Reproduce “A Tale of Two Flows: Cooperative Learning of Langevin Flow and Normalizing Flow Toward Energy-Based Model” 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
Jianwen Xie, Yaxuan Zhu, Jun Li, Ping Li
CoopFlow combines a normalizing flow with a short Langevin run for image generation. The flow maps noise to an initial image, then learns from images revised by the Langevin process, reducing sampling steps sought without eliminating them. On CIFAR-10, it reports lower FID than either component alone; its validity claim is feature-statistic moment matching, not full distribution matching.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.