Silver bullets are for werewolves: Exploring and comparing deep learning with traditional machine learning methods

In this topic the master student will explore and compare deep learning with other machine learning methods - when does it work, and when not?
Master

It seems that nowadays deep learning is used for a lot of different problems and that it is very popular. Nevertheless, deep learning cannot be seen as the silver bullet for all problems and comes with a lot of pitfalls that user often are not aware of.
In this topic the master student will explore and compare deep learning with other machine learning methods. Based on previous research and use cases he will especially focus on finding the transition from when deep learning makes sense and works and when it is useless or a waste of resources.

Goal

A general and large scale comparison between deep learning and standard machine learning methods.

Learning outcome

Critical thinking, Machine learning skills

Qualifications

Critical thinking, good programming and mathematical skills

Supervisors

  • Pål Halvorsen
  • Michael Riegler