Seminar - SCM

This talk presents two examples illustrating how the framework of physical learning can provide insight into adaptive processes in soft and living matter. First, I show how tunable physical systems can exploit constrained local learning dynamics to continually learn new tasks without catastrophic forgetting of old ones — a hallmark of biological learning. Second, I show that the interplay of short-term pulsatility and long-term structural adaptation in periodically driven elastic flow networks, inspired by animal vasculature, is key to stabilizing realistic looped architectures and understanding pathological remodelling when pulsatility is disrupted. Together, these examples highlight how the decentralized dynamics and interpretable structure–function relationships characteristic of the physical learning paradigm can uncover general principles underlying emergent memory, adaptation, and function in disordered soft and living matter, while also informing the design of novel bio-inspired, energy-efficient metamaterials.