Strumming Robotic Tunes- AI and Music
While artificial intelligence remains a topic much discussed even outside the scientific community, the talk revolves around aspects like military applications, job takeovers, and life-altering technological changes. Amidst all this, AI has quietly found a place in a more everyday part of our lives- the music industry.
The idea of using algorithms to develop music is not new. In fact, the advent of algorithmic music can be traced back to the 1700s, when the game Musikalisches Würfelspiel generated short piano compositions with choices made by rolling dice. While widely popular, this method could not create large pieces of music autonomously.
Today, the internet houses a vast number of music albums composed either entirely or partly by an AI. From making bits of music using dice to training networks to compose music pieces on its own, artificial intelligence techniques have come a long way.
“Artificial intelligence would be the ultimate version of Google. The ultimate search engine that would understand everything on the web. It would understand exactly what you wanted, and it would give you the right thing. We’re nowhere near doing that now. However, we can get incrementally closer to that, and that is basically what we work on.”
– Larry Page
AI, at its core, is the practice of using algorithms to parse data, learn from it, and then make a predictive model of real-world problems, thus enabling machines to learn. Stanford researcher John McCarthy defines AI as “the science and engineering of making intelligent machines, especially intelligent computer programs”. AI is accomplished by roughly mimicking the techniques of the human brain.
Machine learning makes use of something called a neural network to do all the heavy lifting. A Neural network, in its essence, is a layer of neurons whose activation varies based on its input and weight parameters. Working on a system of probability, the network makes decisions and predictions based on the data that is fed to it as input. In 1989, the first attempts to generate music with a flavour of neural networks was developed by Peter M. Todd, Michael C. Mozer, and others. The idea took years to turn into reality, owing to the slow computers at the time. Advances in computing horsepower have led the charge in the breakthrough of artificial intelligence to tackle myriads of problems.
The concept of the Long Short-Term Memory (LSTM) networks, which are variations of Recurrent Neural Networks, is involved in the process of generating music, through a process called (generative) deep learning. The computer trains itself to generate new melodies by creating a set of rules for it through the analysis of its training data – thousands of samples of a specific genre of music inputted to it.
April 2017 saw the release of I AM AI, the first music album composed and produced entirely by artificial intelligence. Amper, the artificially intelligent composer and producer of this album, was designed to work in collaboration with a human artist who would provide inputs to be used as composing parameters. “One of our core beliefs as a company is that the future of music is going to be created in collaboration between humans and AI. We want that collaborative experience to propel the creative process forward,” says Drew Silverstein, one of the founders of Amper.
With advancements in technology, writing music on paper has long become obsolete. Most artists today write music using digital software that facilitates easy saving and editing. Electronic music gained popularity with the introduction of digital synthesisers that do away with the need for learning instruments. Taking this further, the idea of conducting concerts solely managed by machines is achievable. However, a lot of progress needs to happen before this dream can be realised.
Among the notable developments in the field are music pieces such as “Daddy’s Car” and “Mr Shadow” by Flow Machines, a research project at Sony CSL that is coordinated by François Pachet, Magenta by Google Brain, whose applications include a sound synthesiser N-Synth, and AI-Duet.
Concerns have been raised over the rise of artificial intelligence, notably by personalities like Stephen Hawking and Elon Musk, fearing a dystopia where humans are slaves to AI. The takeover of jobs by machines is a much nearer and more real concern. Even as these debates over the future continues, AI in the present provides mankind with the power to create and synchronise music, opening up a new musical dimension.