Understanding the Math behind XgBoost

I am trying to understand how XgBoost is different from standard gradient boosting of trees. A good start is found here. It does cover how gradient boosted trees work and how they build upon gradients rather than residues. But what it fails to explain is the exact method of approach for XgBoost. How the regularization and boosting works in particular. Any help is appreciated.

© Copyright 2013-2019 Analytics Vidhya