clix11

4 ms·

1. 1.▲
   by clix11 5y ago

   > what advantage do I get from the compute cost of gradient descent over random sampling? Random sampling becomes prohibitive in higher dimensions due to the curse of dimensionality [0]. Gradient descent doesn't have this problem an
2. 2.▲
   by clix11 7y ago

   This is not entirely true as of C# 7.2. That version of the language adds "ref structs" - which can only be allocated on the stack and cannot be boxed [1]. This requires some care though, as the struct may still be copied needless
3. 3.▲
   by clix11 10y ago

   Infiltrating the ranks of your enemy? Basically, if he got most of the smart people in their 20s as users of the site, it would take that longer for someone to uncover his operations. That's what Al Capone did, though I have no idea ab