What are adaptive learning algorithms exactly? I googled it but didn't really find anything. I read this term on an electronics forum and I really want to know what it means since I'm a noob when it comes to computer science :B

I did an online course with Stanford a few years back on AI. Taught by the guy that leads the Google self-driving car team. By the end of the course, I'd decided it should have been called "Advanced Probabilistic Algorithms." Most of what we studied had to do with taking mountains of data and recognizing algorithmically the important correlations in that mountain. So there was a learning phase, but what came out of that was a tuned algorithm that could process new data of the same type quickly, because now it "knew what to look for." It was mostly applications of Bayesian techniques one way or another.

 

I'd think you could apply the term "adaptive" to any algorithm that "learns" what aspects of data are important and what aspects aren't.

AFAIK It is an algorithm that re-factors itself as new inputs and test cases are generated, and thereby makes itself more relevant.