We are thrilled to announce that our paper “R-MMA: Enhancing Vision-Language Models with Recurrent Adapters for Few-Shot and Cross-Domain Generalization” has been accepted at WACV 2026, one of the premier conferences in Computer Vision (CORE Rank A). Congratulations to Md Fahim and Mir Sazzat Hossain (Research Assistants, CCDS). 
This work introduces R-MMA, a lightweight and highly parameter-efficient adapter designed to enhance few-shot and cross-domain generalization in Vision-Language Models, such as CLIP. R-MMA aligns and refines frozen encoder features through a unified attention-driven representation, achieving state-of-the-art performance across base-to-novel, cross-dataset, and domain generalization benchmarks.
Congratulations to all co-authors, collaborators, and supervisors 
