# Machine Learning the Tip of the Red Giant Branch

**Authors:** Mitchell Dennis, Jeremy Sakstein  
**arXiv:** [2303.12069](https://arxiv.org/abs/2303.12069)  
**Date:** 2023-03-21  
**Summary page:** https://www.summarizepaper.com/en/paper/2303.12069v1/

## Key points

- Mitchell Dennis and Jeremy Sakstein present a novel method for investigating the sensitivity of the tip of the red giant branch (TRGB) I band magnitude $M_I$ to stellar input physics.
- The authors compute a grid of approximately 125,000 theoretical stellar models with varying mass, initial helium abundance, and initial metallicity.
- They train a machine learning emulator to predict $M_I$ as a function of these parameters.
- The emulator can be used to theoretically predict $M_I$ in a given galaxy using Monte Carlo sampling.
- The emulator enables a direct comparison of theoretical predictions for $M_I$ with empirical calibrations to constrain stellar modeling parameters using Bayesian Markov Chain Monte Carlo methods.
- New independent measurements of the metallicity in three galaxies are obtained: Large Magellanic Cloud ($Z=0.0117^{+0.0083}_{-0.0055}$), NGC 4258 ($Z=0.0077^{+0.0074}_{-0.0038}$), and $\omega$-Centauri ($Z=0.0111^{+0.0083}_{-00.0056}$).
- Potential applications include constraining helium abundances and studying populations in globular clusters and dwarf galaxies which could lead to new insights into these systems' formation histories and evolution processes.
- This study presents an innovative approach to investigating the sensitivity of TRGB I band magnitude to stellar input physics and demonstrates its potential for obtaining new independent measurements of important astrophysical parameters such as metallicity in galaxies beyond our own Milky Way which could help us gain further understanding into how galaxies form and evolve over time.

## Summary

In a recent study, Mitchell Dennis and Jeremy Sakstein present a novel method for investigating the sensitivity of the tip of the red giant branch (TRGB) I band magnitude $M_I$ to stellar input physics. The authors compute a grid of approximately 125,000 theoretical stellar models with varying mass, initial helium abundance, and initial metallicity. They then train a machine learning emulator to predict $M_I$ as a function of these parameters. The authors demonstrate two potential applications of their methodology. First, their emulator can be used to theoretically predict $M_I$ in a given galaxy using Monte Carlo sampling. As an example, they predict $M_I = -3.84^{+0.14}_{-0.12}$ in the Large Magellanic Cloud. Secondly, their emulator enables a direct comparison of theoretical predictions for $M_I$ with empirical calibrations to constrain stellar modeling parameters using Bayesian Markov Chain Monte Carlo methods. The authors demonstrate this by using empirical TRGB calibrations to obtain new independent measurements of the metallicity in three galaxies: the Large Magellanic Cloud ($Z=0.0117^{+0.0083}_{-0.0055}$), NGC 4258 ($Z=0.0077^{+0.0074}_{-0.0038}$), and $\omega$-Centauri ($Z=0.0111^{+0.0083}_{-00.0056}$). These results are consistent with other measurements obtained from different methods and provide further evidence for their accuracy and reliability.. The authors also discuss several other potential applications of their methodology such as constraining helium abundances and studying populations in globular clusters and dwarf galaxies which could lead to new insights into these systems' formation histories and evolution processes.. Overall, this study presents an innovative approach to investigating the sensitivity of TRGB I band magnitude to stellar input physics and demonstrates its potential for obtaining new independent measurements of important astrophysical parameters such as metallicity in galaxies beyond our own Milky Way which could help us gain further understanding into how galaxies form and evolve over time..