Lily Turner on Teaching a Machine to Play Like Mark O'Connor


Our guest for this episode of the Music at Pitt Podcast is Lily Turner, a freshly minted Pitt graduate from Blacksburg, Virginia. As a student at Pitt she was a double major in Computer Science and Music, specifically pursuing the Global and Popular Music track. Lily is particularly interested in machine learning and artificial intelligence. She’s previously done research into using convolutional neural networks for computational drug discovery and is now interested in applying machine learning to music. She’s passionate about music-making of all kinds, from old-time fiddling to playing baritone in the Pitt Band. She was very involved with a few student organizations, and was president of Urban Gaming Club. She also took part in the Christian Student Fellowship as well as Project Potter, a Harry Potter-themed service organization at Pitt that raises money for and volunteers with children in the Pittsburgh area. Lily also enjoys drinking tea and playing Dungeons and Dragons with friends.

For her senior capstone project, Lily applied machine learning to replicate the playing style of renowned fiddler Mark O’Connor and summarized her findings in a paper titled, “Mark O’Connor Bot: Recurrent Neural Net Generation of Texas-Style Fiddling.”