Training Gaussian Mixture Models at Scale via Coresets
arXiv:1703.08110
Abstract
How can we train a statistical mixture model on a massive data set? In this work we show how to construct coresets for mixtures of Gaussians. A coreset is a weighted subset of the data, which guarantees that models fitting the coreset also provide a good fit for the original data set. We show that, perhaps surprisingly, Gaussian mixtures admit coresets of size polynomial in dimension and the number of mixture components, while being independent of the data set size. Hence, one can harness computationally intensive algorithms to compute a good approximation on a significantly smaller data set. More importantly, such coresets can be efficiently constructed both in distributed and streaming settings and do not impose restrictions on the data generating process. Our results rely on a novel reduction of statistical estimation to problems in computational geometry and new combinatorial complexity results for mixtures of Gaussians. Empirical evaluation on several real-world datasets suggests that our coreset-based approach enables significant reduction in training-time with negligible approximation error.
References in corpus (9)
- Searching for Exotic Particles in High-Energy Physics with Deep Learning
- A Two-round Variant of EM for Gaussian Mixtures
- Polynomial Learning of Distribution Families
- Distributed k-Means and k-Median Clustering on General Topologies
- Settling the Polynomial Learnability of Mixtures of Gaussians
- Strong Coresets for Hard and Soft Bregman Clustering with Applications to Exponential Family Mixtures
- Tradeoffs for Space, Time, Data and Risk in Unsupervised Learning
- VC dimension of ellipsoids
- Linear-time Outlier Detection via Sensitivity