Selected Topics in Analysis of Sound Signals

Students will have the opportunity to explore the use of different methods for pattern recognition and machining learning (for example, deep neural networks) to solve the problems that we encounter when analyzing sound signals, such as identification of events in sound recordings, classification of sound recordings, transcription of music, detection of samples in music, etc. In the course of the semester, students will develop their own algorithm for solving a problem and send it to one of the evaluation campaigns (e.g., Mirex or DCASE), where its performance can be compared with approaches developed by other researchers (mostly PhD students) around the world.