Deep Multimodal Representation Learning for Stellar Spectra

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Buck, Tobias, Schwarz, Christian
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929594687291392
author Buck, Tobias
Schwarz, Christian
author_facet Buck, Tobias
Schwarz, Christian
contents Recently, contrastive learning (CL), a technique most prominently used in natural language and computer vision, has been used to train informative representation spaces for galaxy spectra and images in a self-supervised manner. Following this idea, we implement CL for stars in the Milky Way, for which recent astronomical surveys have produced a huge amount of heterogeneous data. Specifically, we investigate Gaia XP coefficients and RVS spectra. Thus, the methods presented in this work lay the foundation for aggregating the knowledge implicitly contained in the multimodal data to enable downstream tasks like cross-modal generation or fused stellar parameter estimation. We find that CL results in a highly structured representation space that exhibits explicit physical meaning. Using this representation space to perform cross-modal generation and stellar label regression results in excellent performance with high-quality generated samples as well as accurate and precise label predictions.
format Preprint
id arxiv_https___arxiv_org_abs_2410_16081
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Deep Multimodal Representation Learning for Stellar Spectra
Buck, Tobias
Schwarz, Christian
Solar and Stellar Astrophysics
Astrophysics of Galaxies
Instrumentation and Methods for Astrophysics
Computational Physics
Data Analysis, Statistics and Probability
Recently, contrastive learning (CL), a technique most prominently used in natural language and computer vision, has been used to train informative representation spaces for galaxy spectra and images in a self-supervised manner. Following this idea, we implement CL for stars in the Milky Way, for which recent astronomical surveys have produced a huge amount of heterogeneous data. Specifically, we investigate Gaia XP coefficients and RVS spectra. Thus, the methods presented in this work lay the foundation for aggregating the knowledge implicitly contained in the multimodal data to enable downstream tasks like cross-modal generation or fused stellar parameter estimation. We find that CL results in a highly structured representation space that exhibits explicit physical meaning. Using this representation space to perform cross-modal generation and stellar label regression results in excellent performance with high-quality generated samples as well as accurate and precise label predictions.
title Deep Multimodal Representation Learning for Stellar Spectra
topic Solar and Stellar Astrophysics
Astrophysics of Galaxies
Instrumentation and Methods for Astrophysics
Computational Physics
Data Analysis, Statistics and Probability
url https://arxiv.org/abs/2410.16081