Current browse context:
cs.LG
Change to browse by:
References & Citations
Computer Science > Machine Learning
Title: Compiling Stan to Generative Probabilistic Languages and Extension to Deep Probabilistic Programming
(Submitted on 30 Sep 2018 (v1), last revised 11 Apr 2021 (this version, v5))
Abstract: Stan is a probabilistic programming language that is popular in the statistics community, with a high-level syntax for expressing probabilistic models. Stan differs by nature from generative probabilistic programming languages like Church, Anglican, or Pyro. This paper presents a comprehensive compilation scheme to compile any Stan model to a generative language and proves its correctness. We use our compilation scheme to build two new backends for the Stanc3 compiler targeting Pyro and NumPyro. Experimental results show that the NumPyro backend yields a 2.3x speedup compared to Stan in geometric mean over 26 benchmarks. Building on Pyro we extend Stan with support for explicit variational inference guides and deep probabilistic models. That way, users familiar with Stan get access to new features without having to learn a fundamentally new language.
Submission history
From: Louis Mandel [view email][v1] Sun, 30 Sep 2018 15:39:53 GMT (194kb,D)
[v2] Wed, 1 Jul 2020 20:45:47 GMT (90kb,D)
[v3] Mon, 3 Aug 2020 16:29:27 GMT (89kb,D)
[v4] Tue, 12 Jan 2021 20:51:14 GMT (91kb,D)
[v5] Sun, 11 Apr 2021 15:34:02 GMT (351kb,D)
Link back to: arXiv, form interface, contact.