Reproduce “Learning Through Energy Refinement and Manifold Projection: A Cooperative EBM-AE Framework” 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
Ryad Zemouri
The EBM→AE→EBM method generates handwritten digits by alternating Langevin refinement with a jointly trained autoencoder’s projection, then applying further energy refinement. On MNIST, its reported best-observed FID is 2,834.0, compared with 3,463.0 for AE Joint; lower is better. The comparison uses features from a dedicated MNIST classifier and does not establish lower runtime or computation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.