Task augmentation via channel mixture for few-task meta-learning

Open Webpage

Jiangdong Fan, Yuekeng Li, Jiayi Bi, Hui Xu, Jie Shao

Published: 2025, Last Modified: 30 Jan 2026Neural Networks 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•TACM ensures the continuity and integrity of channel feature information in task augmentation.•Established theoretical results are leveraged to validate the design of our method.•Experiments verify the effectiveness of TACM on multiple public datasets.