Dongyu Xue

Implanting a lightweight structural adapter into a pretrained protein language model, with iterative refinement at inference, raises sequence recovery for given folds to 55.65%/56.63% on CATH 4.2/4.3.

X-MOL uses large-scale pre-training on 1.1 billion molecules to improve molecular analysis tasks through fine-tuning, achieving state-of-the-art results in property prediction, reaction analysis, and molecule generation.

This paper introduces FL-QSAR, a federated learning framework for QSAR analysis, demonstrating its effectiveness in collaborative drug discovery while preserving privacy.