2 papers
cs.AI2026
Whose record is this? Diagnosing and authorizing record use in personalized multimodal models
Xinyu Mao, Junsi Li, Chenyang Liu +2
Contextualized visual personalization can retrieve a true record yet apply it to the wrong visual subject. We formalize when a record may condition an answer as \emph{record author…
cs.CV2025
SEPS: Semantic-enhanced Patch Slimming Framework for fine-grained cross-modal alignment
Xinyu Mao, Junsi Li, Haoji Zhang +2
Fine-grained cross-modal alignment aims to establish precise local correspondences between vision and language, forming a cornerstone for visual question answering and related mult…