African-European Masterclasses in AI and Computational Thinking
This course is a unique opportunity for PhD students and Masters students (looking to go on to pursue a research career) to learn about the latest methods in AI, machine learning and computational thinking from leading academics from throughout Africa and Europe. Starting with methods in data science (illustrated in the game of football); moving through questions of responsible AI and the methods behind large language models; getting into the details of parallel and quantum computing; understanding the mathematical theory behind machine learning and decision-making; and looking at ethical and legal aspects of AI – this course offers insight into everything you need to know about this fast-changing landscape.
The lectures are given by leading experts in both Africa and Europe, with the contents reflecting the current thinking on these two continents.
The teaching will take place (roughly) every third Wednesday, starting from November 1st 2023. We welcome members of staff at institutions to locally organise students to participate.
Please take your laptop to the lesson and have an up to date installation of Phython and R. We will actively work on problems during the Masterclass.
Participation
The course is aimed at PhD students and Masters students interested in pursuing a career in research either in academia or industry. To participate you must be part of a local group at your University. You can alert your supervisor to this possibility if you are a PhD student.
The local organiser, a member of staff at the University, will book a (physical) classroom for the lessons and organise the course for students. Lessons will then be in hybrid form between the local classrooms across Africa and Europe.
As a student you should register by contacting your local organiser. These are as follows:
Uppsala University: david.sumpter@it.uu.se
Stellenbosch University: petruccione@sun.ac.za
Addis Ababa University: solomong@aau.edu.et
Makerere University: joyce.nabende@mak.ac.ug
Rhodes University: o.tastanbishop@ru.ac.za
University of Lagos: vodumuyiwa@unilag.edu.ng
University of Nairobi: pweke@uonbi.ac.ke
University of Ljubljana: [contact required]
University of Rwanda: dndanguza@gmail.com
University of Warwick: olalekan.uthman@warwick.ac.uk
University of Dar Es Salam: watsonlevens@yahoo.com
If you are a staff member at an African or European University and would like to give your students a chance to participate, please email Kajsa Hallberg-Adu (kajsa.hallberg.adu @ antro.uu.se). It is also possible to participate in a single master class but to do so a local institutional organisation is still required.
Learning form
The Masterclass will then take place over Zoom, with the following format.
- Lecture (45 minutes): Introducing area by the lecturer.
- Problem (15 minutes): Presenting task for students.
- Work locally in groups (90 minutes): Work on the task presented.
- Summary (30 minutes): Go through what has been learnt. Give material for hand-ins.
Half-day Masterclasses will involve the above programme. Full-day Masterclasses will involve iterating over this format twice (before and after lunch).
Students pass the course as a whole by completing hand-ins for 5 out of 8 Masterclasses (or according to a decision by a local organising institution).
The Masterclasses
Data science and modelling of football
Date: 1 November 2023, 10:15-13:00 Central Africa Time (GMT+2)
9:15-12:00 Central European Time (GMT+1)
A half-day introduction to using machine learning and modelling in football for data Scientists, PhD students and industry professionals. We will go through how to use logistic regression to fit an expected goals model of quality of chances. We will go on to look at models of pass success and value using other ML methods. By the end of the day you will understand how to use football data and to fit simple models to it. This lecture will also provide an introduction to the challenges of putting data science in to practice.
Physical location: 103150, Ångström.
Register for Zoom: here.
Teacher: Prof. David Sumpter (Uppsala University)
Responsible AI, bias and fairness
Date: 17th November 2023, 10:30-13:00 Central Africa Time (GMT+2)
9:30-12:00 Central European Time (GMT+1)
A half-day introduction to responsible AI and bias in machine translation models for African languages, in particular. There is significant growth in the area of building Natural Language Processing (NLP) tools for African languages, despite the lower availability of explicit text data than for English. In this masterclass we introduce concepts on Responsible AI and then dive into building models for machine translation models for African languages. We then discuss the approach taken to evaluate and quantify the gender bias within a Luganda-English machine translation system using Word Embeddings Fairness Evaluation Framework (WEFE).
Physical location: 100155, Theatrum Visuale, Ångström
Register for Zoom: coming soon
Teachers: Dr. Joyce Nakatumba-Nabende (Makerere University) and Dr. Andrew Katumba (Makerere University)
High-performance computing architecture and parallel computing
Date: 29th November 2023, 10:15-13:00 Central Africa Time (GMT+2)
This workshop introduces the Fundamentals of High-Performance Computing (HPC). High-Performance Computing as a discipline of Computational Science is focused on a structured approach to obtaining computational results faster even in the face of an ever-increasing complex problem-solving domain. Consider the process of predicting (forecasting) the weather over a single city or a completely different problem of vehicle and road traffic planning and management for a city. This workshop aims to give the participants a deep insight into the architecture of current HPC systems and introduce the challenges and solutions in obtaining optimal performance on a large number of processors. We will also show how to write parallel code for processor-intensive applications to be run on clusters, the grid, the cloud, or shared infrastructure. After the workshop, participants will be familiar with various computer architectures and how the hardware components affect performance. They will be able to identify performance challenges and the best architecture for different types of scientific computation. The participants will also be able to select the best compiler and optimization strategy for a different type of scientific computation.
Teachers: Professor Onime Clement (ICTP) and Dr Solomon Gizaw (Addis Ababa University)
A framework for developing an AI gender auditing tool
Date: 13th December 2023, 9:15-13:00 Central Africa Time (GMT+2)
A half-day workshop on the phases involved in developing an AI gender audit tool. The framework lends itself to a number of real life scenarios. This workshop will benefit everyone interested in or currently conducting or intending to conduct research in equitable AI, and this includes but is not limited to Data Scientists, AI Researchers, Students, Academics, Industry Professionals and AI enthusiasts.
Teachers: Dr Mary Akinyemi, Dr Chika Yinka-Banjo, Dr Olasupo Ajayi (University of Lagos)
Further Masterclasses
The following Masterclasses will take place in 2024. Time and place to be arranged.
Introduction to Quantum Computing by Prof. Francesco Petruccione (Stellenbosch University and National Institute for Theoretical and Computational Sciences). An introduction to quantum computing assuming no prior knowledge of quantum theory. By the end of the day, you will understand what quantum computing is and be able to program simple algorithms.
Mathematical modelling in biology by TBA .
Argumentation and Visualization with AI by Prof. Onime Clement and Sarah Bianca.
Deep learning and neural networks by Dr. Niklas Wahlström (Uppsala University). This course will show how to build your own deep learning model.
Large Language Models in AI by Dr. Solomon Gizaw , Dr Yaregal Assabie and Dr. Victor Odumuyiwa (Addis Ababa University). An introduction to recent developments in language models. The training covers topics such as the basics of large language models, in-depth reviews of models like BERT, T5, and GPT, promoting language models, scaling and risks, retrieval-based language models, and multimodal language models.
AI in HealthCare by Dr. Joyce Nakatumba-Nabende, Dr. Andrew Katumba (Makerere University) A half-day introduction to building computer vision models for health care.
AI Policy by Dr. Chika Yinka-Banjo (University of Lagos)