What does music written by Artificial Intelligence (AI) sound like? How does AI with no knowledge of what it’s like to be a human at a traditional music session in the west of Ireland come up with a folk tune it names the “The Drunken Pint”? How many folk songs do you need to train an AI to write folk tunes? (Answer: approximately 20,000).
The sixth event of the series, Bob Sturm and Oded Ben-Tal will explore the basic concepts of machine learning and its application to music data. Their "folk-rnn" system applies deep recurrent learning to create a computer model of transcriptions of traditional tunes typically found in Ireland and the UK. This model can be used to generate a limitless number of new tunes. However, their interest is in exploring the potential of such models as creative partners. To that end, they have worked with musicians – including those practicing this type of traditional music – to investigate creative applications of their system.
Bob Sturm is currently a Lecturer in Digital Media at the Centre for Digital Music, School of Electronic Engineering and Computer Science, Queen Mary University of London. His research interests include digital signal processing for sound and music signals, machine listening, evaluation, and algorithmic composition.
Oded Ben-Tal is a composer and researcher working at the intersection of music, computing, and cognition. His composition include both acoustic pieces, works combining instruments with electronics and multimedia work. He is a senior lecturer at the Music Department, Kingston University where he teaches composition, electronic music, and Music Psychology.
About Sound Salon
Artists have been working with the internet for decades but until recently, archives and critical writing have focused on visual cultures. This series aims to fuel critical discussion, new vocabulary and analysis in sound and music, supporting communities and practitioners in these fields to understand how the internet - and new technology more broadly - is shaping the field.
Together with artist Holly Herndon, Jennifer launched the project Post-Internet Sound in 2015. A crowdsourced database of sound and music works, the site is open access and welcomes both academic and non-academic contributions, be it in the form of writing or a source of sound. Anyone can contribute via Google docs.
About the curators
Journalist Adam Harper discusses aesthetics and criticism in music, art and life on his blog Rouge's Foam. He studies musicology, composes music and has written for The Wire magazine, is the author of Infinite Music, Imagining the Next Millennium of Human Music-Making.
Dr. Michael Waugh is a lecturer in Media Cultures at the University of Hertfordshire. His research is in the fields of digital media and popular music and his thesis, titled "'Music that actually matters?', is the first monograph-length study of Post-Internet musicians and their self-conscious explorations of contemporary identity, representation, social media and digital networks. Fifteen prominent artists, including Holly Herndon, Arca, 18+, Jam City, Evian Christ, Fatima Al Qadiri, Aaron David Ross, Ryan Trecartin and SOPHIE, were interviewed as original research for the project.
In collaboration with Brunel University London.
