Maven: A Multimodal Foundation Model for Supernova Science

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Gemma, Helfer, Thomas, Gagliano, Alexander T., Mishra-Sharma, Siddharth, Villar, V. Ashley
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929479168819200
author Zhang, Gemma
Helfer, Thomas
Gagliano, Alexander T.
Mishra-Sharma, Siddharth
Villar, V. Ashley
author_facet Zhang, Gemma
Helfer, Thomas
Gagliano, Alexander T.
Mishra-Sharma, Siddharth
Villar, V. Ashley
contents A common setting in astronomy is the availability of a small number of high-quality observations, and larger amounts of either lower-quality observations or synthetic data from simplified models. Time-domain astrophysics is a canonical example of this imbalance, with the number of supernovae observed photometrically outpacing the number observed spectroscopically by multiple orders of magnitude. At the same time, no data-driven models exist to understand these photometric and spectroscopic observables in a common context. Contrastive learning objectives, which have grown in popularity for aligning distinct data modalities in a shared embedding space, provide a potential solution to extract information from these modalities. We present Maven, the first foundation model for supernova science. To construct Maven, we first pre-train our model to align photometry and spectroscopy from 0.5M synthetic supernovae using a constrastive objective. We then fine-tune the model on 4,702 observed supernovae from the Zwicky Transient Facility. Maven reaches state-of-the-art performance on both classification and redshift estimation, despite the embeddings not being explicitly optimized for these tasks. Through ablation studies, we show that pre-training with synthetic data improves overall performance. In the upcoming era of the Vera C. Rubin Observatory, Maven serves as a Rosetta Stone for leveraging large, unlabeled and multimodal time-domain datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2408_16829
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Maven: A Multimodal Foundation Model for Supernova Science
Zhang, Gemma
Helfer, Thomas
Gagliano, Alexander T.
Mishra-Sharma, Siddharth
Villar, V. Ashley
High Energy Astrophysical Phenomena
Instrumentation and Methods for Astrophysics
Machine Learning
A common setting in astronomy is the availability of a small number of high-quality observations, and larger amounts of either lower-quality observations or synthetic data from simplified models. Time-domain astrophysics is a canonical example of this imbalance, with the number of supernovae observed photometrically outpacing the number observed spectroscopically by multiple orders of magnitude. At the same time, no data-driven models exist to understand these photometric and spectroscopic observables in a common context. Contrastive learning objectives, which have grown in popularity for aligning distinct data modalities in a shared embedding space, provide a potential solution to extract information from these modalities. We present Maven, the first foundation model for supernova science. To construct Maven, we first pre-train our model to align photometry and spectroscopy from 0.5M synthetic supernovae using a constrastive objective. We then fine-tune the model on 4,702 observed supernovae from the Zwicky Transient Facility. Maven reaches state-of-the-art performance on both classification and redshift estimation, despite the embeddings not being explicitly optimized for these tasks. Through ablation studies, we show that pre-training with synthetic data improves overall performance. In the upcoming era of the Vera C. Rubin Observatory, Maven serves as a Rosetta Stone for leveraging large, unlabeled and multimodal time-domain datasets.
title Maven: A Multimodal Foundation Model for Supernova Science
topic High Energy Astrophysical Phenomena
Instrumentation and Methods for Astrophysics
Machine Learning
url https://arxiv.org/abs/2408.16829