Grounding Language Models for Visual Entity Recognition: AutoVER Framework

The author introduces the AutoVER framework, utilizing retrieval-augmented constrained generation to enhance visual entity recognition performance significantly.