Program Overview

Program Overview

The University of Copenhagen offers a course in Deep Learning (DL) as part of its academic programs. This course is designed to provide students with insight into the foundational methods in deep learning and techniques for effectively training deep networks.

Course Details

• Course Code: NDAK24002U
• Credit: 7.5 ECTS
• Level: Full Degree Master
• Duration: 1 block
• Placement: Block 2
• Schedule: C
• Course Capacity: No limitation – unless you register in the late-registration period (BSc and MSc) or as a credit or single subject student.
• Language: English

Learning Outcomes

The course aims to equip students with:

• Knowledge of:
  • Convolutional neural networks
  • Transformers
  • Message passing and graph neural networks
  • Generative neural networks such as variational autoencoders
  • Basic strategies for interpretability of deep neural networks
  • Training methodology
• Skills to:
  • Select appropriate methodology to solve deep learning problems
  • Implement selected deep learning algorithms
  • Design and train deep learning algorithms
• Competences to:
  • Reflect upon the capabilities and limitations of deep learning algorithms
  • Recognise and describe possible applications of deep learning methodology
  • Design, optimise and use deep models
  • Apply the learned methodology to applications in analysis of real-world data such as images, sound and text
  • Analyse deep learning algorithms

Recommended Academic Qualifications

• Linear algebra corresponding to the course Lineær Algebra i datalogi (LinAlgDat)
• Calculus corresponding to the courses Introduktion til matematik i naturvidenskab (MatIntroNat) and Matematisk Analyse (MatAn)
• Basic statistics and probability theory corresponding to the course Sandsynlighedsregning og statistik (SS)
• Machine learning corresponding to Machine Learning A (MLA)
• Programming experience in Python

Teaching and Learning Methods

The course will mix lectures, exercise classes, and project work.

Workload

• Lectures: 32 hours
• Preparation: 68 hours
• Exercises: 46 hours
• Exam: 60 hours
• Total: 206 hours

Assessment

• Type of Assessment: Continuous assessment
• Type of Assessment Details: Continuous assessment of 3-4 written assignments. All assignments must be passed. The final grade is based on an overall assessment.
• Aid: All aids allowed. For programming tasks specifically, this includes AI-based programming tools such as GitHub Copilot or similar.
• Marking Scale: 7-point grading scale
• Censorship Form: No external censorship. Several internal examiners.
• Re-exam: The re-exam is 25 minutes oral examination, without preparation, in full course syllabus.

Course Information

• Study Board: Study Board of Mathematics and Computer Science
• Contracting Department: Department of Computer Science
• Contracting Faculty: Faculty of Science
• Course Coordinators: Stefan Sommer

Additional Information

The course is similar to the discontinued courses Advanced Deep Learning (NDAK22002U and NDAB21009U) and Artificial Intelligence (NDAB20002U). Therefore, you cannot register for this course if you have already passed Advanced Deep Learning (NDAK22002U or NDAB21009U) or Artificial Intelligence (NDAB20002U).