Multimodal entity linking models are vulnerable to visual adversarial perturbations, and the proposed retrieval-augmented LLM method (LLM-RetLink) reportedly improves accuracy by 0.4% to 35.7%.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
On Evaluating the Adversarial Robustness of Foundation Models for Multimodal Entity Linking
Multimodal entity linking models are vulnerable to visual adversarial perturbations, and the proposed retrieval-augmented LLM method (LLM-RetLink) reportedly improves accuracy by 0.4% to 35.7%.