Mitch Wyle

Mitch Wyle
Research Scientist

Mitch Wyle, PhD, studied computer science at ETH Zürich. His dissertation led to a patent in fast, automated problem resolution. Mitch is now leading an applied research team building machine learned predictive models.

2015 International Conference for Machine Learning (ICML)

Bayesian and Empirical Bayesian Forests

Matt Taddy, Chun-Sheng Chen, Jun Yu, Mitch Wyle

We derive ensembles of decision trees through a nonparametric Bayesian model, allowing us to view random forests as samples from a posterior distribution. This insight provides large gains in interpretability, and motivates a class of Bayesian forest (BF) algorithms that yield small but reliable performance gains.

Based on the BF framework, we are able to show that high-level tree hierarchy is stable in large samples. This leads to an empirical Bayesian forest (EBF) algorithm for building approximate BFs on massive distributed datasets and we show that EBFs outperform subsampling based alternatives by a large margin.