Instruction-Tuned From survey PDF — verify links
HAWK
HAWK — Open-World Understanding
NeurIPS · 2024Human-reviewed summary status
Survey summary
HAWK integrates motion modality into a VLM via an auxiliary consistency loss and explicit motion-to-language supervision, trained on over 8,000 anomaly videos with language descriptions plus 8,000 QA pairs, achieving state-of-the-art video description generation and question-answering performance.
Taxonomy placement
Instruction-Tuned
Instruction-TunedOpen-World Understanding
Method characteristics
Instruction-TunedMotion-languageMulti-dataset QA