Swatter

MLFly is a Machine Learning Fly. It uses a neural network to learn how to avoid the swatter. The close-up of MLFly's head show where its eyes see motion. The arrows indicate which direction it's thinking of moving. The arrows grow in size with its certiantiy.

Each time you swat the fly a new generation is trained in the correct reaction. After about 20 generations the fly will become difficult to hit. You still have the upper hand. The speed of your swipe controls the speed of the swatter. It is possible to swipe so fast that MLFly cannot escape!

Swatter was created by software developer Matthew Wilber. For more information, visit GreenZeta.com