I'm a neuroscientist in a doctoral program but I have a growing interest in deep learning methods (e.g., http://deeplearning.net/ ). As a neuroscientist using MR imaging methods, I often rely on tools to help me classify and define brain structures and functional activations. Some of the most advanced tools for image segmentation are being innovated using magical-sounding terms like Adaboosted weak-learners, auto-encoders, Support Vector Machines, and the like.
While I do not have the time to become a computer-science expert in artificial intelligence methods, I would like to establish a basic skill level in the application of some of these methods. Soylenters, "Do I need to know the mathematical foundation of these methods intimately to be able to employ them effectively or intelligently?" and "What would be a good way of becoming more familiar with these methods, given my circumstances?"
(Score: 3, Informative) by Non Sequor on Monday June 29 2015, @06:06PM
I haven't studied AI methods but I think you want to study linear algebra and computational linguistics.
Most of the methods that I've read about that you can run off and implement at home (e.g. image recognition/classification) involve linear algebra, particularly a good understanding of identifying useful subspaces in high dimensional data and understanding the significance of eigenvectors in iterative processes.
Computational linguistics tends to be less practical unless you limit yourself to things that don't really qualify as AI. The most easily applied techniques (regular and context free languages) are too simple for anything other than rigidly applying a set of rules. You need to understand why it's hard to get more knowledge out of these systems than you put into them manually. Key limitations of neural nets are governed by these results.
Write your congressman. Tell him he sucks.
(Score: 3, Interesting) by Non Sequor on Monday June 29 2015, @06:39PM
Also look into Kolmogorov complexity and the notion of a quine. Keep in mind that large portions of infrastructure for human intelligence are dedicated to homeostasis and self-maintenance rather than learning and these problems may be prerequisites for what you really want and they may be much larger problems than you are prepared to solve.
Write your congressman. Tell him he sucks.