Gradient descent inference in empirical risk minimization
Fuente:
arXiv
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911272324300800 |
|---|---|
| author | Han, Qiyang Xu, Xiaocong |
| author_facet | Han, Qiyang Xu, Xiaocong |
| contents | Gradient descent is one of the most widely used iterative algorithms in modern statistical learning. However, its precise algorithmic dynamics in high-dimensional settings remain only partially understood, which has limited its broader potential for statistical inference applications.
This paper provides a precise, non-asymptotic joint distributional characterization of gradient descent iterates and their debiased statistics in a broad class of empirical risk minimization problems, in the so-called mean-field regime where the sample size is proportional to the signal dimension. Our non-asymptotic state evolution theory holds for both general non-convex loss functions and non-Gaussian data, and reveals the central role of two Onsager correction matrices that precisely characterize the non-trivial dependence among all gradient descent iterates in the mean-field regime.
Leveraging the joint state evolution characterization, we show that the gradient descent iterate retrieves approximate normality after a debiasing correction via a linear combination of all past iterates, where the debiasing coefficients can be estimated by the proposed gradient descent inference algorithm. This leads to a new algorithmic statistical inference framework based on debiased gradient descent, which (i) applies to a broad class of models with both convex and non-convex losses, (ii) remains valid at each iteration without requiring algorithmic convergence, and (iii) exhibits a certain robustness to possible model misspecification. As a by-product, our framework also provides algorithmic estimates of the generalization error at each iteration. As canonical examples, we demonstrate our theory and inference methods in the single-index regression model and a generalized logistic regression model, where the natural loss functions may exhibit arbitrarily non-convex landscapes. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_09498 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Gradient descent inference in empirical risk minimization Han, Qiyang Xu, Xiaocong Statistics Theory Information Theory Optimization and Control Methodology Machine Learning Gradient descent is one of the most widely used iterative algorithms in modern statistical learning. However, its precise algorithmic dynamics in high-dimensional settings remain only partially understood, which has limited its broader potential for statistical inference applications. This paper provides a precise, non-asymptotic joint distributional characterization of gradient descent iterates and their debiased statistics in a broad class of empirical risk minimization problems, in the so-called mean-field regime where the sample size is proportional to the signal dimension. Our non-asymptotic state evolution theory holds for both general non-convex loss functions and non-Gaussian data, and reveals the central role of two Onsager correction matrices that precisely characterize the non-trivial dependence among all gradient descent iterates in the mean-field regime. Leveraging the joint state evolution characterization, we show that the gradient descent iterate retrieves approximate normality after a debiasing correction via a linear combination of all past iterates, where the debiasing coefficients can be estimated by the proposed gradient descent inference algorithm. This leads to a new algorithmic statistical inference framework based on debiased gradient descent, which (i) applies to a broad class of models with both convex and non-convex losses, (ii) remains valid at each iteration without requiring algorithmic convergence, and (iii) exhibits a certain robustness to possible model misspecification. As a by-product, our framework also provides algorithmic estimates of the generalization error at each iteration. As canonical examples, we demonstrate our theory and inference methods in the single-index regression model and a generalized logistic regression model, where the natural loss functions may exhibit arbitrarily non-convex landscapes. |
| title | Gradient descent inference in empirical risk minimization |
| topic | Statistics Theory Information Theory Optimization and Control Methodology Machine Learning |
| url | https://arxiv.org/abs/2412.09498 |