notesum.ai

Published at November 27

Using different sources of ground truths and transfer learning to improve the generalization of photometric redshift estimation

astro-ph.IM

astro-ph.GA

cs.LG

Released Date: November 27, 2024

Authors: Jonathan Soriano¹, Srinath Saikrishnan¹, Vikram Seenivasan¹, Bernie Boscoe², Jack Singal³, Tuan Do¹

Aff.: ¹Physics and Astronomy Department, UCLA, Los Angeles, CA 90024; ²Computer Science Department, Southern Oregon University, Ashland, OR 97520; ³Physics Department, University of Richmond, Richmond, VA 23173

Arxiv: http://arxiv.org/abs/2411.18054v1

Dataset	Number of	Redshift	Median Redshift	i-band mag	No.
	Sources	90th percentile	Uncertainty	90th percentile	Filters
TransferZ	116,335	1.9	0.03	25	5
GalaxiesML	286,401	1.2	0.0002	22	5
Combo Data	402,408	1.5	0.0006	24	5