The Infinity Machine platform uses neural networks to map audio to an array of strobe lights. The client application consists of a Java Swing application used to set light states in a collection of audio files for training purposes. Neural network computation occurs in either the client application or via a RESTful API to a server application. Project utilized Java, Jersey, JAX-RS, Hibernate OGM, MongoDB, Java Encog, Java Swing, Apache Math Commons, and Microsoft Azure.
The project also explored the use of distributed genetic algorithms for neural network training.
The project was developed by a team of 12 undergraduates to fulfill the requirements of a Machine Learning class at the University of California, Davis.