An example of a candidate image processed in real time from a past production run of DWF. Each panel is a small 121 x 121 pixel image, equivalent to ~30 x 30 arcseconds of sky centered on the candidate. The left panel is the (deeper) template image taken before the DWF observation. The middle panel is the current scientific image of the sky taken a few minutes ago, and the subtracted image is the digital subtraction of the two images. Any constant flux is subtracted, e.g. flux differences from transient sources remain in the subtracted image. — astro-ph.IM
As the volume and availability of astronomical data grows rapidly, astronomers will soon rely on the use of machine learning algorithms in their daily work.
This proceedings aims to provide an overview of what machine learning is, delve into different types of learning algorithms, and explore two astronomical use cases. Machine learning has opened up a world of possibilities for us astronomers working with large amounts of data, but users can fall into common pitfalls if they are not careful.
Here we focus on solving problems related to time-series light curve data and optical imaging data, primarily from the Deeper, Wider, Faster Program (DWF). Online notes demonstrating these various techniques are provided alongside written examples. This guide is intended to help you build a small toolkit of knowledge and tools to take home with you for use in future machine learning projects.
Sarah A. Webb, Simon R. Goode
Comment: 12 pages, 5 figures, International Astronomical Union Proceedings Series 368
Subject: Instruments and Methods for Astrophysics (astro-ph.IM)
Quoted as: arXiv:2304.00512 [astro-ph.IM] (or arXiv:2304.00512v1 [astro-ph.IM] for this version)
Submission history
From: Sarah Webb
[v1] Sunday, April 2, 2023 11:03:48 UTC (2,339 KB)
https://arxiv.org/abs/2304.00512
