Random forests reproduce simulated halo gas density profiles to roughly 80-90% accuracy, and Sobol analysis of those forests ranks halo mass and central gas mass as the dominant predictors across EAGLE, IllustrisTNG, and Simba.
Building Trustworthy Machine Learning Models for Astronomy
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
Astronomy is entering an era of data-driven discovery, due in part to modern machine learning (ML) techniques enabling powerful new ways to interpret observations. This shift in our scientific approach requires us to consider whether we can trust the black box. Here, we overview methods for an often-overlooked step in the development of ML models: building community trust in the algorithms. Trust is an essential ingredient not just for creating more robust data analysis techniques, but also for building confidence within the astronomy community to embrace machine learning methods and results.
fields
astro-ph.GA 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Interpretable machine learning of halo gas density profiles: a sensitivity analysis of cosmological hydrodynamical simulations
Random forests reproduce simulated halo gas density profiles to roughly 80-90% accuracy, and Sobol analysis of those forests ranks halo mass and central gas mass as the dominant predictors across EAGLE, IllustrisTNG, and Simba.