Computer Engineering MA, Machine Learning with Visual Media Applications, 6 credits

Aim

The aim of this course is to equip students with theoretical knowledge and practical skills in state-of-the-art machine learning (ML) and deep learning (DL) models for visual media applications and real-world research challenges. Through lectures and hands-on projects, students will develop a strong connection between theory and practice within visual ML/DL fields relating to image and video processing, computer vision, generative AI, and more.

Course objectives

Upon completion of the course the student should be able to:

• Understand and assess deep learning models and techniques for visual media applications.
• Develop, optimize, and fine-tune deep neural networks, including models such CNNs, visual transformers, VAEs, GANs, and others.
• Apply advanced ML/DL methods and utilize transfer learning strategies to adopt architectures for specialized visual media applications.
• Design and execute deep learning-based projects for real-world and research applications.
• Critically evaluate research trends and ethical considerations in ML/DL applications for visual media.

Content

Introduction to state-of-the-art trends and fundamental methods, architectures (CNNs, visual transformers, diffusion models, VAEs, GANs, etc.) as well as techniques for optimization and fine-tuning, with a balanced emphasis on theoretical underpinnings and practical approach.

Critical examination of transfer learning, data augmentation, and hyperparameter optimization for specialized applications, highlighting trade-offs and performance benchmarks.

Efficient and scalable ML techniques for handling large-scale visual data, including model compression, pruning, and deployment strategies.

Examples from various visual AI application areas illustrate both strengths and limitations of current methods. Project-based implementation and evaluation of ML/DL solutions reflecting real-world challenges within the field.

Review of current research trends and ethical aspects in ML/DL for visual media applications.

Entry requirements

Computer Engineering or Electrical Engineering, 60 credits, including programming, 10 credits, Signal and Image Processing, 6 credits, Computer Vision and Multiple View Geometry, 6 credits, Neural Networks and Deep Learning, 6 credits, and Image Analysis, 6 credits

Mathematical subjects, 30 credits, including Probability Theory and Statistics, and Linear Algebra.

Selection rules and procedures

The selection process is in accordance with the Higher Education Ordinance and the local order of admission.

Teaching form

The course is delivered through lectures and project work. Lectures cover state-of-the-art machine learning models for visual media, while project work allows students to apply these concepts to a research-oriented problem. Independent study is essential, with students expected to engage with lecture materials, research papers, and project tasks with limited supervision.

Teaching can take place in Swedish or English.

Examination form

P101: Project, with written report, 3 Credits
Grade scale: Seven-grade scale, A-F o Fx

T101: Written exam, 3 Credits
Grade scale: Seven-grade scale, A-F o Fx

Link to grading criteria: https://www.miun.se/gradingcriteria.

The examiner has the right to offer alternative examination arrangements to students who have been granted the right to special support by Mid Sweden University’s disabilities adviser.

Examination restrictions

Students registered on this version of the syllabus have the right to be examined 3 times within 1 year according to specified examination forms. After that, the examination form applies according to the most recent version of the syllabus.

Grading system

Seven-grade scale, A-F o Fx