Instruction-Tuned From survey PDF — verify links
VAU-R1
VAU-R1 — Anomaly Reasoning
arXiv · 2025Human-reviewed summary status
Survey summary
VAU-R1 uses reinforcement fine-tuning on MLLMs to improve anomaly reasoning, decomposing understanding into multiple-choice QA, temporal grounding, and reasoning tasks; paired with VAU-Bench, the first chain-of-thought benchmark for video anomaly reasoning with rationales and temporal annotations.
Taxonomy placement
Instruction-Tuned
Instruction-TunedAnomaly Reasoning
Method characteristics
Instruction-TunedReinforcement fine-tuningGRPOMulti-task