Breaking the curse of kernelization

Online algorithms that process one example at a time are advantageous when dealing with very large data or with data streams. Stochastic Gradient Descent (SGD) is such an algorithm and it is an att...