Reproduce “Kernelized Cumulants: Beyond Kernel Mean Embeddings” 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
Patric Bonnier, Harald Oberhauser, Zoltán Szabó
The authors extend kernel mean embeddings with cumulants that compare higher-order structure in kernel feature space, defining them by subtracting contributions from lower-order moments and computing comparisons through kernel values. Their degree-two statistic outperforms MMD sooner in a tested moment-matched two-sample example. The full cumulant sequence characterizes distributions under stated kernel assumptions; finite-degree tests do not guarantee broader gains.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.