English

Analytical Correction for Subsampling Bias in Drifting Models

Machine Learning 2026-05-01 v1

Abstract

Drifting models are capable one-step generative models trained to follow a drifting field. The field combines attractive and repulsive softmax-weighted centroids over the data and current-generator distributions. In practice, only a minibatch of nn samples from each distribution is available, and each centroid is approximated by an empirical estimate. In this paper, we begin by showing that the minibatch centroid is in general a biased estimator of the target centroid, with a pointwise O(1/n)O(1/n) bias arising from softmax self-normalization. Correcting this bias requires the expectation over the full distribution, which is intractable. We instead approximate the leading bias term from in-batch statistics and propose Analytical Bias Correction (ABC), a closed-form plug-in adjustment. We prove that ABC reduces the bias from O(1/n)O(1/n) to O(1/n2)O(1/n^2), introduces no first-order increase in total variance, and preserves convex-hull containment of the corrected centroid. In practice, ABC requires only two additional lines of code and has negligible wall-time overhead under compiled execution. Toy experiments confirm the theoretical O(1/n)O(1/n) and O(1/n2)O(1/n^2) scaling. On CIFAR-10, ABC reduces FID and trains faster, with the largest gains at small nn, where the bias is most significant.

Cite

@article{arxiv.2604.27239,
  title  = {Analytical Correction for Subsampling Bias in Drifting Models},
  author = {Jiaru Zhang and Zeyun Deng and Juanwu Lu and Ziran Wang and Ruqi Zhang},
  journal= {arXiv preprint arXiv:2604.27239},
  year   = {2026}
}
R2 v1 2026-07-01T12:42:29.598Z