I want to talk about a gradient boosting algorithm which I developed in Rust. It is different from other GBM algorithms because it doesn't need hyperparameter tuning. github.com/perpetual-ml/perpetual