Course syllabus

Teachers

Bernhard Mehlig (lecturer and examiner)  Bernhard.Mehlig at gu.se
Johann Flemming Gloy (TA)
Thorsteinn Freygardsson (TA)
Isak Bengtsson (TA)

News

The course starts on Thursday September 4 (not earlier as previously advertised).

The course book is availabe at ChalmersStore and CUP. You can take the printed book to the exam. It is your responsibility to order the book in time if you want to have it in the exam.

Course representatives

TBD

Welcome

Welcome to Artificial Neural Networks 2025. How to prepare? Login to the online system OpenTA (link on the menu on the left) to do the preparatory linear-algebra exercises. Sign up for the discussion forum and pose any OpenTA questions there.

Signup to discussion forum.

Link to discussion forum.

Important dates

First lectures  on  September 4 at 13 and 15 in KE.

Deadlines for homework problems: see OpenTA.

Contents

1. Introduction

Part I Hopfield models
2. Deterministic Hopfield networks
3. Stochastic Hopfield networks.
4. The Boltzmann distribution

Part II Supervised learning
5. Perceptrons
6. Backpropagation
7. Deep learning
8. Convolutional networks
9. Recurrent networks

Part III Learning without labels
10. Unsupervised  learning
11. Reinforcement learning

Preliminary schedule

See TimeEdit for schedule.

Lecture 1 Sept. 4 13:15 Introduction (Chapter 1, Slides) Bernhard Mehlig
Lecture 2
Sept. 4 15:15 Hopfield model, one-step error probability (Chapter 2)  Bernhard Mehlig
Lecture 3  Sept. 5 13:15 Hopfield model continued (Chapter 2) Bernhard Mehlig
                                
For discussion of energy function, see separate video.
Lecture 4 Sept. 9 13:15 Hopfield model (Chapter 3)  Bernhard Mehlig
                              
For derivation of mean-field theory, see separate video
                               For calculation of critical storage capacity, see separate video

Lecture 5
Linear algebra (repetition, summary) Bernhard Mehlig
Lecture 6 Monte-Carlo simulation (Chapter 4)  Bernhard Mehlig
Lecture 7  Boltzmann machines (Chapter 4)  Bernhard Mehlig
Lecture 8  Simple perceptrons (Chapter 5)  Bernhard Mehlig
Lecture 9 (Sep. 17) 13:15  Sections 5.3, 5,5, 5.6  Bernhard Mehlig
                               
For capacity of simple perceptron, see separate video.
Lecture 10 (Sep. 19) 15:15 Section 6.1, 6.2, 6.3 Bernhard Mehlig
Lecture 11 (Sep. 20) 15:15  Sections 6.4, 6.5  Bernhard Mehlig
Lecture 12 (Sep. 24) 13:15  Section 7.1  Bernhard Mehlig
Lecture 13 (Sep. 26) 13:15 Section 7.2, 7.3, and 7.4   Bernhard Mehlig
Lecture 14 (Oct. 1) 13:15 Section 7.5   Bernhard Mehlig 
                          
Please read Chapter 8 at home. I'll take questions on Oct. 4 and 8.

Lecture 15 (Oct. 3) 15:15 Section 7.6  Bernhard Mehlig

Lecture 16 (Oct. 8) 13:15 Section 9.1 Bernhard Mehlig
Lecture 17 (Oct. 10) 13:15 Sections 9.2 and 9.3  Bernhard Mehlig
Lecture 18 (Oct. 15) 13:15 Reservoir computers (Section 9.5) Ridge regression)
                         and Summary unsupervised learning  Bernhard Mehlig
Lecture 19 (Oct. 15) 15:15   Sections 10.1, 10.2 Bernhard Mehlig
Lecture 20 (Oct. 17) 13:15   Section 10.3 Bernhard Mehlig
Lecture 21 (Oct. 18) 13:15  Sections 10.4, 10.5, and 10.6, Bernhard Mehlig
Lecture 22 (Oct. 22) 13:15  Chapter 11  Bernhard Mehlig

Lecture 23 (Oct. 24) 15:15  Chapter 11 continued (PDF), transformers (PDF) and exam preparation (in particular energy function in Hopfield model, pp. 27, 28) Bernhard Mehlig

Exercise classes (homeworks and exam preparation to be determined)

Chapters and Sections refer to the course book below.

