3 papers
astro-ph.IM2026
Don't Cut Corners: How Training Outside the Prior Makes Simulation-Based Inference More Robust
Chaipat Tirapongprasert, Matthew Ho
Large astrophysical simulation campaigns often generate training data by sampling parameters across a Uniform prior box. Due to the proposal's sharp edge, neural posterior estimato…
astro-ph.IM2026
DegenDetector: Symbolic Recovery of Parameter Degeneracies in Bayesian Posteriors
Chaipat Tirapongprasert, Matthew Ho
We introduce DegenDetector, a framework for identifying and characterizing parameter degeneracies in posterior distributions as closed-form symbolic equations. By combining mutual…
astro-ph.IM2026
Learning at the Edge: Tailed-Uniform Sampling for Robust Simulation-Based Inference
Chaipat Tirapongprasert, Matthew Ho
We introduce the Tailed-Uniform proposal distribution for generating training simulations in simulation-based inference. Instead of sampling parameters uniformly within bounded reg…