Reproduce “Learning on-the-fly quasi-stationarity diagnostics from low-dimensional surrogates” 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
Noé Blassel
The authors introduce an on-the-fly recurrent neural network to decide when replicas in metastable states reach a quasi-stationary distribution. It uses successive collective-variable histograms, learned from low-dimensional surrogates. On synthetic benchmarks, its ROC-AUC—the probability a positive ranks above a negative—was 0.96 versus 0.82 for Gelman–Rubin; with a nonslow variable, early-decorrelation risk was 57% at α=0.1.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.