Production-ready K-Means clustering for Apache Spark with pluggable Bregman divergences (KL, Itakura-Saito, L1, etc). 6 algorithms, 740 tests, cross-version persistence. Drop-in replacement for MLlib with mathematically correct distance functions for probability distributions, spectral data, and count data.

Fast and memory-efficient clustering + coreset construction, including fast distance kernels for Bregman and f-divergences.

The official source code to: Uncertainty Estimates of Predictions via a General Bias-Variance Decomposition (AISTATS'23)

Python implementation of Bregman Hard Clustering and Bregman Soft Clustering as a scikit-learn module.

Approximate Bregman proximal gradient algorithm

SWIG bindings for best subset, best partition solvers, multiple flavors, multiple objectives

Mirror Descent on Sphere Geometry