Open jfan3 opened 5 years ago
It's a distinct possibility! In the brain, synaptic plasticity is believed to be gated by signals such as dopamine, and one could imagine ANNs in which "dopamine" projects to various parts of the ANN, increasing learning in these targeted areas. I'm stealing this idea from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5690421/.
I think it is quite established that learning encourages the growth of new denderitic spines. I was wondering the formation of new XdG during learning can be incorporated into the model as well?