Ensemble learning is indeed a powerful tool--more than just the sum of its parts.
This is a cool ranking that lists various learning algos and how they do on MNIST (which may or my not be a good benchmark for what you're doing). Links to the corresponding papers are provided, including a few ensemble methods (committees), which tend to perform rather well. http://yann.lecun.com/exdb/mnist/
I found this very cool! I clicked on the article thinking it would be about crowdsourcing, but it's actually about a cool machine learning technique I didn't know about. Plus, it was very well explained. Thanks for writing it!
This is a cool ranking that lists various learning algos and how they do on MNIST (which may or my not be a good benchmark for what you're doing). Links to the corresponding papers are provided, including a few ensemble methods (committees), which tend to perform rather well. http://yann.lecun.com/exdb/mnist/