May, 2024
沃瑟斯坦梯度提升:一种应用于后验回归的通用框架
Wasserstein Gradient Boosting: A General Framework with Applications to
Posterior Regression
TL;DRWasserstein gradient boosting is a novel ensemble method that uses Wasserstein gradient to approximate a target probability distribution and produce a distributional estimate of the output-distribution parameter, outperforming existing methods in probabilistic prediction.