BriefGPT.xyz
Mar, 2023
忘却是否是正向迁移的良好归纳偏差?
Is forgetting less a good inductive bias for forward transfer?
HTML
PDF
Jiefeng Chen, Timothy Nguyen, Dilan Gorur, Arslan Chaudhry
TL;DR
研究探讨了持续学习算法中前向传递的表现力,并考察了连续学习算法在图像分类领域的表现。结果发现,保留过去信息的表现力更好,同时这种表现力更加多样和可辨识。
Abstract
One of the main motivations of studying
continual learning
is that the problem setting allows a model to accrue knowledge from past tasks to learn new tasks more efficiently. However, recent studies suggest that the key metric that
→