Overview of the sessions
- Dive into Machine Learning: Learning by Implementing – French & English
- Machine Learning for Biology: Learning the Language of Life
- Responsible AI
- Introduction to Probabilistic Thinking and Programming
- Recommender Systems or Why Your Phone Isn’t Actually Spying on You (kinda)
- Frozen Lake: An Icy Adventure Using Reinforcement Learning!
- Mathematics for Machine Learning – French & English
- LLMs for everyone
- Building your own Stable Diffusion
- Let’s map Africa! Introductory Tutorial to Geospatial Machine Learning
Dive into Machine Learning: Learning by Implementing – French & English
Description:
This tutorial offers an immersive exploration of the world of machine learning. Our primary goal is to demystify complex concepts, presenting them in a simplified manner. We adopt an interactive approach, fostering a gradual and intuitive understanding that enables learners to construct their very first machine-learning model step by step.
Aims/Learning Objectives: TBA.
Prerequisites: TBA.
Link: TBA.
Machine Learning for Biology: Learning the Language of Life
Description:
“Imagine a flashy spaceship lands in your backyard. The door opens and you are invited to investigate everything to see what you can learn. The technology is clearly millions of years beyond what we can make.
This is biology.”
–Bert Hubert, “Our Amazing Immune System”
In this tutorial we invite you to explore the language of proteins using machine learning.. No previous bio experience needed, just an open-mind and a curious spirit.
Aims/Learning Objectives: TBA.
Prerequisites: TBA.
Link: TBA.
Responsible AI
Description: TBA.
Aims/Learning Objectives: TBA.
Prerequisites: TBA.
Link: TBA.
Introduction to Probabilistic Thinking and Programming
Description:
Thinking probabilistically and working with probability distributions can be very powerful tools for any machine learning practitioner. Unfortunately, they are tools that are often disregarded due to their perceived complexity. In this practical we hope to demistify these ideas by building intuition, provided practical tips, and introducing a very powerful framework for embracing the probabilistic approach – probabilistic programming. We’ll both motivate why we need probabilistic programming and give an introduction for using it in practice.
This prac is aimed at all knowledge levels! No matter what your prior experience with probabilistic thinking and/or programming, we are sure that you will be able to take away some useful knowledge from this practical. However, this means that depending on your level, some of the content will not be aimed at you. Don’t worry, this will be clearly marked at all points. We reccomend that you try and stick to our suggestions in order to get the most out of this prac in the given time, but if curiosity gets the better of you that’s also great!
Aims/Learning Objectives:
- [Beginner] Understand what random variables and probability distributions are.
- [Beginner] Be able to work with probability distributions using numpyro.
- [Intermediate] Understand the difference between MLE, MAP, and Bayesian learning.
- [Advanced] Understanding the challenges involved in computing Bayes rule, and how probabilistic programming solved these.
- [Advanced] Be able to implement a simple probabilistic program with numpyro.
Prerequisites:
- Basic machine learning (e.g., simple supervised and unsupervised machine learning techniques).
- Basic calculus (e.g., computing an integral and taking derivatives to solve min/max optimisation problems).
- Python programming (with jax and numpy).
Link: TBA.
Recommender Systems or Why Your Phone Isn’t Actually Spying on You (kinda)
Description:
Recommender Systems are probably one of the most ubiquitous type of machine learning model that we encounter in our online life. They influence what we see in our social media feeds, the products we buy, the music we listen to, the food we eat, and the movies we watch. Sometimes they’re so good that people feel that their phone is spying on their conversations! In this prac, we hope to convince you that this isn’t the case (mostly) as well as taking you through some of the techniques popularly used in industry that recommends the content you see online.
Aims/Learning Objectives: TBA.
Prerequisites: TBA.
Link: TBA.
Frozen Lake: An Icy Adventure Using Reinforcement Learning!
Description:
In this tutorial, we will be learning about Reinforcement Learning, a type of Machine Learning where an agent learns to choose actions in an environment that lead to maximal reward in the long run. RL has seen tremendous success on a wide range of challenging problems such as learning to play complex video games like Atari, StarCraft II and Dota II.
We will show how RL can be used to help our agent cross a Frozen Lake, using several different RL approaches, ranging from tabular Q-learning to more modern methods, such as DQN (Deep Q-Networks). Along the way, you will be introduced to some of the most important concepts and terminology in RL.
Aims/Learning Objectives: TBA.
Prerequisites: TBA.
Link: TBA.
Mathematics for Machine Learning – French & English
Description: TBA.
Aims/Learning Objectives: TBA.
Prerequisites: TBA.
Link: TBA.
LLMs for everyone
Description:
Welcome to “LLMs for Everyone,” a practical exploration into the captivating world of Language Models! This entire block of text was crafted only by ChatGPT, showcasing the remarkable capabilities of these type models. Throughout this practical, we will delve into the underlying fundamentals of transformers, the powerful technology that drives models like GPT, and learn how to fine-tune and train our very own Large Language Models. Let’s embark on this exciting journey of understanding and creating LLMs, and discover how such impressive AI text generation is made possible!
Aims/Learning Objectives: TBA.
Prerequisites: TBA.
Link: TBA.
Building your own Stable Diffusion
Description:
Denoising Diffusion Models are a type of generative modelling which serves backbone of recent advances in image synthesis including Dall-E 2, Stable Diffusion, and Midjourney. These models utilise an iterative denoising process during inference to produce high quality samples. In this talk we explore the fundamentals of diffusion models, the intuition behind them, how they work in practice, and how they may be generalised to a wide range of applications.
Aims/Learning Objectives: TBA.
Prerequisites: TBA.
Link: TBA.
Let’s map Africa! Introductory Tutorial to Geospatial Machine Learning
Description:
Climate change is a pressing issue affecting the entire planet, with the Global South bearing a disproportionate burden of its impacts despite contributing less to its causes compared to more developed nations in the Global North. Extreme climate events such as droughts, floods, storms, and heatwaves have led to food insecurity and poverty. On the other hand, satellite imagery and machine learning (ML) can help address climate related challenges including food and water insecurity, biodiversity, energy and public health. To this end, this practical is designed to provide an introductory tutorial on geospatial machine learning for agriculture, particularly to classify farm-level crop types in Kenya using Sentinel-2 satellite imagery. Starting with an introductory session on key concepts of geospatial ML, the tutorial delves into ML development, validation and performance evaluation techniques. Moreover, we aim to kick-start a build-up of an African GeoAI community that strives and collaborates to solve related challenges in the continent.
Aims/Learning Objectives:
- Learn the basics of Geospatial: differentiating between the primary data types, such as vector and raster primitives. This knowledge will form the foundation for any geospatial analysis or modeling task.
- Comprehensive Knowledge on Geospatial Machine Learning Workflow: learn how to frame a geospatial problem, acquire and preprocess relevant data, and fit a model.
- Diversity in Modeling Approaches: By studying two distinct approaches – tabular learning with LightGBM and deep learning using a Sequence-to-One model you will appreciate the versatility of tools and techniques available in the geospatial machine learning domain, allowing to select the best approach for different types of problems
Prerequisites: Basic Machine Learning Concepts, Python Programming, Familiarity with Deep Learning, Hands-on Experience with Data Preprocessing
Link: geospatial_deep_learning.ipynb – Colaboratory (google.com)
