AI-based instability-driven curvature fields enable multiscale structure inverse design

Fluid–solid instabilities are typically suppressed in biofabrication, limiting their use as structural design mechanisms. Here, we show that viscoelastic coiling instability can be harnessed as a deterministic process to encode hierarchical structures. Using a muscle-derived viscoelastic self-assembling bioink (MVSB), we demonstrate that programmable curvature fields generated by coiling dynamics guide nanofiber organization and transmit structural information across scales. By integrating physics-based modeling with a physics-guided residual neural network (PG-ResNet), we enable inverse design of instability-generated geometries from target curvature fields. This strategy produces auxetic biological scaffolds with coupled mechanical properties and cellular organization. Our findings establish fluid instability as a programmable physical framework linking flow dynamics, multiscale structure, and biological function.

Dr. Yuanhao Wu is currently a Full Professor and Plastic Surgeon at Wuhan Union Hospital. She has been selected as a recipient of the Outstanding Youth Science Foundation (Overseas) and appointed as a Distinguished Professor of Huazhong Outstanding Scholars. Her research focuses on the frontier areas of supramolecular self-assembly and biomimetic intelligent systems, with a particular emphasis on applications in regenerative medicine. She has successfully established an interdisciplinary research team that bridges clinical medicine and biomedical engineering. In recent years, as first or corresponding author, she has published high-impact research articles in leading international journals such as Matter, Nature Communications, and Advanced Functional Materials. Over the past three years, she has been approved to lead six research projects, with total funding amounting to 3.28 million RMB. Additionally, as the first inventor, she has filed ten patents, with technology transfer revenues exceeding 1 million RMB.