Structure

[k_golyaev]

Originally published at Konstantin's Private Blog. You can comment here or there.

Previously I mentioned what machine learners do better than most economists. Now I want to bring up a slightly more controversial distinction. Many machine learning applications are fundamentally less willing to make assumptions. This may sound a bit abstract and

Comments

[yyy_dnuoh_flow]

there are simply too many moving parts and too few degrees of freedom
Если я правильно понял идею то тут скорее надо говорить не о степенях свободы, а о слишком мало количестве управляющих переменных.