Joint Multimodal Embedding and Backtracking Search in Vision-and-Language Navigation
Due to the development of computer vision and natural language processing technologies in recent years, there has been a growing interest in multimodal intelligent tasks that require the ability to concurrently understand various forms of input data such as Grooming images and text.Vision-and-language navigation (VLN) require the alignment and grou