UNIMO: Towards Unified-Modal Understanding and Generation via Cross-Modal Contrastive Learning
TowardsUnified-ModalUnderstandingandGenerationviaCross-ModalContrastiveLearning(通过跨模态对比学习实现统一模态的理解和生成)1.现有方法的不足:目前的方法要么只能专注于单模态的任务,要么只能专注于多模态任务,而不能将这二者有机地统一在一起互相补充2.创新点:提出UNIMO这一统一模态的预训练架构