Abstract
We develop a learning algorithm for complex spring networks, aimed at adjusting their physical parameters so as to ensure a desired mechanical behaviour in response to physical input (control) stimuli. The algorithm is based on the gradient descent paradigm and has been tested on our computer implementation. The systems output by our software conform to real-world physics and thus are also suitable for hardware implementation.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have