Using GPUs in Machine Learning

Machine learning algorithms have been known to perform better with more free parameters to tune and more training data. But learning algorithms are often too slow for large scale applications and thus the size of training models (free parameters) and data is limited in practice. So usage of GPUs to improve the speeds of these algorithms has attracted a lot of attention recently. I focused on a couple of papers analyze the performance improvements by adapting the learning algorithms for GPUs. I ran the experiments using the source code provided publicly under Theano framework