Symbol Grounding Association in Multimodal Sequences with Missing Elements
2015·Arxiv