Master's level
- Bachelor of Science (at least 180 higher education credits) in computer science or related subjects such as mathematics, informatics, telecommunications, electrical engineering, physics.
- Knowledge equivalent to English 6 at Swedish upper secondary level.
- At least 15 credits in programming.
- At least 7.5 credits in mathematics.
- Passing grade in the course Statistical methods for Data Science (MA660E)
Participation in the course also requires knowledge obtained from the course Artificial intelligence for Data Science (DA631E).
The course is part of the programme Computer Science: Applied Data Science, master’s programme, and can be included in the master's degree in computer science (120 credits).
The course contains the following elements:
- Data transformation, Data Augmentation, adjustment/calibration of model parameters (including Advanced Feature Extraction, Hyper-parameter Optimisation)
- Interactive machine learning methods (including Human-Machine Collaboration, Active Learning, Online learning, Incremental Learning, Learning from Data Streams)
- Meta-learning algorithms and Ensemble Methods
- Advanced algorithms for supervised learning and unsupervised learning (with emphasis on discriminative and generative Deep Learning architectures)
- Reinforcement learning (including Policy Search, Policy Iteration, Value Iteration, Q-learning)
- Trends and current front line research in machine learning
**Knowledge and understanding
**For a passing grade the student shall be able to:
- Explain advanced machine learning methods and how they are used in practice
Competence and abilities
For a pass grade the student shall be able to:
- Implement advanced machine learning algorithms, both individually and in groups
- Assimilate and use published research results in machine learning
- Explore recent developments in commercial machine learning applications
- Evaluate and compare the suitability of different methods for addressing a given problem
- Interpret the relevance of machine learning results
Evaluation abilities and approach
For a pass grade the student shall be able to:
- Analyse and evaluate academic publications in machine learning
- Critically analyse strengths and weaknesses in scientific arguments for both theoretical and experimental results.
Lectures, data laboratories, seminars, project work (Kaggle competition)
The students' achievements are assessed through a report on group projects (5 credits, UG), laboratory assignments (5 credits, UG) and written examination (5 credits, UA).
An A-E passing grade requires that all parts have been completed and passed.
The final grade is based on the written examination.
- Aurlien Gron. 2017. Hands-On Machine Learning with Scikit-Learn and Tensorflow: Concepts, Tools, and Techniques to Build Intelligent Systems (1st ed.). O'Reilly Media, Inc.
- Han, J., Kamber, M., & Pei, J. (2011). Data mining: Concepts and techniques (3rd ed.). Waltham: Morgan Kaufmann.
- Ian Goodfellow, Yoshua Bengio, and Aaron Courville. 2016. Deep Learning. The MIT Press.
- Russell, Stuart Jonathan & Norvig, Peter (2010). Artificial intelligence: a modern approach. (3rd ed.) Boston: Pearson Education.
- Witten, Ian H., Frank, Eibe & Hall, Mark A. (2011). Data mining: practical machine learning tools and techniques (3rd ed.) Burlington, MA: Morgan Kaufmann.
• A collection of scientific articles will bed added to the above mentioned literature.
The University provides students who are taking or have completed a course with the opportunity to share their experiences of and opinions about the course in the form of a course evaluation that is arranged by the University. The University compiles the course evaluations and notifies the results and any decisions regarding actions brought about by the course evaluations. The results shall be kept available for the students. (HF 1:14).
When a course is no longer given, or the contents have been radically changed, the student has the right to re-take the examination, which will be given twice during a one year period, according to the syllabus which was valid at the time of registration.
If a student has a Learning support decision, the examiner has the right to provide the student with an adapted test, or to allow the student to take the exam in a different format. The syllabus is a translation of a Swedish source text.