Improved Marginal Unbiased Score Expansion (MUSE) via Implicit Differentiation

September 21, 2022 · Entered Twilight · 🏛 arXiv.org

Repo contents: README.md, poetry.lock, pyproject.toml, scaling.ipynb, toy_problems.ipynb

Authors Marius Millea arXiv ID 2209.10512 Category stat.ML: Machine Learning (Stat) Cross-listed cs.AI, cs.LG Citations 4 Venue arXiv.org Repository https://github.com/marius311/muse-implicit-paper ⭐ 5 Last Checked 3 months ago

Abstract

We apply the technique of implicit differentiation to boost performance, reduce numerical error, and remove required user-tuning in the Marginal Unbiased Score Expansion (MUSE) algorithm for hierarchical Bayesian inference. We demonstrate these improvements on three representative inference problems: 1) an extended Neal's funnel 2) Bayesian neural networks, and 3) probabilistic principal component analysis. On our particular test cases, MUSE with implicit differentiation is faster than Hamiltonian Monte Carlo by factors of 155, 397, and 5, respectively, or factors of 65, 278, and 1 without implicit differentiation, and yields good approximate marginal posteriors. The Julia and Python MUSE packages have been updated to use implicit differentiation, and can solve problems defined by hand or with any of a number of popular probabilistic programming languages and automatic differentiation backends.