Problem Solving in Machine Learning (15 September - 3 October 2025)

Claire David (AIMS South Africa) and Harrison Prosper (Florida State University)

This course introduces students to the key concepts of machine learning: linear regression, gradient descent, logistic regression, regularisation, over/underfitting. The most common algorithms in supervised learning will be presented, from boosted decision trees to deep neural networks. Students will code from scratch the algorithms as well as the methods to assess the performance of their model (ROC, train/validation learning curve, bias/variance). They will learn how to improve their program through hyperparameter optimization and advanced techniques (momentum, scheduler). In unsupervised machine learning, the approaches of dimensionality reduction and clustering will be introduced and illustrated.

Bayesian Inference (15 September - 3 October 2025)

Sanjoy Mahajan (Lund University)

Bayesian probability and statistics, with applications in mathematics, science, engineering, and philosophy of science. The approach emphasises the mathematical description of, and inference from, incomplete information. Topics include: nature of probability, conditional probability, Bayes' theorem, prior and posterior probabilities, Bayes factors, discrete and continuous distributions, point and compound hypotheses, hypothesis testing, likelihood principle, Shannon information, history of the Bayesian approach, and comparison with orthodox (frequentist) statistics. Applications include: medical testing, the problem of old evidence, the reproducibility crisis, drug evaluation, sequential sampling, p values, confidence intervals, legal evidence and reasoning, the Monty Hall problem, and plausible reasoning in mathematics.

Introduction to Quantum Computing (6-24 October 2025)

Ryan Sweke (AIMS South Africa)

This course will introduce students to the basic ideas and formalism of quantum computing. In the first week we will cover the basic formalism of both single classical and quantum systems, as well as combinations of multiple such systems. This will then allow us to define the notion of entanglement, and demonstrate some of the unique features and properties of entangled quantum systems. In the second week, we will then shift our focus to quantum algorithms, and we will cover some of basic quantum algorithms, from the Deutsch-Joza algorithm through to Grover’s algorithm and the variational quantum eigensolver. Finally, in the third week, we will take a physicist's perspective, and reframe what we learned in the first week through the eyes of a physicist. This will then allow us to give a formal description of many-body quantum systems, and introduce a variety of learning theoretic problems associated with many-body quantum systems.

Machine Learning Operations (MLOps) (6-24 October 2025)

Oliver Angelil (Ishango.ai), Cyrille Feudjio (Ishango.ai) and Vincent Frimpong (Ishango.ai).

This course covers essential applied skills to bridge the gap between academia and industry in the year 2025. We follow best coding practices (e.g. PEP8; CI/CD; type-hinting) and industry standards to build useful machine learning applications that are deployed and hosted on Google Cloud Platform. Our lecture series covers technical debt; microservice architectures and containerization; RESTful APIs; basic User Interfaces (UIs); Data Engineering; distributed versus single node compute; the machine learning lifecycle; and agentic AI.

Applied Machine Learning at Scale (3 - 21 November 2025)

Ulrich Paquet (Google DeepMind, AIMS South Africa)

ML and AI drive the back-ends and front-ends of many large online companies, and are set to play a transformative role in the “internet of things”. This is a practical module that looks at how ML is applied to internet-scale systems. In this module, students will build their own recommender systems from scratch. Topics covered will include A/B testing, ranking, recommender systems, and the modelling of users and entities that they engage with online (like news stories).

Monte Carlo Methods for Machine Learning (3 - 21 November 2025)

Hugo Touchette (Stellenbosch University)

This course gives an introduction to Monte Carlo and stochastic approximation methods with applications in machine learning. The focus is on Markov chain Monte Carlo methods used for sampling probabilistic models and stochastic gradient descent methods used for minimisation.

Natural Language Processing and Large Language Models (24 November - 12 December 2025)

Shamsuddeen Hassan Muhammad (Imperial College, London) and Idris Abdulmumin (University of Pretoria)

This course introduces Natural Language Processing (NLP) and transformer-based Large Language Models (LLMs). Students will explore foundational NLP concepts, including tokenization, word embeddings, and language modelling. They will learn the core mechanics of LLMs, such as architecture, training, fine-tuning, reasoning, evaluation, and deployment strategies. The curriculum includes practical applications such as text classification, machine translation, summarization, and zero-/few-shot prompting. Through hands-on work with real-world datasets, students will design NLP pipelines and evaluate model performance in multilingual settings, with particular emphasis on low-resource and under-represented languages. By the end of the course, students will also build a simple language model from scratch.

