DAT255 Deep learning engineering

Knowledge

• Able to explain fundamental concepts, architectures, and training algorithms in deep learning, including optimization, regularization, and neural network design principles.
• Able to explain how deep learning can be used to solve practical problems from a variety of domains.
• Able to explain the limitations and challenges of using deep learning for real-world tasks.
• Understanding of the difference between deep learning as a discipline in itself, and deep learning engineering. This includes how deep learning relates to software engineering more broadly and an understanding of machine learning engineering and MLOps, including model versioning, monitoring, and maintenance.
• Understanding of deep learning model evaluation metrics and validation strategies specific to different application domains.
• Understanding of the main drivers of progress in artificial intelligence.

Skills

• Able to solve concrete, practical problems from topic such as computer vision and natural language processing, using deep learning.
• Has hands-on experience with model deployment pipelines, including both local deployment and cloud-based solutions, with consideration for scalability and performance.
• Can develop and use modern, state-of-the-art software tools and frameworks for data analysis, machine learning, and visualization.

General competence

• Able to formulate and complete deep learning-based projects.
• Able to evaluate and select appropriate deep learning approaches for given problems, considering both technical and practical constraints.
• Able to present deep learning projects through reports and technical documentation, including well-structured code, experimental results, and performance analyses.

The student should have programming skills at the bachelor level in a computer science or computer engineering program. Experience with Python will be a significant advantage. Exposure to core parts of software engineering will be advantageous. Several online resources for acquiring the recommended background knowledge will be provided at the beginning of the course.

You should have completed at least one course in machine learning at the level of DAT158 (HVL) or INF264 (UiB).

Lectures, discussions and hands-on labs, where you will work on topics covered by the lectures and your course project, assisted by the lecturer and fellow students.

In Førde and Haugesund, the course will be given as a fully digital offering.

The exam has two parts:

1. A group project report that counts for 50% of the final grade
2. Written school exam, 2 hours, that counts for 50% of the final grade.

Both parts of the exam must result in a passing grade to get a final grade in the course. If a student fails one of the parts, the part can be retaken separately.

Grading scale is A-F where F is fail.