Conference
Aug 2026
Adaptive Cross-Modal Fusion With Instance-Level Gating for Vision-Language Understanding
Multimodal deep learning integrates heterogeneous data sources such as images and text to enable machines to understand complex real-world contexts. Although recent vision-language models have achieved significant progress, most existing approaches rely on rigid fusion strategies that combine modalities either at early...
Unnati A. Patel, Sanskruti Patel, J. Nanavati et al.
· International Conference on... · 0 citations