Bio-acoustics (24 November - 12 December 2025)

Tom Denton (Google DeepMind) and Eunice Manyasi (Strathmore University)

Bioacoustics is the study of the sounds of animals. In recent years, it has become possible to collect thousands (or even millions) of hours of audio automatically in natural spaces, allowing unprecedented views into wildlife ecology. Extracting knowledge from this data requires a combination of machine learning and statistics, and close collaboration with domain experts in animal communication. In this course, we will learn about the training, evaluation, and application of a bioacoustics foundation model to novel problems, the geometry of high-dimensional embedding spaces, and unique statistical problems. The course will focus on a combination of reading and discussing core papers, as well as implementation of key ideas in computer lab sessions.

Neuroscience and AI (12-30 January 2026)

Christopher Currin (inait.ai) and James Ferguson (Institute of Science and Technology, Austria)

This intensive three-week programme provides a foundational understanding of NeuroAI for researchers in machine learning and mathematics. The course bridges the gap between biological neural systems and artificial intelligence, exploring how principles of brain function have inspired and can continue to drive innovation in AI. By the end of this course, students will be able to critically evaluate AI models from a neuroscientific perspective and identify promising avenues for biologically-inspired research.

Explainable AI (XAI) (12-30 January 2026)

Inga Strümke (Norwegian University of Science and Technology)

This course introduces explainable artificial intelligence (XAI), and covers:

• Taxonomy and terminology of XAI, and discussions of interpretable models.

• Local and global feature importance attribution methods, model-specific XAI methods.

• Basics and implementation of: counterfactuals, local and global surrogate models including LIME, the Shapley decomposition and SHAP, PDP, ALE, LOFO, SAGE, CAV and TCAV, attention weights interpretation and Grad-SAM, CNN model interpretation, activation maximization and CAM, Grad-CAM and Integrated Gradients.

CUDA programming for NVIDIA GPUs (12-30 January 2026)

Mike Giles (University of Oxford)

This is a 3-week course to learn how to develop parallel applications to run on NVIDIA GPUs. The aim is that by the end of the course students will be able to write relatively simple parallel programs, and will feel confident to continue learning to use CUDA through studying the code samples provided by NVIDIA on GitHub.

Fluid Dynamics (12-30 January 2026)

Richard Katz (University of Oxford)

Fluids are all around us, from the air we breathe to the oceans that determine our climate and from oil that powers our industries to metals that are cast into machinery. The study of fluid dynamics requires sophisticated applications of mathematics and the ability to translate physical problems into mathematical language and back again. The course begins by building a fundamental understanding of viscous fluid flows in the context of unidirectional flows. In more general, higher dimensional flows, pressure gradients are generated within a fluid to deflect the flow around obstacles rather than the fluid being compressed in front of them, and an understanding of the coupling between momentum and mass conservation through the pressure field is key to the understanding and analysis of fluid motions. We will use simple experiments to illustrate and motivate our mathematical understanding of fluid flow. Prerequisite for the course is fluency with differential equations and vector calculus. No previous knowledge of fluid dynamics will be assumed.

AI for Climate Change (2-20 February 2025)

Neil Hart (University of Oxford), Shruti Nath (University of Oxford)

When was the last time you or your family was affected by a weather or climate event? Did the rains fail to come? Did too much rain come in one day? Did the heat make daily life difficult? Understanding the changing risk of such events is at the core of climate science research today. Building dynamical and statistical methods to better predict these events is the core of atmospheric prediction science. Fundamental research on how the Earth System works continues to support both climate science and prediction science.

In this module, we will explore the ways in which machine learning tools have been used to make sense of our weather and climate and how the rapid advances of modern AI techniques are accelerating advances across this research area. The goal of this module is to equip you with sufficient knowledge of the Earth System for you to find matches between your growing AI skill set and weather-climate problems.

