I have thought about this problem, and I think the better approach is something way more domain specific. Start with an application in mind, see which machine l

I have thought about this problem, and I think the better approach is something way more domain specific. Start with an application in mind, see which machine learning algorithms they are using and just try to get them working. Once you have something working, notice the strengths of the algorithms for that particular domain. There is usually some nice learning theory to explain that. Then move on to another domain and repeat. As you continue doing this patterns will emerge across the domains. This is usually the best time to start working through a machine learning textbook. When you take this approach you defer a chunk of the theory for practical examples, but can immediately start to apply the algorithms and have some results.