PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
June 26, 2026ElectronicsOpen Access

A Contrastive and Uncertainty–Aware Framework for Multimodal Named Entity Recognition

View Full Paper
Ask AI
Bookmark
Share

Authors

XYXiao YangRZRuixue ZhaoHLHonglei Li

Discussion

Loading...

Member takes

Overview

Randomized trial demonstrates improved entity detection in social media texts, suggesting enhanced MNER performance.

Key Points

  • This study aims to enhance multimodal named entity recognition by addressing issues in text-image alignment and entity representation.
  • Proposed a contrastive uncertainty-aware framework (CUA-MNER) for MNER.
  • Implemented hierarchical vision-text alignment for improved token, phrase, and sentence correspondences.
  • Used variational uncertainty-aware fusion to manage modality contributions and enhance entity recognition.
  • Achieved F1 scores of 76.97% and 89.66% on Twitter2015 and Twitter2017 benchmarks, respectively.
  • Outperformed competitive baselines by 0.66 and 1.95 F1 points.
  • Identified that the model's components provide complementary benefits, indicating robustness in multimodal recognition.

Cite This Study

Yang et al. (2026) studied this question.

synapsesocial.com/papers/6a3e17d3030ad1a9b3091310https://doi.org/10.3390/electronics15132770
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Visual Clue Guidance and Consistency Matching Framework for Multimodal Named Entity Recognition2024 · 3 citations
  2. 2Multimodal Named-Entity Recognition Based on Symmetric Fusion with Contrastive Learning2026
  3. 3VEC-MNER: Hybrid Transformer with Visual-Enhanced Cross-Modal Multi-level Interaction for Multimodal NER2024 · 13 citations
  4. 4Multimodal Named Entity Recognition with Prior Knowledge from Multimodal Large Models and Text-Directed Fusion2026
  5. 52M-NER: Contrastive Learning for Multilingual and Multimodal NER with Language and Modal Fusion2024 · 1 citations