Found 1 document(s)
Page 1 of 1
-
Video Question Grounding (VideoQG) requires both accurate question answering and precise temporal localization. Existing supervised methods rely on costly temporal annotations, while existing zero-shot approaches are pri...