Reproduce “Compositional Score Modeling for Simulation-based Inference” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Tomas Geffner, George Papamakarios, Andriy Mnih
F-NPSE estimates posteriors from multiple observations in simulation-based inference. It trains a conditional score network on individual observation–parameter pairs, using one simulator call per training case, then composes observation-specific scores and uses annealed Langevin dynamics to sample. Across four tasks, F-NPSE or PF-NPSE performs best in most settings; intermediate grouping often helps, but results are not uniform.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.