Geometry and clustering with metrics derived from separable Bregman divergences

2018/10/25 by Gomes-Gonçalves, Erika, Gzyl, Henryk, Nielsen, Frank
#FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML)

paper · doi:10.48550/arxiv.1810.10770

Abstract

Separable Bregman divergences induce Riemannian metric spaces that are isometric to the Euclidean space after monotone embeddings. We investigate fixed rate quantization and its codebook Voronoi diagrams, and report on experimental performances of partition-based, hierarchical, and soft clustering algorithms with respect to these Riemann-Bregman distances.

Related

• A scaled Bregman theorem with applications
• Agglomerative Bregman Clustering
• Bregman-Hausdorff divergence: strengthening the connections between computational geometry and machine learning
• Convex Relaxations of Bregman Divergence Clustering
• Strong Coresets for Hard and Soft Bregman Clustering with Applications\n to Exponential Family Mixtures