Program Overview

Program Overview

The program in question is "深度學習於音樂分析及生成" (Deep Learning for Music Analysis and Generation), offered by the Department of Electrical Engineering, National Taiwan University.

Program Details

• Course Number: CommE5070
• Course Credits: 3 credits
• Course Type: Elective
• Department: Electrical Engineering and Telecommunications Engineering
• Instructor: 楊奕軒 (Yang Yi-Xuan)
• Class Schedule: Thursday, 6-8 pm
• Classroom: 學新118 (Xue Xin 118)
• Language: Chinese
• Total Enrollment: 60 students

Course Description

"Music Information Research" (MIR) is an interdisciplinary research field that concerns the analysis, retrieval, processing, and generation of musical content or information. This course focuses on the application of machine learning, particularly deep learning, to address music-related problems. The course is divided into two parts: analysis and generation.

Course Objectives

1. Understanding of different aspects of music: timbre, rhythm, pitch, harmony, and structure, and the use of domain knowledge for corresponding music signal analysis tasks.
2. Understanding of and hands-on experiences with deep learning techniques for music audio signal analysis.
3. Understanding of and hands-on experiences with deep generative models for both musical audio and text-like music data such as MIDI.
4. A taste of the fun of research.

Course Requirements

• Students are expected to have a good background in machine learning and mathematics.
• Students are expected to have good coding experience in Python and a deep learning framework such as PyTorch.
• Students are expected to have a great interest in music.

Evaluation Method

• Homeworks (60%): 3-4 coding assignments related to building ML/DL models.
• Final Project (40%): For teams of 2 or 3, oral presentation + written report.

References

• Jakub M. Tomcza, Deep Generative Modeling. 978--2. Springer, 2022.