1 paper · 1 filter
Aodi Li, Liansheng Zhuang, Xiao Long +2
Domain generalization aims to learn a model from multiple training domains and generalize it to unseen test domains. Recent theory has shown that seeking the deep models, whose par…