Reproduce “DIPHINE: Diffusion-based $Φ$-ID Neural Estimator” 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
Simon Pedro Galeano Munoz, Mustapha Bounoua, Giulio Franzese, Pietro Michiardi, Maurizio Filippone
DIPHINE estimates all 16 information-dynamic atoms for ergodic systems with two partitions, using one diffusion-based network to estimate nine mutual-information terms before applying minimum-mutual-information redundancy and Möbius inversion. In d=10 synthetic V-A-R(1) tests, its atom M-A-E was 0.038–0.044, below InfoNCE’s 0.055–0.063, while its intermediate M-I estimates did not outperform every baseline. These results do not establish real-world accuracy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.