Statistical Deep Learning

The course treats basic as well as modern concepts of statistical learning in terms of artificial neural networks (deep learning), with applications in statistical data analysis.

Topics treated include feedforward networks, regularization and optimization of networks with many layers, convolutional networks, recurrent networks and validation methods. In addition, mathematical interpretations of networks are given, such as nonlinear regression with different link functions for the outcome variable. The course includes some of the following topics; autoencoders, representation learning, deep generative methods, and information theoretic concepts of deep learning.

The course consists of two parts: theory and hand-in assignments.


Teaching Format

Instruction is given in the form of lectures, exercise sessions and supervision.


Assessment

The course is assessed through a written exam and home assignments.

Examiner

A list of examiners can be found on

Exam information

The schedule will be available no later than one month before the start of the course. We do not recommend print-outs as changes can occur. At the start of the course, your department will advise where you can find your schedule during the course.


Note that the course literature can be changed up to two months before the start of the course.

Goodfellow, Bengio, Courville: Deep Learning. MIT Press.

List of course literature Department of Mathematics

Course reports are displayed for the three most recent course instances.

New student
During your studies

Course web

We do not use Athena, you can find our course webpages on kurser.math.su.se.