Hauptinhalt

Topinformationen

Deep Learning for Music Processing

8.30503

Dozenten

Ulf Krumnack, Ph.D.

Beschreibung

Course Description:

In this seminar we will investigate the potential of deep learning based music processing on a theoretical and practical level. We will cover several topics from analysis and synthetis, including music fingerprinting, retrieval, classification, tracking, synchronization, separation, transcription, generation, and composition. The goal of the course is to achieve a solid understanding of common representation and processing schemes, problem settings, learning scenarios, and training data. On the practical level, demo implementations allow to explore the strengths and weaknesses of current systems.

Learning Objectives:

By the end of this course, students will:
1. have got an overview of different topics in music processing
2. have seen different approaches to process music with deep networks
3. have gathered experience with practically employing such systems

Prerequisites:

The course is intended for advanced bachelor and master students. We assume a basic knowledge of deep learning, as for example provided in the course "ANNs with TensorFlow", including familiarity with at least one of the common deep learning frameworks, like TensorFlow or PyTorch. Some background or interest in music could also be beneficial.

Course Format:
- Delivery Method: in-person, no recordings or online sessions
- Type of Contact & Contact Hours: weekly seminar sessions Tuesdays 10:00-12:00
- Selection Process: none planned, assuming a reasonable number of participants show up

Assessment and Grading:

To pass this course, you should actively participate in discussions, both in class as well as in the online forum. Every participant should present one topic and prepare a practical demonstration. Topics will be announced in the first session.

Will this class be offered again/regularly?
Unsure

Weitere Angaben

Ort: 69/117
Zeiten: Di. 10:00 - 12:00 (wöchentlich)
Erster Termin: Dienstag, 15.04.2025 10:00 - 12:00, Ort: 69/117
Veranstaltungsart: Seminar (Offizielle Lehrveranstaltungen)