In this module we will together explore, and reproduce, applications of unsupervised learning, computer vision and causal inference to such problems. Examples will be provided of new techniques in Explainable AI (XAI) to use AI as a research partner in climate science. Consideration will also be given to the very latest breakthroughs in weather prediction using GraphCast, AIFS, and similar.

Selected Topics in AI in Science (2-20 February 2026)

Kristofer Linton-Reid (Imperial College, London), Aidan Crilly (Imperial College, London), Dan Akarca (Imperial College, London), Sam Hillman (Imperial College, London), Ben Moseley (Imperial College, London), Elli Heyes (Imperial College, London), Marcus Ghosh (Imperial College, London), Sara Veneziale (Imperial College, London), Austin Mroz (Imperial College, London), Lloyd Fung (Imperial College, London), Calum Gabbutt (Imperial College, London), Jon Langford (Imperial College, London) and Laura Helleckes (Imperial College, London)

This course introduces a range of artificial intelligence (AI) and machine learning (ML) methods and explores how these approaches are applied to scientific research across different disciplines. Each week, a new topic will be presented by members of the teaching team, who are domain experts in fields such as chemistry, biology, physics, and mathematics.

The first week focuses on neural networks and simulators, followed by structured, generative and geometric machine learning in the second week. The final week covers Bayesian inference.

Each week includes a hands-on mini-project, allowing students to engage with Python code and apply the concepts learned in real time. By the end of the course, students will have gained both practical experience and a deeper understanding of how artificial intelligence can drive innovation in scientific discovery.

Week 1. Neural Networks and Simulators

By: Kristofer Linton-Reid, Aidan Crilly, Dan Akarca and Sam Hillman.

Week 2. Structured, Generative & Geometric Machine Learning

By: Ben Moseley, Elli Heyes, Marcus Ghosh and Sara Veneziale.

Week 3. (Bayesian) Inference

By: Austin Mroz, Lloyd Fung, Calum Gabbutt, Jon Langford and Laura Helleckes.

Matrix and Tensor Factorization Techniques for Machine Learning (2-20 February 2026)

Guillaume Rabusseau (Université de Montréal)

This course provides an in-depth exploration of tensor methods, starting with a foundation in matrix factorization techniques and progressing to advanced tensor decomposition methods and tensor networks. The course covers both theoretical and practical aspects, including computational algorithms, efficiency considerations, and real-world applications. Special attention is given to tensor networks, a powerful framework for representing and manipulating high-dimensional data, with applications in areas such as probabilistic modeling, neural network compression, and quantum system simulations. Through lectures, hands-on exercises, and discussions of open research problems, students will gain a comprehensive understanding of the role of tensors in modern computational sciences.

Objectives

• Develop a strong foundation in matrix factorization techniques and their applications in machine learning.

• Understand the fundamentals of tensor networks and tensor factorization, including key operations and decomposition models.

• Explore advanced tensor decomposition algorithms and their role in machine learning applications.

• Examine advanced topics such as computing gradients of tensor networks, simulating quantum systems, and open research problems in the field.

Reinforcement Learning (2-20 March 2026)

Arnu Pretorius (InstaDeep)

This course provides an introduction to Reinforcement Learning (RL). It will cover fundamental topics such as Markov decision processes, dynamic programming, Monte Carlo and value-based methods including temporal difference learning, function approximation and policy gradient methods. The presentation of topics will be roughly split equally between theory and code, using notebooks and additional material. Although the focus will be on understanding the foundations of RL, the course will conclude by giving an overview of some more advanced topics and applications.

Principles of Imaging for Radio Astronomy (2-20 March 2026)

Marta Spinelli (Université Côte d'Azur) and Landman Bester (South Africa Radio Astronomy Observatory; Rhodes University)

Our best theory for the evolution of the Universe is based on the existence of a “dark” form of matter that has only gravitational interactions and an unknown form of “dark” energy that causes its accelerated expansion. Unveiling the nature of the dark sector is the next fundamental question in Cosmology and requires mapping increasingly larger volumes of the observable Universe.

