Bump hunting in latent space
Unsupervised anomaly-detection could be crucial in future analyses searching for rare phenomena in large datasets, as for example collected at the LHC. To this end, we introduce a physics inspired variational autoencoder (VAE) architecture which performs competitively and robustly on the LHC Olympic...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
3 June 2022
|
| In: |
Physical review
Year: 2022, Volume: 105, Issue: 11, Pages: 1-8 |
| ISSN: | 2470-0029 |
| DOI: | 10.1103/PhysRevD.105.115009 |
| Online Access: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1103/PhysRevD.105.115009 Verlag, lizenzpflichtig, Volltext: https://link.aps.org/doi/10.1103/PhysRevD.105.115009 
|
| Author Notes: | Blaž Bortolato, Aleks Smolkovič, Barry M. Dillon, Jernej F. Kamenik |
| Summary: | Unsupervised anomaly-detection could be crucial in future analyses searching for rare phenomena in large datasets, as for example collected at the LHC. To this end, we introduce a physics inspired variational autoencoder (VAE) architecture which performs competitively and robustly on the LHC Olympics Machine Learning Challenge datasets. We demonstrate how embedding some physical observables directly into the VAE latent space, while at the same time keeping the anomaly-detection manifestly agnostic to them, can help to identify and characterize features in measured spectra as caused by the presence of anomalies in a dataset. |
|---|---|
| Item Description: | Gesehen am 20.01.2023 |
| Physical Description: | Online Resource |
| ISSN: | 2470-0029 |
| DOI: | 10.1103/PhysRevD.105.115009 |