Neural embedding: learning the embedding of the manifold of physics data
Abstract 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 anal...
| Hoofdauteurs: | , , |
|---|---|
| Andere auteurs: | |
| Formaat: | Artikel |
| Taal: | English |
| Gepubliceerd in: |
Springer Berlin Heidelberg
2023
|
| Online toegang: | https://hdl.handle.net/1721.1/151120 |
| _version_ | 1826203459941040128 |
|---|---|
| author | Park, Sang E. Harris, Philip Ostdiek, Bryan |
| author2 | Massachusetts Institute of Technology. Department of Physics |
| author_facet | Massachusetts Institute of Technology. Department of Physics Park, Sang E. Harris, Philip Ostdiek, Bryan |
| author_sort | Park, Sang E. |
| collection | MIT |
| description | Abstract
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. |
| first_indexed | 2024-09-23T12:37:41Z |
| format | Article |
| id | mit-1721.1/151120 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T12:37:41Z |
| publishDate | 2023 |
| publisher | Springer Berlin Heidelberg |
| record_format | dspace |
| spelling | mit-1721.1/1511202024-01-19T22:00:27Z Neural embedding: learning the embedding of the manifold of physics data Park, Sang E. Harris, Philip Ostdiek, Bryan Massachusetts Institute of Technology. Department of Physics Abstract 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. 2023-07-17T12:49:25Z 2023-07-17T12:49:25Z 2023-07-12 2023-07-16T03:11:15Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/151120 Journal of High Energy Physics. 2023 Jul 12;2023(7):108 PUBLISHER_CC en https://doi.org/10.1007/JHEP07(2023)108 Creative Commons Attribution http://creativecommons.org/licenses/by/4.0/ The Author(s) application/pdf Springer Berlin Heidelberg Springer Berlin Heidelberg |
| spellingShingle | Park, Sang E. Harris, Philip Ostdiek, Bryan Neural embedding: learning the embedding of the manifold of physics data |
| title | Neural embedding: learning the embedding of the manifold of physics data |
| title_full | Neural embedding: learning the embedding of the manifold of physics data |
| title_fullStr | Neural embedding: learning the embedding of the manifold of physics data |
| title_full_unstemmed | Neural embedding: learning the embedding of the manifold of physics data |
| title_short | Neural embedding: learning the embedding of the manifold of physics data |
| title_sort | neural embedding learning the embedding of the manifold of physics data |
| url | https://hdl.handle.net/1721.1/151120 |
| work_keys_str_mv | AT parksange neuralembeddinglearningtheembeddingofthemanifoldofphysicsdata AT harrisphilip neuralembeddinglearningtheembeddingofthemanifoldofphysicsdata AT ostdiekbryan neuralembeddinglearningtheembeddingofthemanifoldofphysicsdata |