Reproduce “Machine Learning the H-theorem” 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
Ruben Lier
The authors train a permutation-invariant model to rank velocity sets from a gas of 1,000 hard disks, using temporal order without the Boltzmann H-functional as a training target. The learned scalar is compared with independently calculated H. After separate affine fits, it tracks relaxation in four runs, more smoothly at τ = 0.1; this is not a derivation of the H-theorem.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.