Course book

B. Mehlig, Machine learning with neural networks, Cambridge University Press (2021).
Errata for Machine Learning with Neural Networks (October 18, 2022)

Examination

Credits for this course are obtained by solving the homework problems (solutions of examples and programming projects) and by a written examination. There are three sets of homework problems. Each of the three gives at most 3 points. The exam gives at most 15 points, resulting in a maximum of 24 points.

To pass the course,  it is necessary to obtain at least 6 points in the written exam, and to have at > 13.5 points in total.

Passing grades:
Chalmers: 3: >13.5p; 4: >17p, 5: >21.5p
GU: G: >13.5p; VG: >19.5p
ECTS: C: >13.5p; B: >17p; A: >21.5p

OpenTA

This course uses the OpenTA online system for exercises, homework, and exam preparation.

Rules for homework submissions

Same rules as for written exams apply: it is not allowed to copy any material from anywhere unless reference is given: all sources must be stated in writing (e.g. old solutions, fellow students, internet, ChatGPT).  All students must write their own computer programs and submit their own solutions and program code via OpenTA.

Keep a backup of your solutions to the OpenTA questions, of your submitted PDF files as well as the answers you typed in. The system does not store your answers after December 2025. If you take a re-exam in January or August 2026 you will be asked to re-submit all answers and OpenTA scores.

Your OpenTA points are valid for the two re-exams in January and August 2026. Please contact any of the teachers if you need guidance for your exam preparation, or if you have questions about the coming re-exams. To pass the course in future academic years (for instance 2026/2027) you need to redo the OpenTA problems for that academic year.

For some OpenTA problems, you are asked to submit your answer in the form of a PDF file. It must be a single A4 page with 12pt single-spaced text, and with 2cm margins. LateX template. The page may contain at most one Figure and/or one Table with the corresponding Figure and/or Table caption, in addition to the text discussing the results shown in the Figure/Table. It is not necessary to write a full page, but you must explain/describe what you have done, clearly state your results/answers, your conclusions, and cite the sources used. When necessary, you must discuss possible errors and inaccuracies in your results. If you are asked to plot results/make graphs, you do this in a Figure with legible axis labels and tic labels. All symbols and lines must be explained in the Figure or in a caption. The Figure may consist of separate panels. Refer to them as 'left panel', 'right panel', 'bottom panel', etc. (or alternatively label them 'a', 'b',...).

In addition you need to upload a PDF file with the computer code you used to generate your results. No length restriction applies for this file.

OpenTA forwards your submissions to URKUND.   Information about the URKUND system can be found here.

Deadlines are sharp. Late submissions are not accepted.

Rules for homework resubmissions

Most homework tasks are automatically graded. For these tasks, you only get the green light from OpenTA. You can try as often as you want (before the deadline).

Some homework tasks require you to submit a one-page PDF file with your results and discussion. These are manually graded. If we find errors or problems in these solutions you will be notified (text message on OpenTA) and you are allowed to resubmit. The final deadline for resubmissions is to be determined, a few days after the exam. After that date, the system does not accept any resubmissions. If you resubmit before that deadline and your solution is correct, you get full points for that task.

Written exam

Allowed material for exam: course book, only printed by Cambridge University Press, no written annotations (underlining/highlighting allowed).

The exam covers the material presented in the lectures as well as the homework problems. Old exam questions are given in the course book. Solutions.

Old exams with solutions
October 2023
August 2023
January 2023
October 2022
October 2021
January 2021
October 2020

Old exams without solutions
August 2024
January 2024

January 2020
October 2019
January 2019

Note: from October 2022 onwards, students are allowed to have the course book (as printed by CUP) in the exam. 

Date for written exam, deadline for registration for exam. Please follow this link. Course code FFR135.

If date & time of the exam collide with another exam you must take, then you must follow the steps outlined here.

If you don't pass the exam

Your OpenTA points are valid for the two re-exams in January and August 2026. Please contact any of the teachers if you need guidance for your exam preparation, or if you have questions about the coming re-exams. To pass the course in future academic years, you need to redo the OpenTA problems for that academic year.

Changes from last year

 

Course summary:

Date Details Due