Non-linear motor control by local learning in spiking neural networks
This addresses the challenge of motor control in robotics or neuroscience using spiking neural networks, but it appears incremental as it builds on existing local learning methods for specific dynamics.
The paper tackled the problem of learning weights in spiking neural networks with hidden neurons using local, stable, and online rules to control non-linear body dynamics, by employing a supervised scheme called FOLLOW to train a network to control a two-link arm and reproduce a desired state trajectory, achieving the lowest test error with a differential feedforward architecture.
Learning weights in a spiking neural network with hidden neurons, using local, stable and online rules, to control non-linear body dynamics is an open problem. Here, we employ a supervised scheme, Feedback-based Online Local Learning Of Weights (FOLLOW), to train a network of heterogeneous spiking neurons with hidden layers, to control a two-link arm so as to reproduce a desired state trajectory. The network first learns an inverse model of the non-linear dynamics, i.e. from state trajectory as input to the network, it learns to infer the continuous-time command that produced the trajectory. Connection weights are adjusted via a local plasticity rule that involves pre-synaptic firing and post-synaptic feedback of the error in the inferred command. We choose a network architecture, termed differential feedforward, that gives the lowest test error from different feedforward and recurrent architectures. The learned inverse model is then used to generate a continuous-time motor command to control the arm, given a desired trajectory.