AutoStan demonstrates a CLI coding agent that autonomously builds and iteratively improves interpretable Bayesian models in Stan.
March 31, 2026
Original Paper
AutoStan: Autonomous Bayesian Model Improvement via Predictive Feedback
arXiv · 2603.27766
The Takeaway
Shows that agents can handle highly specialized probabilistic programming tasks, moving from simple linear models to complex hierarchical and mixture models using only predictive feedback (NLPD) and sampler diagnostics. This bridges the gap between black-box ML and interpretable statistics.
From the abstract
We present AutoStan, a framework in which a command-line interface (CLI) coding agent autonomously builds and iteratively improves Bayesian models written in Stan. The agent operates in a loop, writing a Stan model file, executing MCMC sampling, then deciding whether to keep or revert each change based on two complementary feedback signals: the negative log predictive density (NLPD) on held-out data and the sampler's own diagnostics (divergences, R-hat, effective sample size). We evaluate AutoSt