Goto

Collaborating Authors

 synapse






Acontrastiveruleformeta-learning

Neural Information Processing Systems

Our rule may be understood as ageneralization of contrastive Hebbian learning to meta-learning and notably, it neither requires computing second derivativesnorgoing backwardsintime,twocharacteristic features of previous gradient-based methods that are hard to conceive in physicalneuralcircuits.


a1ada9947e0d683b4625f94c74104d73-Paper.pdf

Neural Information Processing Systems

Many networks in the brain are sparsely connected, and the brain eliminates synapses during development and learning. How could the brain decide which synapses to prune?