Welcome to the Advanced Computation and Artificial Intelligence (ACAI) Lab at Purdue University, led by Dr. Shengfeng Yang. Our vision is to build Reliable Physical AI that learns in simulation and performs in the real world.

We develop Physical AI models that bridge the gap between simulation and reality. By combining high-fidelity physics simulations, advanced machine learning, and real-world validation, we build predictive models that are both accurate and computationally efficient. Our approach reduces the reliance on resource-intensive experiments while ensuring real-world reliability. Our research spans across scales: from predicting atomic-level defects in materials to enabling complex robot manipulation in unstructured environments.

Our Three Research Areas:

Defect-Aware Predictive Digital Twins

Scientific Machine Learning

Physical Intelligence for Robotics