dmlrs - Fast Double / Debiased Machine Learning Implemented in Rust

Double / debiased machine learning (Chernozhukov et al., 2018, <doi:10.1111/ectj.12097>) for causal inference: partially linear, interactive, instrumental variable and difference-in-differences models with cross-fitted nuisance functions, multiplier bootstrap and sensitivity analysis for omitted confounding. Estimation, including the nuisance learners (lasso, logistic regression, random forests, gradient boosting), runs in Rust through the 'dml-core' crate, in parallel. Results match the Python package 'dml-rs', which is validated against 'DoubleML'.