Reproduce “On the Power Distance Representation of Bregman Divergences” 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
Frank Nielsen
The authors represent Bregman divergences with weighted points, pairing primal parameters with gradient-space representations. The divergence is half a power distance, defined as squared Euclidean separation minus the weights; argument order preserves possible asymmetry. This converts queries to weighted geometry. In two dimensions, closest-pair and diameter problems have theoretical O(n log n) algorithms. The Bregman minimum enclosing ball requires an exact double-sided power reformulation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.