Reproduce “Symmetry Guarantees Statistic Recovery in Variational Inference” 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
Daniel Marks, Dario Paccagnan, Mark van der Wilk
The authors explain when variational inference can recover a statistic even if its approximation family cannot represent the target. When target and family respect a transformation group, a unique minimizer shares that symmetry, constraining the statistic to a fixed set. In a sphere example, the axis is recovered at η=1 but not at η=2, where minimizers are non-unique.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.