![[2602.17543] genriesz: A Python Package for Automatic Debiased Machine Learning with Generalized Riesz Regression](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.17543] genriesz: A Python Package for Automatic Debiased Machine Learning with Generalized Riesz Regression
Summary
The article presents 'genriesz', an open-source Python package designed for automatic debiased machine learning using generalized Riesz regression, facilitating efficient estimation of causal parameters.
Why It Matters
This package addresses the growing need for robust machine learning tools that can automate complex statistical estimations, making advanced methodologies more accessible to practitioners and researchers in the field. Its focus on debiasing enhances the reliability of machine learning outcomes, which is crucial for applications in econometrics and causal inference.
Key Takeaways
- genriesz automates debiased machine learning using generalized Riesz regression.
- The package includes features like covariate balancing and density ratio estimation.
- It provides a modular interface for various model specifications and estimators.
- Key applications include estimating average treatment effects and marginal effects.
- The package is available on PyPI, promoting accessibility for users.
Statistics > Machine Learning arXiv:2602.17543 (stat) [Submitted on 19 Feb 2026] Title:genriesz: A Python Package for Automatic Debiased Machine Learning with Generalized Riesz Regression Authors:Masahiro Kato View a PDF of the paper titled genriesz: A Python Package for Automatic Debiased Machine Learning with Generalized Riesz Regression, by Masahiro Kato View PDF HTML (experimental) Abstract:Efficient estimation of causal and structural parameters can be automated using the Riesz representation theorem and debiased machine learning (DML). We present genriesz, an open-source Python package that implements automatic DML and generalized Riesz regression, a unified framework for estimating Riesz representers by minimizing empirical Bregman divergences. This framework includes covariate balancing, nearest-neighbor matching, calibrated estimation, and density ratio estimation as special cases. A key design principle of the package is automatic regressor balancing (ARB): given a Bregman generator $g$ and a representer model class, genriesz} automatically constructs a compatible link function so that the generalized Riesz regression estimator satisfies balancing (moment-matching) optimality conditions in a user-chosen basis. The package provides a modulr interface for specifying (i) the target linear functional via a black-box evaluation oracle, (ii) the representer model via basis functions (polynomial, RKHS approximations, random forest leaf encodings, neural embeddings, and ...
