【论文笔记】Towards Corrective Deep Imitation Learning in Data Intensive Environments: Helping robots to learn faster by leveraging human knowledge Abstract Interactive imitation learning refers to learni

这是一篇2019CVPR（oral）的文章，也是我几个月前看的，有人说这篇文章有点水，给CVPR颁“金酸莓奖”，知乎网友热议最差论文，战火烧到Reddit论坛，所以我当时也没认真看了。 论文地址：Accel: A Corrective Fusion Network forEfficient Semantic Segmentation on Video code：https://github.co