Presentation on samedi at 2:15 après-midi to 2:45 après-midi in Room 1170.
This talk shows how Python was used to trigger an epic sound and light show whenever the Montreal Canadiens hockey team scored a goal in last season's playoffs.
The author trained a machine learning model to detect in real-time that a goal was just scored by the Habs based on the live audio feed of a game and to trigger a light show using Philips hues in his living room. The system was built using various Python modules, more specifically scikit-learn, pyaudio, librosa, phue and bottle.
The talk will cover every step of the hack from the creation of the dataset, the feature generation and the machine learning model training, the real-time audio feed processing to do real-time scoring and the programming of the light show.
François Maillet is a computer scientist specializing in machine learning. He currently leads the machine learning team at Datacratic, a Montreal startup building the Machine Learning Database (MLDB).
He previously built an audio fingerprinter used by a TV advertisement monitoring company. He also did research at the Université de Montréal applying machine learning to solve problems related to music recommendation, play list generation and track auto-tagging.