Spider webs are incredible biological structures, comprising thin but strong silk filament and arranged into complex hierarchical architectures with striking mechanical properties (e.g., lightweight but high strength, achieving diverse mechanical responses, multiple biological functions).
Published in
@PNASNews today, we are excited to share three graph-focused deep learning techniques (sparse and dense diffusion model, autoregressive transformer) to capture and model the complex design language of large 3D graph architectures—here exemplified for 3D spider webs. We use the methods to generate a diverse array of de novo bioinspired structural designs.
Paper: W. Lu, N.A. Lee, M. Buehler,
pnas.org/doi/10.1073/pnas.23…