Nice historical discussion around the #doubledescent curve. Thanks Dmitry Kobak @hippopedoid. Next week, we'll publish a new blog post explaining how to reproduce the behavior of this curve in Python with LightOn's #OPU and #RandomFeatures.
Dmitry Kobak @hippopedoidSpent some time investigating history of "double descent". As a function of model complexity, I haven't seen it described before 2017. As a function of sample size, it can be traced to 1995; earlier research seems less relevant. Also: I think we need a better term. Thread. (1/n)
015
