EVIDENT: Routing MLLM Adaptation through Entity-Grounded Visual Evidence for Cross-Domain Video Temporal Grounding 文章

ArXiv CS.CV2026-05-26NEWSen作者: Geo Ahn, Jiwook Han, Youngrae Kim, Joonseok Lee, Jinwoo Choi
查看原文 →
关系图谱

EVIDENT: Routing MLLM Adaptation through Entity-Grounded Visual Evidence for Cross-Domain Video Temporal Grounding · 相关技术

相关技术

LLMEVIDENTVTGEntity-Binding Distillation lossEntity Bottleneck AdapterVideo Temporal Grounding