Music Information Retrieval (MIR)

Music Information Retrieval (MIR) is an interdisciplinary field concerned with the extraction of information from music and its analysis. It aims to study the processes, systems, and knowledge representations required to retrieve information from music, which can be in symbolic format (such as MIDI files), audio format (such as mp3 files), or vector format (such as scanned sheet music).It studies the use of techniques and knowledge from signal processing, machine learning, music cognition, database management, human-computer interaction, music archiving or the sociology of music.

MIR started to be recognized as a scientific field in the early 2000s, and today its many applications have transformed a wide range of technologies around the world: from recommendation systems that help researchers find music in streaming services, to automatic genre and mood recognition systems that generate customized music. MIR combines machine learning with human expert knowledge to use digital music data ("symbolic" data such as sheet music images, MIDI files, audio, and metadata about music items) for information retrieval, classification, and estimation or sequence labeling.