SAMA is a unified augmentation framework using semantic anchors, CME-MLLM with adapters, anchor-preserving diffusion, and dual-constraint filtering to improve MNER, MRE, and MEE in low-resource settings.
Chain-of-thought prompt distillation for mul- timodal named entity recognition and multimodal relation extraction,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SAMA: Semantic Anchor-aligned Augmentation for Unified Low-Resource Multimodal Information Extraction
SAMA is a unified augmentation framework using semantic anchors, CME-MLLM with adapters, anchor-preserving diffusion, and dual-constraint filtering to improve MNER, MRE, and MEE in low-resource settings.