From soap bubbles to viscous paints and from beating human hearts to dexterous octopus tentacles, complex physical systems exist in every corner of the natural world. Understanding the mechanics and exploring the limits of these systems is key to making scientific discoveries in many areas. However, this process is challenging, due to both the inherent complexities of the systems and the infinite possible parameter combinations in trial-and-error experiments. 

By combining simulation-based and data-driven approaches, I create fully automated computational frameworks to explore the complex physical world. I take three particular steps in exploring these approaches. First, I develop simulation methods based on novel geometric representations and PDE solvers to model systems exhibiting intricate structures. Second, I build algorithms that map structures to functional properties and efficiently explore the limits of these properties. Third, I devise machine learning algorithms to unveil the intrinsic mechanisms and new patterns associated with these extremal properties. I will demonstrate these approaches by giving three examples, including modeling complex fluids, designing functional soft bodies, and discovering new microstructural materials. Finally, I will discuss the future directions of applying these automated computational frameworks on making scientific discoveries in different areas.

Bio:
Bo Zhu is a postdoctoral associate in the Computer Science and Artificial Intelligence Laboratory (CSAIL) at Massachusetts Institute of Technology (MIT). He received his Ph.D. in Computer Science from Stanford University in 2015. His research interests encompass computer graphics, computational physics, and computational fabrication. In particular, he focuses on building computational approaches to automate the process of exploring complex physical systems.