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

arXiv: https://arxiv.org/abs/2411.18054 · HTML: https://mlsanity.com/paper/2411.18054/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 1 of its 50 references that are in the map.
- First posted on arXiv 27 Nov 2024.

## Builds on (most cited first, 1 of 1)

- [Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimization](https://mlsanity.com/paper/1603.06560/) (2016)

## Built on by (most cited first, 3 of 3)

- [Multi-Modal Masked Autoencoders for Learning Image-Spectrum Associations for Galaxy Evolution and Cosmology](https://mlsanity.com/paper/2510.22527/) (2025)
- [Combining datasets with different ground truths using Low-Rank Adaptation to generalize image-based CNN models for photometric redshift prediction](https://mlsanity.com/paper/2601.00146/) (2026)
- [AquiLLM: Evaluating Faithfulness in Open-Weight RAG-LLM Systems for Scientific Research](https://mlsanity.com/paper/2609.16519/) (2026)

## Abstract (arXiv)

In this work, we explore methods to improve galaxy redshift predictions by combining different ground truths. Traditional machine learning models rely on training sets with known spectroscopic redshifts, which are precise but only represent a limited sample of galaxies. To make redshift models more generalizable to the broader galaxy population, we investigate transfer learning and directly combining ground truth redshifts derived from photometry and spectroscopy. We use the COSMOS2020 survey to create a dataset, TransferZ, which includes photometric redshift estimates derived from up to 35 imaging filters using template fitting. This dataset spans a wider range of galaxy types and colors compared to spectroscopic samples, though its redshift estimates are less accurate. We first train a base neural network on TransferZ and then refine it using transfer learning on a dataset of galaxies with more precise spectroscopic redshifts (GalaxiesML). In addition, we train a neural network on a combined dataset of TransferZ and GalaxiesML. Both methods reduce bias by ~ 5x, RMS error by ~ 1.5x, and catastrophic outlier rates by 1.3x on GalaxiesML, compared to a baseline trained only on TransferZ. However, we also find a reduction in performance for RMS and bias when evaluated on TransferZ data. Overall, our results demonstrate these approaches can meet cosmological requirements.