Accelerated Coordinate Descent with Adaptive Coordinate Frequencies

Coordinate descent (CD) algorithms have become the method of choice for solving a number of machine learning tasks. They are particularly popular for training linear models, includ- ing linear support vector machine classication, LASSO regression, and logistic regression. We propose an extension of the CD algorithm, called the adaptive coordinate frequencies (ACF) method. This modied CD scheme does not treat all coordinates equally, in that it does not pick all coordinates equally often for optimization. Instead the relative frequen- cies of coordinates are subject to online adaptation. The resulting optimization scheme can result in signicant speed-ups. We demonstrate the usefulness of our approach on a number of large scale machine learning problems.