Adversarial random forests for density estimation and generative modelling

We propose methods for density estimation and data synthesis using a novel form of unsupervised random forests. Inspired by generative adversarial networks, we implement a recursive procedure in which trees gradually learn structural properties of the data through alternating rounds of generation and discrimination. The method is provably consistent under minimal assumptions. Unlike existing tree-based alternatives, our approach provides smooth unconditional densities and allows for fully synthetic data generation. We achieve comparable or superior performance to state-of-the-art deep learning models on various tabular data benchmarks while executing about two orders of magnitude faster on average. All algorithms are implemented in easy-to-use $\texttt{R}$ and Python packages.