Most meta-learning methods assume that the (very small) context set used to
establish a new task at test time is passively provided. In some settings,
however, it is feasible to actively select which points to la
提出了一种名为 Learning To Sample 框架的新型基于学习的主动学习方法,它能够通过不确定性和多样性的优化整合来主动选择最具代表性和信息量的样本,并在图像分类、薪资水平预测和实体解析等三个任务上验证了其有效性,并且特别适用于类别高度不平衡的数据集,还能够有效解决现有主动学习方法中出现的冷启动问题。