Generalized and Optimal Straight-Through Estimators

James Hooper, Alexander ShekhovtsovAISTATS 2026 · OpenReview

Abstract

Modern ML models often utilize discrete components within their computational graphs, making training challenging. In such cases, approximate-chain-rule gradient estimators can be applied. They work reasonably well but are obtained by combining diverse rationales with ad-hoc choices. In this work, we propose a principled axiomatic approach to define a general family of gradient estimators and show that it subsumes many existing methods. Within this family, we derive optimal estimators with respect to a minimum variance criterion subject to interpretable bias-limiting constraints, addressing integer and one-hot categorical discrete variables. We empirically demonstrate that our estimator can achieve a better bias-variance trade-off than existing ones on synthetic problems and outperforms them on training variational auto-encoders with discrete latent variables.
BibTeX
@inproceedings{hooper2026generalized,
  author = {James Hooper and Alexander Shekhovtsov},
  title = {{Generalized and Optimal Straight-Through Estimators}},
  booktitle = {International Conference on Artificial Intelligence and Statistics},
  year = {2026},
  url = {https://openreview.net/forum?id=5LZQfHUVER},
}

Loading related content...

Figure