brAIn, Sungkyunkwan’s Student AI Society

Field notes

The society organizes its program as a sequence from foundational machine learning to specialized NLP and computer vision tracks, then to frontier papers and student research. The research section names five projects, including anomaly detection with Mamba, neural-codec audio watermarking, VLM hallucination handling, open-set counting, and medical multimodal fusion. Several entries include reported metrics, such as Exact-Match moving from 28.2% to 45.5% and medical-fusion AUROC/AUPRC figures of 0.9276 and 0.9387.