Jia-Xin Zhu

AI2 methods can now be transformed from bespoke protocols into reusable workflows, significantly enhancing the efficiency of molecular simulations in complex chemical systems.

Fixed-$δ$ auditing reveals that method-dependent thresholds can obscure critical utility-fairness trade-offs in multi-task learning.

Training MLIPs on mixed surface charge states for metal/electrolyte interfaces can lead to inconsistent predictions, suggesting careful charge-specific training is crucial for accurate simulations.