Mechanical Engineering: Stewart Harris Seminar Series
The Shape of Intelligence: Designing Materials that Cloak, Morph, and Move
Liwei Wang, Ph.D
Assistant Professor
Department of Mechanical Engineering,
Carnegie Mellon University
Abstract:
Engineered systems have traditionally relied on electronics to perform complex sensing, actuation, and control tasks. Manufacturing innovations have made it possible to embed these intelligent functions directly into materials and structures, giving rise to chip-free physical intelligence. In this emerging paradigm, engineered materials can autonomously adapt and move in response to external stimuli in a programmable manner, offering transformative potential for biomedical, environmental, and robotics applications. This new manufacturing flexibility, however, also introduces significant challenges for inverse design. Designers must now reason over complex physical behaviors and navigate vast, heterogeneous design spaces that encompass materials, structures, stimuli, and manufacturing processes.
In this talk, we explore how computational intelligence, including machine learning and computational optimization, can be leveraged to establish inverse design frameworks for physical intelligence. The first part of the talk focuses on the co-design of materials and structures in 3D-printed heterogeneous metamaterials. We present a framework that integrates deep generative modeling with multiscale optimization to program macroscopic property distributions and deformations, enabling unfeelability cloaks that conceal objects from mechanical detection. The second part extends the framework to responsive material systems enabled by 4D printing, where materials, structures, stimuli, and manufacturing processes are co-designed to perform on-demand tasks. By combining differentiable simulation, machine learning, and physical principles, we enable materials to exhibit a range of controllable intelligent behaviors, from shape morphing to dynamic motion under thermal, optical, and magnetic stimuli.
Bio:
Liwei Wang is an Assistant Professor of Mechanical Engineering at Carnegie Mellon University. He received his B.S. and Ph.D. degrees in Mechanical Engineering from Shanghai Jiao Tong University. Prior to his current position, he was a postdoctoral scholar and a visiting predoctoral scholar at Northwestern University. His lab at CMU focuses on developing machine learning and optimization methods to establish computational design frameworks for functional materials, metamaterials, and programmable material systems. He serves as an editorial board member of the Structural and Multidisciplinary Optimization (SMO), and as guest editor for ASME Journal of Mechanical Design and Thin-Walled Structures. He is the recipient of the inaugural ASME Design Automation Dissertation Award and the ASME Design Automation Conference Best Paper Award.