Reproduce “Understanding Deep Generative Models with Generalized Empirical Likelihoods” 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
Suman Ravuri, Mélanie Rey, Shakir Mohamed, Marc Deisenroth
The authors propose generalized empirical likelihood diagnostics that reweight observed data and generated examples to satisfy chosen feature-based moment conditions, using individual weights to reveal distributional mismatches. In a controlled CIFAR-10 mode-dropping experiment, kernel generalized empirical likelihood achieved Hellinger distance 0.2732, versus 0.4805 for coverage and 0.6137 for improved recall, with eight classes omitted; this finding is limited to that setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.