核心概念
GenView improves self-supervised learning by enhancing positive view quality through pretrained generative models and a quality-driven contrastive loss.
統計資料
Self-supervised learning has achieved remarkable success in acquiring high-quality representations from unlabeled data.
GenView significantly improves MoCov2 performance by 2.5%/2.2% on ImageNet linear/semi-supervised classification.
Moreover, GenView performs much better than naively augmenting the ImageNet dataset with Laion400M or ImageNet21K.