Part of Advances in Neural Information Processing Systems 25 (NIPS 2012)
Amr Ahmed, Sujith Ravi, Alex Smola, Shravan Narayanamurthy
Clustering is a key component in data analysis toolbox. Despite its importance, scalable algorithms often eschew rich statistical models in favor of simpler descriptions such as $k$-means clustering. In this paper we present a sampler, capable of estimating mixtures of exponential families. At its heart lies a novel proposal distribution using random projections to achieve high throughput in generating proposals, which is crucial for clustering models with large numbers of clusters.