Fly colony relies on artificial intelligence software run on a Raspberry Pi

David Bowen, who is an Associate Professor of Sculpture and Physical Computing at the University of Minnesota, Duluth, has created a project called FlyAI. In this artificial intelligence experiment, Bowen has established a fly colony inside a transparent globe and is using a piece of learning software called TensorFlow. The Raspberry Pi captures images of the fly environment and then processes it through image recognition libraries and TensorFlow to determine if a fly was present in the photograph. If the fly is determined as being present, a pump delivers water and nutrients to the colony. If the fly is not definitely there, no nutrients are delivered. The system will run indefinitely, just to see what happens. You can read more about the system on Bowen’s blog or watch a video below.

Leave a Reply

Your email address will not be published. Required fields are marked *