Bin Wen

Context-augmented RL lets smaller MLLMs punch *way* above their weight, rivaling much larger models on reasoning tasks while dodging reward hacking.

Achieve zero-collision embedding tables in production recommenders without sacrificing training speed, unlocking better personalization via fresher and higher-quality item embeddings.

Overcome Alzheimer's speech detection's data scarcity with FAL-AD, a federated learning framework that hits 91.52% accuracy by generating synthetic speech samples and aligning acoustic and textual features.

Achieve scalable and consistent multi-reference image editing by dynamically serializing reference images into a coherent latent sequence, outperforming existing diffusion-based methods.