Diffusion Boosted Trees

Xizewen Han, Mingyuan Zhou

Combining the merits of both denoising diffusion probabilistic models and gradient boosting, the diffusion boosting paradigm is introduced for tackling supervised learning problems…