Volume 4, Issue 3
Optimistic Estimate Uncovers the Potential of Nonlinear Models

Yaoyu Zhang, Zhongwang Zhang, Leyang Zhang, Zhiwei Bai, Tao Luo & Zhi-Qin John Xu

J. Mach. Learn. , 4 (2025), pp. 192-222.

Published online: 2025-09

[An open-access article; the PDF is free to any online user.]

Export citation
  • Abstract

We propose an optimistic estimate framework to evaluate the potential of nonlinear models in fitting target functions at overparameterization. In our framework, such potential is quantified by estimating the smallest possible sample size needed for a model to recover a target function, referred to as an optimistic sample size. Following the framework, we derive the optimistic sample sizes for matrix factorization models, deep models, and two-layer neural networks (NNs) with fully-connected or convolutional architectures. For each nonlinear model, we confirm via experiments that the target functions can be fitted at overparameterization as predicted by our analysis. Our results suggest a hierarchical inductive bias of nonlinear models towards simple functions with smaller optimistic sample sizes intrinsic to their architecture. The dynamical realization of the suggested inductive bias remains an open problem for the further study.

  • AMS Subject Headings

  • Copyright

COPYRIGHT: © Global Science Press

  • Email address
  • BibTex
  • RIS
  • TXT
@Article{JML-4-192, author = {Zhang , YaoyuZhang , ZhongwangZhang , LeyangBai , ZhiweiLuo , Tao and Xu , Zhi-Qin John}, title = {Optimistic Estimate Uncovers the Potential of Nonlinear Models}, journal = {Journal of Machine Learning}, year = {2025}, volume = {4}, number = {3}, pages = {192--222}, abstract = {

We propose an optimistic estimate framework to evaluate the potential of nonlinear models in fitting target functions at overparameterization. In our framework, such potential is quantified by estimating the smallest possible sample size needed for a model to recover a target function, referred to as an optimistic sample size. Following the framework, we derive the optimistic sample sizes for matrix factorization models, deep models, and two-layer neural networks (NNs) with fully-connected or convolutional architectures. For each nonlinear model, we confirm via experiments that the target functions can be fitted at overparameterization as predicted by our analysis. Our results suggest a hierarchical inductive bias of nonlinear models towards simple functions with smaller optimistic sample sizes intrinsic to their architecture. The dynamical realization of the suggested inductive bias remains an open problem for the further study.

}, issn = {2790-2048}, doi = {https://doi.org/10.4208/jml.231109}, url = {http://global-sci.org/intro/article_detail/jml/24380.html} }
TY - JOUR T1 - Optimistic Estimate Uncovers the Potential of Nonlinear Models AU - Zhang , Yaoyu AU - Zhang , Zhongwang AU - Zhang , Leyang AU - Bai , Zhiwei AU - Luo , Tao AU - Xu , Zhi-Qin John JO - Journal of Machine Learning VL - 3 SP - 192 EP - 222 PY - 2025 DA - 2025/09 SN - 4 DO - http://doi.org/10.4208/jml.231109 UR - https://global-sci.org/intro/article_detail/jml/24380.html KW - Nonlinear regression, Optimistic sample size estimate, Matrix factorization model, Two-layer neural network, Hierarchical inductive bias. AB -

We propose an optimistic estimate framework to evaluate the potential of nonlinear models in fitting target functions at overparameterization. In our framework, such potential is quantified by estimating the smallest possible sample size needed for a model to recover a target function, referred to as an optimistic sample size. Following the framework, we derive the optimistic sample sizes for matrix factorization models, deep models, and two-layer neural networks (NNs) with fully-connected or convolutional architectures. For each nonlinear model, we confirm via experiments that the target functions can be fitted at overparameterization as predicted by our analysis. Our results suggest a hierarchical inductive bias of nonlinear models towards simple functions with smaller optimistic sample sizes intrinsic to their architecture. The dynamical realization of the suggested inductive bias remains an open problem for the further study.

Zhang , YaoyuZhang , ZhongwangZhang , LeyangBai , ZhiweiLuo , Tao and Xu , Zhi-Qin John. (2025). Optimistic Estimate Uncovers the Potential of Nonlinear Models. Journal of Machine Learning. 4 (3). 192-222. doi:10.4208/jml.231109
Copy to clipboard
The citation has been copied to your clipboard