Accumulated decoupled learning with gradient staleness mitigation for convolutional neural networks

• Main Authors: Zhuang, Huiping, Weng, Zhenyu, Luo, Fulin, Toh, Kar-Ann, Li, Haizhou, Lin, Zhiping
• 其他作者: School of Electrical and Electronic Engineering
• 格式: Conference or Workshop Item
• 語言: English
• 出版: 2024
• 主題: Computer and Information Science; Convolutional neural networks; Delayed gradients-based methods
• 在線閱讀: https://hdl.handle.net/10356/174480; https://icml.cc/virtual/2021/index.html
• 機構: Nanyang Technological University

Gradient staleness is a major side effect in decoupled learning when training convolutional neural networks asynchronously. Existing methods that ignore this effect might result in reduced generalization and even divergence. In this paper, we propose an accumulated decoupled learning (ADL), wh...