Seminar 12: Alternating direction method of multipliers for large scale optimization

This is the 12th seminar in the weekly seminars of machine learning reading group at KTH (MLRG-KTH). This presentation covers the well-know Alternating Direction Method of Multipliers (ADMM) and its applications to large-scale machine learning problems. The talk is based on the following well-written tutorial:

- N Parikh, S Boyd, "Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers", FnT

[ArmanAli-wwml]

How can we calculate y^k which is in equation of x^k+1? Why did you do it’s zero? Isn’t it the same as dual problem ir equal to max (g(y))???