to study how best to represent jet information. By exploiting power counting, we show how to simplify the EFP basis for analysis tasks involving quark and gluon jets.…”

Section: Jhep09(2022)021mentioning

confidence: 99%

Power counting energy flow polynomials

Cal

Thaler

Waalewijn

2022

J. High Energ. Phys.

View full text Add to dashboard Cite

Power counting is a systematic strategy for organizing collider observables and their associated theoretical calculations. In this paper, we use power counting to characterize a class of jet substructure observables called energy flow polynomials (EFPs). EFPs provide an overcomplete linear basis for infrared-and-collinear safe jet observables, but it is known that in practice, a small subset of EFPs is often sufficient for specific jet analysis tasks. By applying power counting arguments, we obtain linear relationships between EFPs that hold for quark and gluon jets to a specific order in the power counting. We test these relations in the parton shower generator Pythia, finding excellent agreement. Power counting allows us to truncate the basis of EFPs without affecting performance, which we corroborate through a study of quark-gluon tagging and regression.

show abstract

Section: Jhep09(2022)021mentioning

confidence: 99%

Power counting energy flow polynomials

Cal

Thaler

Waalewijn

2022

J. High Energ. Phys.

View full text Add to dashboard Cite

show abstract

“…Now the network learns how to process high-dimensional correlations in the data, and thus the representations learned by these networks can be very useful for downstream tasks. We introduced the self-supervised JetCLR method in [53] and demonstrated its ability to construct highly expressive representations for classification tasks. In [54] this same technique was used to construct representations for CWoLa-based anomaly detection.…”

Section: Introductionmentioning

confidence: 99%

“…In addition to these works, other self-supervised / representation learning techniques have been applied in particle physics [55,56] and in other scientific disciplines such as astrophysics [57][58][59][60]. In [53,54] the augmentations corresponded to transformations of the event to which the underlying physics should be invariant to rotations or translations, but also soft-collinear parton splittings.…”

Section: Introductionmentioning

confidence: 99%

Anomalies, Representations, and Self-Supervision

Dillon¹,

Favaro²,

Feiden³

et al. 2023

Preprint

View full text Add to dashboard Cite

We develop a self-supervised method for density-based anomaly detection using contrastive learning, and test it using event-level anomaly data from CMS ADC2021. The Anomaly-CLR technique is data-driven and uses augmentations of the background data to mimic non-Standard-Model events in a model-agnostic way. It uses a permutation-invariant Transformer Encoder architecture to map the objects measured in a collider event to the representation space, where the data augmentations define a representation space which is sensitive to potential anomalous features. An AutoEncoder trained on background representations then computes anomaly scores for a variety of signals in the representation space. With AnomalyCLR we find significant improvements on performance metrics for all signals when compared to the raw data baseline.

show abstract

“…Embedding into a lower-dimensional space can also be seen as an alternate way of building a simpler space to perform physics measurements, searches, and classification, such as in [48]. We bypass the need to do latent variable modeling by directly building the space through embedding with optimal transport distances on the original space, yielding a new handle in how to organize and classify data.…”

Section: Introductionmentioning

confidence: 99%

Neural Embedding: Learning the Embedding of the Manifold of Physics Data

Park¹,

Harris²,

Ostdiek³

2022

Preprint

View full text Add to dashboard Cite

In this paper, we present a method of embedding physics data manifolds with metric structure into lower dimensional spaces with simpler metrics, such as Euclidean and Hyperbolic spaces. We then demonstrate that it can be a powerful step in the data analysis pipeline for many applications. Using progressively more realistic simulated collisions at the Large Hadron Collider, we show that this embedding approach learns the underlying latent structure. With the notion of volume in Euclidean spaces, we provide for the first time a viable solution to quantifying the true search capability of model agnostic search algorithms in collider physics (i.e. anomaly detection). Finally, we discuss how the ideas presented in this paper can be employed to solve many practical challenges that require the extraction of physically meaningful representations from information in complex high dimensional datasets.

show abstract

Symmetries, safety, and self-supervision

Cited by 24 publications

References 67 publications

Power counting energy flow polynomials

Power counting energy flow polynomials

Anomalies, Representations, and Self-Supervision

Neural Embedding: Learning the Embedding of the Manifold of Physics Data

Contact Info

Product

Resources

About