[Back]

P. Berger, G. Hannak, G. Matz:
"Coordinate descent accelerations for signal recovery on scale-free graphs based on total variation minimization";
Talk: 25th European Signal Processing Conference (EUSIPCO) 2017, Kos, Greece (invited); 08-28-2017 - 09-02-2017; in: "2017 25th European Signal Processing Conference (EUSIPCO)", (2017), ISBN: 978-0-9928626-8-8; 1689 - 1693.

We extend our previous work on learning smooth graph signals from a small number of noisy signal samples. Minimizing the signal's total variation amounts to a non-smooth convex optimization problem. We propose to solve this problem using a combination of Nesterov's smoothing technique and accelerated coordinate descent. The resulting algorithm converges substantially faster, specifically for graphs with vastly varying node degrees (e.g., scale-free graphs).

convex programming; graph theory; minimisation; signal processing; signal recovery; scale-free graphs; total variation minimization; smooth graph signals; noisy signal samples; nonsmooth convex optimization problem; coordinate descent accelerations;

https://publik.tuwien.ac.at/files/publik_276662.pdf