Dr Eng. Mateusz Modrzejewski from the Department of Computer Graphics at the Faculty of Electronics and Information Technology of the Warsaw University of Technology (PW) conducts international research on the use of artificial intelligence and deep neural networks for analysis and music creation.

Dr Modrzejewski is dealing with music information retrieval. He wonders if artificial intelligence algorithms can replace the conductor, what is the chance that the instruments will learn to play on their own, and what would be the difference between machine-written and human-written music.

The research by Dr Mateusz Modrzejewski is based on the use of deep neural models, including recursive and generative GAN networks and modern representations based on a transfer learning approach.

New methods are developed with the use of large data sets in audio and symbolic formats such as MIDI (Musical Instruments Digital Interface). After processing individual songs with the use of artificial intelligence algorithms, it is possible to generate completely new music that meets certain assumptions or to accurately classify existing music according to given criteria, such as identification of the singer, mood and instruments, or recognition of the musical genre.

The scientist from the Warsaw University of Technology is also a drummer giving concerts in China, Vietnam, Germany, Scotland, England, Estonia, Ukraine, and at many festivals and music events in Poland.

“In my research, I combine my two greatest passions; music and issues related to deep neural networks that are the basis of machine learning and artificial intelligence applications. As a jazz drummer with extensive stage experience, I know and understand the structure of music, and as an engineer, I write it not only in the form of notes but also in the computer language. And this is an excellent set of input data for machine processing, analysis and composition of works by artificial intelligence algorithms”, explains Dr Mateusz Modrzejewski.

Adrian Andrzejewski