- Aug 2020
-
codeforces.com codeforces.com
-
Linova wants to choose kkk cities which can maximize the sum of happinesses of all envoys
k : min path<br> n - k : max path
-
city 111 is the capital of the kingdom.
1 at head
-
There are nnn cities and n−1n−1n-1 two-way roads connecting pairs of cities in the kingdom. From any city, you can reach any other city by walking through some roads.
means tree
-
- Mar 2020
-
cs231n.github.io cs231n.github.io
-
Since Neural Networks are non-convex
neither convex neither non-convex
"The fact that J has multiple minima can also be interpreted in a nice way. In each layer, you use multiple nodes which are assigned different parameters to make the cost function small. Except for the values of the parameters, these nodes are the same. So you could exchange the parameters of the first node in one layer with those of the second node in the same layer, and accounting for this change in the subsequent layers. You'd end up with a different set of parameters, but the value of the cost function can't be distinguished by (basically you just moved a node, to another place, but kept all the inputs/outputs the same)."
Tags
Annotators
URL
-
- Jan 2020
-
cs231n.github.io cs231n.github.io
-
∇wyiLi=−(∑j≠yi1(wTjxi−wTyixi+Δ>0))xi
grad wrt wyi will be -xi only when then stuff inside max turns out to be true else 0
-
- Dec 2019
-
cs231n.github.io cs231n.github.io
-
L=1N∑iLidata loss+λR(W)regularization loss
data loss is over all examples, similarly, regularisation loss is also over all examples for all classes
-
-
web.archive.org web.archive.org
- Nov 2019
-
saturncloud.io saturncloud.io
-
Temporal dependencies
wut?
-
