Look Closer to Ground Better: Weakly-Supervised Temporal Grounding of Sentence in Video
2020·Arxiv