“This work attempts to create lyrics from academic papers and appropriate melodies to go with them. We believe this system can also be modified to use different initial data sources, be it text sources for the lyrics or music sources for the music style. We chose academic papers as input due to their diversity and availability. Furthermore, due to their usual seriousness, it was our opinion that it would be amusing, not only for readers but also for authors, to see these works in a different light.”
The SMUG: Scientific Music Generator has been developed by Marco Scirea, Gabriella A. B. Barros, and Noor Shaker of the Center for Computer Games Research, IT University of Copenhagen, Denmark, and Julian Togelius at the Department of Computer Science and Engineering, New York University, US.
TRY IT: You can try it yourself, online, and hear the delights of such music.
“Bul let tins first ar gu ments na tu re an is an same thus hist re ca pi tu la tion con tro ver sial ahh p man in ha bi tants”
Their paper is presented today at ICCC 2015: The Sixth International Conference on Computational Creativity, Park City, Utah, US, June 29 – July 2, 2015. The conference programme, including papers on AI cocktail generation, AI computer-aided humour and AI automatic painting, can be found here.
UPDATE: The authors have kindly alerted us to the existence of an online version of SMUG. Improbable has uploaded the paper ‘Frictional Coefficient under Banana Skin’ (by the winners of the 2014 Ig Nobel physics prize, Kibyyoshi Mabuchi, Kensei Tanaka, Daichi Uchijima and Rina Sakai) The paper is now available as a lyrical and musical rendition.
Upload your paper of choice (in .pdf format) here.