The SKA Observatory is an intergovernmental organisation for ground-based astronomy in charge of building and operating the world’s largest radio facilities. One of the two observatories will be in South Africa, in the Karoo Desert. The successful exploitation of the data from SKA and its precursors will shed new light on our understanding of the formation and evolution of our Universe. This new fundamental research needs on the one hand new and innovative data analysis techniques and, on the other hand, the construction of realistic end-to-end simulations, going from the sky emission to cosmological parameter constraints.

This course will cover the timely and innovative subject of Radio Cosmology with a particular focus on the role of neutral hydrogen in unveiling the large-scale structure of the Universe. It will also discuss how AI has improved the quality of data analysis and the construction of simulations.

Course outline

• Summary of the key concepts of Modern Cosmology: evidence in favour of the Big Bang, inflation, dark matter, and dark energy. Description of the main phase of the evolution of the Universe. Concept of distance.

• Basics of Bayesian Approach in Cosmology.

• A brief summary of how we simulate the Universe. AI methods to speed up this process.

• Radio Cosmology: what we can learn about the structure and evolution of our Universe using Radio Astronomy data. Description of the SKA Observatory project. Review of the main cosmological probes accessible in the Radio band.

• 21cm Intensity Mapping: what is it, how we can do it, and what we can learn with it.

• ML techniques to analyze the data: distinguish the signal from the foreground emission.

AI for Biomedicine (2-20 March 2026)

Tiago Azevado (University of Cambridge)

This course introduces the application of artificial intelligence to biomedicine, structured around two complementary domains. The first part focuses on computer vision methods for digital pathology and related biomedical imaging modalities, exploring how deep learning models extract diagnostic insights from tissue images and spectral data. The second part turns to neuroscience, covering AI approaches for brain data such as structural and functional imaging, and graph-based representations of neural connectivity. Through these perspectives, students will learn how modern AI techniques can address key challenges across biomedical research, gaining both theoretical understanding and practical skills to apply machine learning methods to complex biological data.

Responsible AI (23 March - 10 April 2026)

Umang Bhatt (University of Cambridge)

This course equips students with the algorithmic and statistical tools needed to build AI systems that are fair, transparent, and effective in collaboration with humans. It is organised into three modules, each spanning one week. The first week covers fairness. Students will study fairness in classification and risk assessment. We will progress from statistical definitions to equality of opportunity. They will implement interventions such as post-processing and evaluate trade-offs in case studies like hiring. The second week addresses transparency and interpretability. Students will learn to audit black-box models using methods such as SHAP, LIME, and quantitative input influence. They will study transparency in practice, including datasheets, model cards, and nutritional labels, and examine how discrimination arises in online ad targeting and delivery. The third week covers human-AI collaboration. Students will study uncertainty estimation, calibration, and selective prediction as the foundations for deciding when a model should act and when it should defer. They will then examine learning-to-defer frameworks, routing policies, and human-AI complementarity, before turning to online methods, including contextual bandits, for personalisation. Students will be able to design collaborative systems, evaluate when models should defer to humans, and analyse the effects of AI assistance on judgment, trust, and skill retention. A distinguishing feature of this course is its holistic treatment of the machine learning pipeline: this covers practical workflows for deploying systems and considerations for responsible AI use more generally. 

The Evolution of Deep Learning (23 March - 10 April 2026)

Razvan Pascanu (Google DeepMind)

The course will focus on the evolution of neural networks, starting from the basics and going all the way up to modern LLM architecture. It will aim at providing both an account of how the field evolved over time, highlighting different milestones, but also cover most technical aspects required to understand modern machine learning tools. In particular, it will cover topics such as expressivity of neural networks, activation functions, optimisation and learning, signal propagation, generalization, strong generalization, MLPs, convnet, recurrent neural networks, graph neural networks and transformers, scaling laws, brief introduction to reinforcement learning, continual learning, meta-learning and in-context learning. For all of these concepts the aim is to provide an intuitive understanding of their role in a modern system, but also sufficient technical depth to allow students to debug architectures as well as to build their own.

Research Project (December 2025 soft start, for 13 April - 12 June 2026)