#artificial intelligence
Sep 2025
Align Where the Words Look: Cross-Attention-Guided Patch Alignment with Contrastive and Transport Regularization for Bengali Captioning
A compute-aware Bengali captioning pipeline trained on LaBSE-verified EN-BN pairs and 110k bilingual-prompted synthetic images, with a core novelty that tightens real-synthetic patch alignment, reduces spurious matches, and drives strong gains on Flickr.
Riad Ahmed Anonto, Sardar Md. Saffat Zabin, M. S. Rahman
· arXiv.org · 2 citations