MEAformer: Multi-modal Entity Alignment Transformer for Meta Modality Hybrid
This work improves multi-modal entity alignment for knowledge graph integration by enhancing robustness against noise, but it is incremental as it builds on existing transformer-based methods with a novel fusion strategy.
The paper tackles the problem of multi-modal entity alignment across knowledge graphs by addressing the limitations of existing KG-level modality fusion strategies, which ignore entity-specific modality preferences and are vulnerable to noisy data. The proposed MEAformer model dynamically predicts modality correlations for fine-grained fusion, achieving state-of-the-art performance in various training scenarios with efficient runtime and interpretability.
Multi-modal entity alignment (MMEA) aims to discover identical entities across different knowledge graphs (KGs) whose entities are associated with relevant images. However, current MMEA algorithms rely on KG-level modality fusion strategies for multi-modal entity representation, which ignores the variations of modality preferences of different entities, thus compromising robustness against noise in modalities such as blurry images and relations. This paper introduces MEAformer, a multi-modal entity alignment transformer approach for meta modality hybrid, which dynamically predicts the mutual correlation coefficients among modalities for more fine-grained entity-level modality fusion and alignment. Experimental results demonstrate that our model not only achieves SOTA performance in multiple training scenarios, including supervised, unsupervised, iterative, and low-resource settings, but also has a limited number of parameters, efficient runtime, and interpretability. Our code is available at https://github.com/zjukg/MEAformer.