Sound Applications in Machine Learning

The future of programming will rely on machine learning. Machine learning is what gives a program the ability to learn and improve based on its own experience. Although this has been better known for improving medical and statistical research, it’s potentiality in the arts is often overlooked. It is overlooked because it is hard for artists to understand how to use machine learning to their advantage. Since it is a new topic, there are not a lot of quick-and-easy tutorials for artistic programmers who are finding themselves lost. Fortunately, there is a clear and concise way to explain these complex topics. There are machine learning libraries that are easy to learn from that can be installed onto one’s computer. These libraries include Scikit-Learn and Tensorflow, which do most of the dirty work for those who are interested in getting their hands on this technology without having to build it from scratch. Although some may prefer jumping immediately into the application in music and sound, it is important to recognize that most of the work involves understanding how machine learning works. The programming language Python will be used because it is the most popular language for this topic due to it’s easy and concise nature. Machine learning can develop original music. For instance, it is possible to give a program a folder full of music by The Beatles and have it output what it would predict they would perform next. This can also be done with a folder full of random songs that are not all the same artist. Machine learning can develop original sounds as well. A programmer can give a program a folder full of voice recordings and have the program output an original sound inspired by their voice. The tutorials, explanations and real-world applications will serve as motivation for those who are interested in expanding their horizons in music and sound.