François Fleuret, Olivier Canévet
We propose a novel approach to efficiently select informative samples for large-scale learning. Instead of directly feeding a learning algorithm with a very large amount of samples, as it is usually done to reach state-of-the-art performance, we have devel ...
2014
