Learning Bounds for Open-Set Learning

             Zhen Fang * 1 Jie Lu 1 Anjin Liu * 1 Feng Liu 1 Guangquan Zhang 1

             Abstract                 et al., 2018; Yang et al., 2020).
   Traditional supervised learning aims to train a        However, the closed-set assumption is not realistic during
   classifier in the closed-set world, where training      the testing phase (i.e., there are no labels in the samples)
   and test samples share the same label space. In        since it is no ...