When Neural Networks Learn Like Brains: Locality Beats Weight Sharing in Hebbian Vision Models
A new computational neuroscience study shows that biologically constrained Hebbian learning can build useful visual representations without exact weight sharing, ...
