作者位次 #3/3ACL 2026Findings· Findings of ACL
DICA : Dual-Indicator Guided Contrastive Alignment in Multimodal Large Language Models
Hao Yang · Jin Wang · Xuejie Zhang
摘要
Human visual reasoning typically follows a coarse-to-fine attention process, starting from global scene understanding and gradually focusing on question-relevant regions. However, multimodal large language models may deviate from this pattern due to attention drift and the underutilization of vis...