CLIP-BCA-Gated: A Dynamic Multimodal Framework for Real-Time Humanitarian Crisis Classification with Bi-Cross-Attention and Adaptive Gating
During humanitarian crises, social media generates over 30 million multimodal tweets daily, but 20% textual noise, 40% cross-modal misalignment, and severe class imbalance (4.1% rare classes) hinder effective classification. This study presents CLIP-BCA-Gated, a dynamic multimodal framework that int...
Saved in:
| Main Authors: | Shanshan Li, Qingjie Liu, Zhian Pan, Xucheng Wu |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-08-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/15/8758 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Similar Items
-
Stock movement prediction with multimodal stable fusion via gated cross-attention mechanism
by: Chang Zong, et al.
Published: (2025-07-01) -
Cross-modal gated feature enhancement for multimodal emotion recognition in conversations
by: Shiyun Zhao, et al.
Published: (2025-08-01) -
A Dual-Enhanced Hierarchical Alignment Framework for Multimodal Named Entity Recognition
by: Jian Wang, et al.
Published: (2025-05-01) -
Transferring CLIP for visual grounding in remote sensing images
by: Linlin Liang, et al.
Published: (2025-08-01) -
Revealing Depression Through Social Media via Adaptive Gated Cross-Modal Fusion Augmented With Insights From Personality Traits
by: Gede Aditra Pradnyana, et al.
Published: (2025-01-01)