Timezone: »

Bundle Methods for Machine Learning
Alexander Smola · Vishwanathan S V N · Quoc V Le

Tue Dec 04 03:20 PM -- 03:30 PM (PST) @
We present a globally convergent method for regularized risk minimization problems. Our method applies to Support Vector estimation, regression, Gaussian Processes, and any other regularized risk minimization setting which leads to a convex optimization problem. SVMPerf can be shown to be a special case of our approach. In addition to the unified framework we present tight convergence bounds, which show that our algorithm converges in $O(1/\epsilon)$ steps to $\epsilon$ precision for general convex problems and in $O(\log \epsilon)$ steps for continuously differentiable problems. We demonstrate in experiments the performance of our approach.

Author Information

Alexander Smola (Amazon)

**AWS Machine Learning**

Vishwanathan S V N (National ICT Australia)
Quoc V Le (Stanford)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors