Saved in:
Bibliographic Details
Main Authors: Randone, Francesca, Doz, Romina, Tribastone, Mirco, Bortolussi, Luca
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.15167
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914270726324224
author Randone, Francesca
Doz, Romina
Tribastone, Mirco
Bortolussi, Luca
author_facet Randone, Francesca
Doz, Romina
Tribastone, Mirco
Bortolussi, Luca
contents We present DeGAS, a differentiable Gaussian approximate semantics for loopless probabilistic programs that enables sample-free, gradient-based optimization in models with both continuous and discrete components. DeGAS evaluates programs under a Gaussian-mixture semantics and replaces measure-zero predicates and discrete branches with a vanishing smoothing, yielding closed-form expressions for posterior and path probabilities. We prove differentiability of these quantities with respect to program parameters, enabling end-to-end optimization via standard automatic differentiation, without Monte Carlo estimators. On thirteen benchmark programs, DeGAS achieves accuracy and runtime competitive with variational inference and MCMC. Importantly, it reliably tackles optimization problems where sampling-based baselines fail to converge due to conditioning involving continuous variables.
format Preprint
id arxiv_https___arxiv_org_abs_2601_15167
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DeGAS: Gradient-Based Optimization of Probabilistic Programs without Sampling
Randone, Francesca
Doz, Romina
Tribastone, Mirco
Bortolussi, Luca
Programming Languages
We present DeGAS, a differentiable Gaussian approximate semantics for loopless probabilistic programs that enables sample-free, gradient-based optimization in models with both continuous and discrete components. DeGAS evaluates programs under a Gaussian-mixture semantics and replaces measure-zero predicates and discrete branches with a vanishing smoothing, yielding closed-form expressions for posterior and path probabilities. We prove differentiability of these quantities with respect to program parameters, enabling end-to-end optimization via standard automatic differentiation, without Monte Carlo estimators. On thirteen benchmark programs, DeGAS achieves accuracy and runtime competitive with variational inference and MCMC. Importantly, it reliably tackles optimization problems where sampling-based baselines fail to converge due to conditioning involving continuous variables.
title DeGAS: Gradient-Based Optimization of Probabilistic Programs without Sampling
topic Programming Languages
url https://arxiv.org/abs/2601.15167