Along these lines, I think these classes of algorithms might have much broader applications and could possibly solve problems that we don't even have good defin

Along these lines, I think these classes of algorithms might have much broader applications and could possibly solve problems that we don't even have good definitions for yet. I'm particularly curious to see how the field develops new techniques for systems that start off with poor quality results but evolve with new data to produce better results. btw, I vouched for and upvoted the parent. It makes an excellent point and, frankly, I'm baffled by the downvotes/dead status.