Spiking neural networks take control

Back in September I had an review article in Science Robotics published, discussing new work from Abadia et al, 2021 titled A cerebellar-based solution to the nondeterministic time delay problem in robotic control. In my review I talk about the parallel’s between the brain inspired approach Abadia et al used to create a neural circuit that can effectively control a compliant, highly nonlinear robotic system in the face of variable feedback delay, and the history and application of convolutional neural networks. The basic premise is, to no one’s surprise I’m sure, that there is still much to learn from the brain in building effective artificial neural networks. If you’d like to read the article, you can now check it out here: Spiking neural networks take control.

Advertisement