Yinwei Wei

HABIT achieves superior image retrieval performance by simulating human habit formation, effectively tackling the Noise Triplet Correspondence problem that plagues traditional methods.

INTENT achieves robust image retrieval by effectively mitigating both cross-modal and modality-inherent noise, outperforming traditional methods in the presence of annotation errors.

LLMs can handle complex, multi-turn conversations far more effectively by ditching the flat context window for a dynamic discourse tree.