Advancements in quantum computing may present sustainable creative opportunities for artists experimenting with machine learning.\u00a0IL\u0100\u2019s latest album \u2018Quantum Computer Music\u2019 aims to show how.<\/p>\n\n\n\n
The project saw the London-based artist, producer and vocalist approach machine-learning systems as collaborators rather than automated generators.<\/p>\n\n\n\n
Using a combination of quantum reservoir computing (QRC), DNA sonification and AI synthesis (more on that in a second), with her latest piece <\/em>IL\u0100 has built on her earlier project \u2018Recurse\u2019 \u2014 a collaboration with quantum-technology company MOTH and Harvard University that was described as the \u201cworld\u2019s first infinite mix\u201d<\/a>.\u00a0<\/p>\n\n\n\n\n What is QRC, and how does it apply to music, you may ask? In brief: it\u2019s a machine learning method in which data can be poured into a system \u2014 called the reservoir \u2014 causing the system to react in a number of complex ways.\u00a0<\/p>\n\n\n\n IL\u0100\u2019s process involves training the system on her original vocal recordings, which the system produces its own \u201ctake\u201d on. IL\u0100 then sculpts a new version, sometimes then feeding it back in for further iteration.<\/p>\n\n\n\n \u201cIt\u2019s an evolving, self-referential dialogue between me and the machine that blurs the line between author and output,\u201d IL\u0100 told Music Ally. \u201cThe outputs are not random generative fragments. In a way they\u2019re reflections of myself through a quantum lens.\u201d\u00a0<\/p>\n\n\n\n\n Only available to highly specialised academic researchers until very recently, QRC is being newly utilised among creatives experimenting with small, high quality data sets.<\/p>\n\n\n\n \u201cA big issue with many Large Language Models is that they\u2019re trained on vast amounts of data scraped from the internet \u2014 often without artists\u2019 permission,\u201d IL\u0100 told BBC. \u201c[QRC] requires far less data to create meaningful results, making it more sustainable and collaborative by nature.\u201d<\/p>\n