Reproduce “Low-Temperature Sampling via Equilibrium Disagreements” 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
Andreas Galanis, Daniel Stefankovic
The authors analyze low-temperature Glauber dynamics within a phase by studying disagreements between independent equilibrium samples and constructing local paths that switch small disagreement components. Across hard-core, Ising, and random-cluster models on fixed-degree random regular graphs, this gives n^{3/2+o(1)} updates from natural extreme starts, with high probability and arbitrarily small inverse-polynomial error, not worst-case mixing across phases.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.