
This course is about convex optimization. The image on the left illustrates how we can build a ‘‘sparse graphical model’’ based on Senate voting data, revealing an inner structure of the two political parties. The graph is obtained using a convex approximation described here.
The course covers the following topics.
Convex optimization: convexity, duality.
Algorithms: emphasis on firstorder, largescale methods, distributed optimization.
Selected topics: robustness, algebraic geometry.

Link to UC Berkeley Schedule of Classes: here.
Notes:
To communicate, we use bspace and Piazza.
EE 227BT replaces the class previously offered as EE 227A. In the future EE 227BT will be renamed EE 227B, and will be crosslisted again. The ‘‘T’’ means temporary — UC Berkeley has complicated rules about course numbers…
This is not an entrylevel graduate class. If you never took an introductory graduate class in optimization, I strongly recommend taking EE 127, or its graduatelevel version EE 227AT (offered concurrently next Spring 2013). In particular, I will expect you to be proficient in linear algebra.
Lectures: Tu,Th 111230P, 247 CORY (NOTE: CLASS ROOM HAS CHANGED).
Discussion section: Fri 910am, 289 Cory.
