Comment on Every Model Learned by Gradient Descent Is Approximately a Kernel MachineparentComments−mrtranscendence5yThis really threw me off when I first started learning about ANNs, coming from a traditional econometrics background. B-but ... the parameters aren’t identified!
Comments
This really threw me off when I first started learning about ANNs, coming from a traditional econometrics background. B-but ... the parameters aren’t identified!