Ephemeral learning: augmenting triggers with online-trained normalizing flows
The large data rates at the LHC require an online trigger system to select relevant collisions. Rather than compressing individual events, we propose to compress an entire data set at once. We use a normalizing flow as a deep generative model to learn the probability density of the data online. The...
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Dokumenttyp: | Article (Journal) |
| Sprache: | Englisch |
| Veröffentlicht: |
07-10-2022
|
| In: |
SciPost physics
Year: 2022, Jahrgang: 13, Heft: 4, Pages: 1-17 |
| ISSN: | 2542-4653 |
| DOI: | 10.21468/SciPostPhys.13.4.087 |
| Online-Zugang: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.21468/SciPostPhys.13.4.087 Verlag, lizenzpflichtig, Volltext: https://scipost.org/10.21468/SciPostPhys.13.4.087 
|
| Verfasserangaben: | Anja Butter, Sascha Diefenbacher, Gregor Kasieczka, Benjamin Nachman, Tilman Plehn, David Shih and Ramon Winterhalder |
| Zusammenfassung: | The large data rates at the LHC require an online trigger system to select relevant collisions. Rather than compressing individual events, we propose to compress an entire data set at once. We use a normalizing flow as a deep generative model to learn the probability density of the data online. The events are then represented by the generative neural network and can be inspected offline for anomalies or used for other analysis purposes. We demonstrate our new approach for a toy model and a correlation-enhanced bump hunt. |
|---|---|
| Beschreibung: | Gesehen am 28.11.2023 |
| Beschreibung: | Online Resource |
| ISSN: | 2542-4653 |
| DOI: | 10.21468/SciPostPhys.13.4.087 |