VRCMC: a semi-supervised method for multi-modal entity alignment with visual refinement and cross-modal calibration
Abstract The goal of multi-modal entity alignment (MMEA) is to automatically discover entities across different multi-modal knowledge graphs (MMKGs) that refer to the same real-world object. However, because of the complexity and diversity of data sources, entities in MMKGs often lack images or poss...
I tiakina i:
| Ngā kaituhi matua: | , , |
|---|---|
| Hōputu: | Artigo |
| Reo: | Inglês |
| I whakaputaina: |
Springer
2026-01-01
|
| Rangatū: | Journal of King Saud University: Computer and Information Sciences |
| Ngā marau: | |
| Urunga tuihono: | https://doi.org/10.1007/s44443-025-00443-2 |
| Ngā Tūtohu: |
Kāore He Tūtohu, Me noho koe te mea tuatahi ki te tūtohu i tēnei pūkete!
|