Building Machines With an Ear for Music
What’s the Latest Development?
Imagine you wanted to find every jazz song that combines the harmonica with the saxophone. One day soon, you may be able to. By improving on current algorithmic recommendation technology, the kind that suggests which books you should buy on Amazon and which movies you should watch on Netflix, engineers are finding ways to classify the vast libraries of music we have on the Internet. Currently, computer-based recommendation systems cannot recommend anything that is not already popular, and human-based classification systems are ill-equipped to handle the 60 hours of multimedia uploaded to YouTube every minute.
What’s the Big Idea?
Computer engineers in California and New York want to create a hybrid of the man and machine techniques for classifying and recommending music. By refining machine-learning software, engineers can teach a computer how to mimic ways that humans classify music, then make recommendations based on those classifications. This kind of naturally-enhanced AI is becoming increasingly popular: “Such man-machine collaborations seem to be taking over many facets of artificial intelligence. … In medicine, law, finance, retailing, manufacturing and even scientific discovery, the key to winning the race is not to race against machines, but to win using machines.”
Photo credit: Shutterstock.com
