Machine Learning – (M.Sc.)

Contents

Machine Learning is an area within Computer Science where computer systems are designed to learn from large sets of examples, similar to the learning strategies of biological systems (like humans). Recently, Machine Learning has gained great importance for the design of search engines, robots, and sensor systems, and for the processing of large scientific data sets. The focus of the programme is on mathematical foundations and methods for Machine Learning.

The programme consists of a basic curriculum followed by one of two tracks: (i) Perception and Cognition, and (ii) Information Retrieval.

Track I: Perception and Cognition

An important aspect of an intelligent system is the ability to observe and understand its environment. All intelligent biological systems have this ability; we humans perceive the world through vision, hearing, taste, smell and touch. Sensing for artificial systems is developed within the fields of Speech Understanding and Image Understanding, also known as Computer Vision.

Machine Learning has come to dominate algorithmic construction in both these fields. The reason for this is the complexity and uncertainty of visual and acoustic signals. In the Perception and Cognition track, the student will be acquainted with different Machine Learning methods for artificial perception, in particular Computer Vision, which is a large and active field of research at KTH.

Another focus of the Perception and Cognition track is the relationship between Machine Learning and Neuroscience, where the functionality of neural systems, such as the human brain, is studied.

Track II: Information Retrieval

During the last decade, there has been an explosive growth in the amount of data available, both on the Internet and in applications such as Economics, Medicine, Meterology, Geoscience, and Astronomy.

Machine Learning has come to be an effective tool for developing methods to search and organise large volumes of data. This process is often referred to as Data Mining, Knowledge Discovery or Information Retrieval. A number of successful companies, such as Google, build their entire product on Machine Learning methods for Information Retrieval.

In the Information Retrieval track, the student will be acquainted with the mathematical foundations of Machine Learning for Information Retrieval, but also with technology for database construction and data security.

Year 1

Semester 1

Compulsory courses:

* Philosophy of science and research methodology
* Artificial intelligence
* Machine learning
* Image analysis and computer vision

Semester 2

Track I: Perception and Cognition

* Neuroscience
* Image based recognition and classification
* Artificial neural networks and other learning systems
* Elective courses

Track II: Information Retrieval

* Statistical methods in applied computer science
* Modern database systems and their applications
* Information retrieval
* Elective courses

Year 2

Semester 1

Track I: Perception and Cognition

* Computational photography
* Elective courses

Track II: Information Retrieval

* Elective courses

Semester 2

Degree project
A part of the second year is dedicated to a degree project of 30 ECTS credits. The purpose of the degree project is for the student to demonstrate the ability to perform independent project work, using the skills obtained from the courses in the programme.