Machine Learning in the arts

Computers Watching Movies by Benjamin Grosser. Takes computer vision algorithms and applies them to movies. At first it’s just kind of cute, but upon a closer look, there’s something really interesting there. For example, look at the Matrix and Inception links vs. Annie Hall and American Beauty. Matrix and Inception cover more of the screen; they look more “epic” maybe. For example, maybe before watching the movie, you watch the computer watching the movie. Or maybe there’s something you can tell about people’s movie tastes based on the CV outputs of what they watch. (I know Netflix would love to know about that.)

Genetic Algorithm Walkers by Rafael Matsunaga.

Genetic algorithms are a lot of fun, but it’s particularly fun when you’re genetic-algorithming something silly like humanoids walking. It shows the algorithm in action in a way that’s pretty easy to understand. Also, it maps the variables to pretty-easily-rememberable names. It’s a neat way to show evolution, to teach how genetic algorithms can come up with something that works pretty well for difficult tasks like walking with joints.