Reproduce “Mutual Information, Neural Networks and the Renormalization Group” 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
Maciej Koch-Janusz, Zohar Ringel
The authors introduce a neural-network procedure that learns coarse variables by maximizing mutual information with a distant environment. In dimer samples, it recovers field-like patterns while ignoring added decoupled noise. For the two-dimensional square-lattice Ising model, it estimates the correlation-length exponent as ν ≈ 1.0 ± 0.15, consistent with exact value 1, in the tested setting.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.