Sequential requirements

Matthias Melcher, Jul 08, 2016
Commentary by Stephen Downes

Discussion of some of my recent comments on whether a neural network needs domain knowledge in order to learn. As Matthias Melcher suggests, many artificial neural networks will have the benefit of training by human experts. But what of human neural networks? And how does this play into the idea that learning is sequential? "I think," writes melcher, "while artificial networks need some prerequisite input, human neuronal networks use recognizing from the very beginning and require no indispensable prerequisites." This is enabled by thinking of learning as a process of recognition rather than of representation. "Recognition explains the deeper mechanism of learning as not linear/ sequential (not via fixed isolated representations) but as laminar/ all-at-once (multiple connected features of a pattern)."

Views: 0 today, 339 total (since January 1, 2017).[Direct Link]