Reproduce “Deep unfolding of MCMC kernels: scalable, modular & explainable GANs for high-dimensional posterior sampling” 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
Jonathan Spence, Tobías I. Liaudat, Konstantinos Zygalakis, Marcelo Pereyra
The authors introduce U-LATINO, a trainable sampler that unfolds MCMC steps while retaining the measurement model. It combines a learned image prior with data consistency, and is trained for task-specific reconstruction. On PROBES radio-interferometry tests, U-LATINO achieves 43.40 dB PSNR, compared with 37.33 for zero-shot LATINO at eight denoiser evaluations; this result is limited to the tested setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.