Truncation-free Hybrid Inference for DPMM

Arnim Bleier

Dirichlet process mixture models (DPMM) are a cornerstone of Bayesian non-parametrics. While these models free from choosing the number of components a-priori, computationally attr…