Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2023
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2311.05589 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866916654030520320 |
|---|---|
| author | Yin, Yida Xu, Zhiqiu Li, Zhiyuan Darrell, Trevor Liu, Zhuang |
| author_facet | Yin, Yida Xu, Zhiqiu Li, Zhiyuan Darrell, Trevor Liu, Zhuang |
| contents | Stochastic Variance Reduced Gradient (SVRG), introduced by Johnson & Zhang (2013), is a theoretically compelling optimization method. However, as Defazio & Bottou (2019) highlight, its effectiveness in deep learning is yet to be proven. In this work, we demonstrate the potential of SVRG in optimizing real-world neural networks. Our empirical analysis finds that, for deeper neural networks, the strength of the variance reduction term in SVRG should be smaller and decrease as training progresses. Inspired by this, we introduce a multiplicative coefficient $α$ to control the strength and adjust it through a linear decay schedule. We name our method $α$-SVRG. Our results show $α$-SVRG better optimizes models, consistently reducing training loss compared to the baseline and standard SVRG across various model architectures and multiple image classification datasets. We hope our findings encourage further exploration into variance reduction techniques in deep learning. Code is available at github.com/davidyyd/alpha-SVRG. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2311_05589 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | A Coefficient Makes SVRG Effective Yin, Yida Xu, Zhiqiu Li, Zhiyuan Darrell, Trevor Liu, Zhuang Machine Learning Optimization and Control Stochastic Variance Reduced Gradient (SVRG), introduced by Johnson & Zhang (2013), is a theoretically compelling optimization method. However, as Defazio & Bottou (2019) highlight, its effectiveness in deep learning is yet to be proven. In this work, we demonstrate the potential of SVRG in optimizing real-world neural networks. Our empirical analysis finds that, for deeper neural networks, the strength of the variance reduction term in SVRG should be smaller and decrease as training progresses. Inspired by this, we introduce a multiplicative coefficient $α$ to control the strength and adjust it through a linear decay schedule. We name our method $α$-SVRG. Our results show $α$-SVRG better optimizes models, consistently reducing training loss compared to the baseline and standard SVRG across various model architectures and multiple image classification datasets. We hope our findings encourage further exploration into variance reduction techniques in deep learning. Code is available at github.com/davidyyd/alpha-SVRG. |
| title | A Coefficient Makes SVRG Effective |
| topic | Machine Learning Optimization and Control |
| url | https://arxiv.org/abs/2311.05589 |