ART adapts VLMs for visual relation classification via instruction tuning with adaptive, uncertainty-based instance selection, improving generalization to unseen and rare relations.
Flamingo: A visual language model for few-shot learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ART: Adaptive Relation Tuning for Generalized Relation Prediction
ART adapts VLMs for visual relation classification via instruction tuning with adaptive, uncertainty-based instance selection, improving generalization to unseen and rare relations.