An in-silico Neural Model of Dynamic Routing through Neuronal Coherence
Devarajan Sridharan · Brian Percival · john arthur · Kwabena A Boahen

Wed Dec 5th 05:20 -- 05:30 PM @ None
We describe a neurobiologically plausible mechanism to implement dynamic routing using the concept of neuronal communication through neuronal coherence. The model, implemented on a neuromorphic chip, incorporates a three-tier neural network architecture: the lowest tier comprises of raw input representations, the middle tier of the routing neurons, and the topmost tier of invariant output representation. The correct mapping between input and output representations is realized by an appropriate alignment of the phases of their background oscillations by the routing neurons. We demonstrate that our method is able to dramatically reduce the number of connections required from O($N^{3}$) to O($N^{2}$)

Author Information

Devarajan Sridharan (Neurosciences Program, Stanford University)
Brian Percival (Stanford University)
john arthur (Stanford University)
Kwabena A Boahen (Stanford University)

More from the Same Authors