Visual prompts on images replace textual mention words in a new multimodal entity linking task, with a new dataset (VPWiki) and a framework (IIER) that achieves the best reported results on it.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
VP-MEL: Visual Prompts Guided Multimodal Entity Linking
Visual prompts on images replace textual mention words in a new multimodal entity linking task, with a new dataset (VPWiki) and a framework (IIER) that achieves the best reported results on it.