Yang Zhong

First-principles quantum transport simulations no longer stall at a few hundred atoms: predicting Hamiltonians with an $E(3)$-equivariant GNN unlocks linear scaling to 10,000+ atom devices with a $1,000\times$ speedup over conventional DFT-NEGF.

Existing vulnerability detection tools struggle with a mere 33.3%-40.1% F1 score, underscoring the urgent need for better benchmarks like VICBench.

Endowing VLMs with intrinsic 3D geometric awareness and physical interaction cues via XEmbodied substantially boosts performance on spatial reasoning and embodied tasks, surpassing existing 2D image-text pretrained models.