Despite recent advances of AI, story understanding remains an open and under-investigated problem. We collect, preprocess, and publicly release a video-language story dataset, Synopses of Movie Narratives (SYMON), containing 5,193 video summaries of popular movies and TV series. SYMON captures naturalistic story-telling videos for human audience made by human creators. As a prototypical and naturalistic story dataset, SYMON features high coverage of multimodal story events, abundant mental-state descriptions, and large semantic gaps between the visual and the textual modalities. We establish benchmarks on video-text retrieval and zero-shot alignment on movie summary videos, which showcase the importance of in-domain data in story understanding. With SYMON, we hope to lay the groundwork for progress in multimodal story understanding.
View on arXiv