I'm an EE/applied physics student (probably interested in an academic path involving signal processing, programming, computation, mathy stuff) with a hole in my upcoming schedule. The choice is basically between these two courses - numeric linear algebra and a course called "neural networks and self-learning systems". The first is a continuation on an introductory linear analysis course we're doing right now, and the other contains machine learning, pattern recognition, clustering, etc. They both sound pretty fantastic. Guessing the first one is more generally useful, but machine learning is such an interesting subject. Does anyone have any input to share?