Invariant representation driven neural classifier for anti-QCD jet tagging
Abstract We leverage representation learning and the inductive bias in neural-net-based Standard Model jet classification tasks, to detect non-QCD signal jets. In establishing the framework for classification-based anomaly detection in jet physics, we demonstrate that, with a well-calibrated and pow...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2022-10-01
|
| Series: | Journal of High Energy Physics |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/JHEP10(2022)152 |
| _version_ | 1797991793020633088 |
|---|---|
| author | Taoli Cheng Aaron Courville |
| author_facet | Taoli Cheng Aaron Courville |
| author_sort | Taoli Cheng |
| collection | DOAJ |
| description | Abstract We leverage representation learning and the inductive bias in neural-net-based Standard Model jet classification tasks, to detect non-QCD signal jets. In establishing the framework for classification-based anomaly detection in jet physics, we demonstrate that, with a well-calibrated and powerful enough feature extractor, a well-trained mass-decorrelated supervised Standard Model neural jet classifier can serve as a strong generic anti-QCD jet tagger for effectively reducing the QCD background. Imposing data-augmented mass-invariance (and thus decoupling the dominant factor) not only facilitates background estimation, but also induces more substructure-aware representation learning. We are able to reach excellent tagging efficiencies for all the test signals considered. In the best case, we reach a background rejection rate of 51 and a significance improvement factor of 3.6 at 50% signal acceptance, with the jet mass decorrelated. This study indicates that supervised Standard Model jet classifiers have great potential in general new physics searches. |
| first_indexed | 2024-04-11T08:57:53Z |
| format | Article |
| id | doaj.art-f2c92e77811a4d7ab928c1009ba118b2 |
| institution | Directory Open Access Journal |
| issn | 1029-8479 |
| language | English |
| last_indexed | 2024-04-11T08:57:53Z |
| publishDate | 2022-10-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Journal of High Energy Physics |
| spelling | doaj.art-f2c92e77811a4d7ab928c1009ba118b22022-12-22T04:33:09ZengSpringerOpenJournal of High Energy Physics1029-84792022-10-0120221013410.1007/JHEP10(2022)152Invariant representation driven neural classifier for anti-QCD jet taggingTaoli Cheng0Aaron Courville1Mila — Quebec Artificial Intelligence InstituteMila — Quebec Artificial Intelligence InstituteAbstract We leverage representation learning and the inductive bias in neural-net-based Standard Model jet classification tasks, to detect non-QCD signal jets. In establishing the framework for classification-based anomaly detection in jet physics, we demonstrate that, with a well-calibrated and powerful enough feature extractor, a well-trained mass-decorrelated supervised Standard Model neural jet classifier can serve as a strong generic anti-QCD jet tagger for effectively reducing the QCD background. Imposing data-augmented mass-invariance (and thus decoupling the dominant factor) not only facilitates background estimation, but also induces more substructure-aware representation learning. We are able to reach excellent tagging efficiencies for all the test signals considered. In the best case, we reach a background rejection rate of 51 and a significance improvement factor of 3.6 at 50% signal acceptance, with the jet mass decorrelated. This study indicates that supervised Standard Model jet classifiers have great potential in general new physics searches.https://doi.org/10.1007/JHEP10(2022)152AutomationJets and Jet Substructure |
| spellingShingle | Taoli Cheng Aaron Courville Invariant representation driven neural classifier for anti-QCD jet tagging Journal of High Energy Physics Automation Jets and Jet Substructure |
| title | Invariant representation driven neural classifier for anti-QCD jet tagging |
| title_full | Invariant representation driven neural classifier for anti-QCD jet tagging |
| title_fullStr | Invariant representation driven neural classifier for anti-QCD jet tagging |
| title_full_unstemmed | Invariant representation driven neural classifier for anti-QCD jet tagging |
| title_short | Invariant representation driven neural classifier for anti-QCD jet tagging |
| title_sort | invariant representation driven neural classifier for anti qcd jet tagging |
| topic | Automation Jets and Jet Substructure |
| url | https://doi.org/10.1007/JHEP10(2022)152 |
| work_keys_str_mv | AT taolicheng invariantrepresentationdrivenneuralclassifierforantiqcdjettagging AT aaroncourville invariantrepresentationdrivenneuralclassifierforantiqcdjettagging |