The Dimension Strikes Back with Gradients: Generalization of Gradient Methods in Stochastic Convex Optimization
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_ | 1866929218426765312 |
|---|---|
| author | Schliserman, Matan Sherman, Uri Koren, Tomer |
| author_facet | Schliserman, Matan Sherman, Uri Koren, Tomer |
| contents | We study the generalization performance of gradient methods in the fundamental stochastic convex optimization setting, focusing on its dimension dependence. First, for full-batch gradient descent (GD) we give a construction of a learning problem in dimension $d=O(n^2)$, where the canonical version of GD (tuned for optimal performance of the empirical risk) trained with $n$ training examples converges, with constant probability, to an approximate empirical risk minimizer with $Ω(1)$ population excess risk. Our bound translates to a lower bound of $Ω(\sqrt{d})$ on the number of training examples required for standard GD to reach a non-trivial test error, answering an open question raised by Feldman (2016) and Amir, Koren, and Livni (2021b) and showing that a non-trivial dimension dependence is unavoidable. Furthermore, for standard one-pass stochastic gradient descent (SGD), we show that an application of the same construction technique provides a similar $Ω(\sqrt{d})$ lower bound for the sample complexity of SGD to reach a non-trivial empirical error, despite achieving optimal test performance. This again provides an exponential improvement in the dimension dependence compared to previous work (Koren, Livni, Mansour, and Sherman, 2022), resolving an open question left therein. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_12058 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | The Dimension Strikes Back with Gradients: Generalization of Gradient Methods in Stochastic Convex Optimization Schliserman, Matan Sherman, Uri Koren, Tomer Machine Learning We study the generalization performance of gradient methods in the fundamental stochastic convex optimization setting, focusing on its dimension dependence. First, for full-batch gradient descent (GD) we give a construction of a learning problem in dimension $d=O(n^2)$, where the canonical version of GD (tuned for optimal performance of the empirical risk) trained with $n$ training examples converges, with constant probability, to an approximate empirical risk minimizer with $Ω(1)$ population excess risk. Our bound translates to a lower bound of $Ω(\sqrt{d})$ on the number of training examples required for standard GD to reach a non-trivial test error, answering an open question raised by Feldman (2016) and Amir, Koren, and Livni (2021b) and showing that a non-trivial dimension dependence is unavoidable. Furthermore, for standard one-pass stochastic gradient descent (SGD), we show that an application of the same construction technique provides a similar $Ω(\sqrt{d})$ lower bound for the sample complexity of SGD to reach a non-trivial empirical error, despite achieving optimal test performance. This again provides an exponential improvement in the dimension dependence compared to previous work (Koren, Livni, Mansour, and Sherman, 2022), resolving an open question left therein. |
| title | The Dimension Strikes Back with Gradients: Generalization of Gradient Methods in Stochastic Convex Optimization |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2401.12058 |