Enhancing Progressive Ensemble Learning via Normalized Extra-Gradient Initialization
Published in Neural Networks, 2026
This work develops Normalized Extra-Gradient Initialization to accelerate progressive training of sparse MoE models.
Recommended citation: Zheshun Wu, Yu Pan, Dun Zeng, Zenglin Xu, Qifan Wang, and Jie Liu. (2026). Enhancing Progressive Ensemble Learning via Normalized Extra-Gradient Initialization. Neural Networks.