Research map / Vision-language and multimodal / Cross-modal learning
rMMEA: Robust Multi-Modal Entity Alignment with Missing and Noise Visual Modality
Lingbing Guo, Zhuo Chen, Yichi Zhang, Wenbin Guo, Haonan Yang, Zhao Li, Zirui Chen, Xin Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add ranking-based distillation (ablation) | rMMEA · DBP15K ZH-EN (visual missing 0.95) | Hits@1: .723 -> .759 | Improved |
| add mutual information estimation (ablation) | rMMEA · DBP15K ZH-EN (visual missing 0.95) | Hits@1: .736 -> .759 | Improved |
| add other-modality missing/noisy training data (ablation) | rMMEA · DBP15K ZH-EN (visual missing 0.95) | Hits@1: .757 -> .759 | No clear effect |
| use ranking-based distillation instead of KL divergence distillation (ablation) | rMMEA · DBP15K ZH-EN (visual missing 0.95) | Hits@1: .724 -> .759 | Improved |
| use ranking-based distillation instead of KL divergence distillation (ablation) | rMMEA · OpenEA EN-FR (visual missing 0.95) | Hits@1: .607 -> .628 | Improved |
| add ranking-based distillation and MI estimation (rMMEA vs UMAEA) | anti-missing MMEA · DBP15K ZH-EN (visual missing 0.95) | Hits@1: .720 -> .759 | Improved |
| self-distillation instead of generative multi-stage training | rMMEA vs UMAEA · DBP15K ZH-EN training cost | training time: 1026.4 s vs 2542.8 s (rMMEA vs UMAEA, as ordered in Figure 4) | Improved |
| increase visual missing ratio | MCLEA · DBP15K ZH-EN | Hits@1 | Mixed |
| add chaotic (zero/random) data to other modalities | rMMEA vs UMAEA · four datasets (rvm 0.95) | MRR: gain over UMAEA rises from 9.9% to 27.9% (DBP15K ZH-EN) | Improved |
Closest papers by meaning
- Tackling Uncertain Correspondences for Multi-Modal Entity Alignment (NeurIPS 2024 · similarity 0.92)
- Pseudo-Label Calibration Semi-supervised Multi-Modal Entity Alignment (AAAI 2024 · similarity 0.91)
- Multiple Heads are Better than One: Mixture of Modality Knowledge Experts for Entity Representation Learning (ICLR 2025 · similarity 0.88)
- MESED: A Multi-Modal Entity Set Expansion Dataset with Fine-Grained Semantic Classes and Hard Negative Entities (AAAI 2024 · similarity 0.87)
- A Dual-Way Enhanced Framework from Text Matching Point of View for Multimodal Entity Linking (AAAI 2024 · similarity 0.87)
- Tokenization, Fusion, and Augmentation: Towards Fine-grained Multi-modal Entity Representation (AAAI 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
