Sound Recognition Machine Learning in Max/MSP

Hi there,

Does anybody know of any existing libraries, or have any suggestions as to how I could implement machine learning for sound recognition into a Max patch? I'm not looking for speech recognition per se, but rather for the ability to train my patch on a kind of sound (imagine a training set of 10 or so recordings) that I could then use as a sound recognition platform. So if I trained my patch on something like 10 dog bark recordings, when I then present it with another dog bark recording it should be able to recognize that this is a "similar sound" (1), and when I present it with a cat meow recording it should be able to recognize that this is a "dissimilar sound" (0).

Any help or suggestions would be much appreciated!