Reproduce “Spectra: Exact Component Transport for Test-Time Prior Adaptation in 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

Xin Zhao, Nico Scherf, Robert Trampel, Kerrin J. Pine, Nikolaus Weiskopf

Spectra adapts a frozen diffusion model for simulation-based inference to changed priors without retraining. It transforms the model’s score query using the ratio between target and training priors; exact single-factor transport applies to isotropic exponential–quadratic factors, while mixtures extend coverage with estimated weights. On strong-shift SLCP, C2ST was 0.608 versus PriorGuide’s 0.637; this is not unrestricted prior adaptation.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.