Introduction to AI and Machine Learning for Biomedical Research
Get familiarised with AI models and state-of-the-art biomedical research based on machine learning! VAIA and its partners are hosting their successful introductory AI training course for biomedical research already for the third time.
Practical information:
Want to register?
- Register until: 19 Nov 2024
- Price: free of charge
Certificate of attendance
Georganiseerd door:
Leertraject
Huge amounts of data are available today for biomedical research. These data arise in different forms like images, omics-data, electronic medical files, etc. Machine Learning algorithms can help researchers to find patterns, classify the data or help to make well-founded decisions, based on these data. Typical examples are the analysis of CT scans for cancer diagnoses or the identification of the proper treatment for MS patients.
Programme
Part 1: Self-paced online course KULeuvenX: ‘AI in Healthcare. Hype or Help’
- Chapters 1, 2 and 4 are mandatory. Chapter 9 is recommended.
Part 2: Live online classes
- Module 1: Supervised Learning
Axel-Jan Rousseau (UHasselt, VAIA) & Christos Chatzichristos (KU Leuven, VAIA)
26 November, 13.00-15.00 - Module 2: Unsupervised Learning
Celine Vens (KU Leuven)
28 November, 13.00-15.00 - Module 3: Deep Learning and Neural Networks
3 December, 13.00-15.00
Matthias De Lange (Superlinear) - Module 4: Reinforcement Learning
6 December, 13.00-15.00
Pieter Libin (VUB AI Lab) - Module 5: Fairness and Privacy of AI
10 December, 13.00-15.00
Maarten Buyl (UGent)
Part 3: Presentation of use cases (online)
- Module 6: Use cases from the biomedical sector
12 December, 13.00-15.00
Learning objectives
The participant
- has an overview of the terminology and algorithms of AI that are applied in current research.
- is aware of the possibilities as well as the challenges concerning AI techniques (bias, fairness, data management, ethics, etc.).
- is able to discuss problems related to their research with an AI expert using appropriate terminology.
- is able to judge if AI techniques might be interesting to apply in their research.
- is familiar with some contemporary research topics in the biomedical field that use AI-related techniques.
Part 1: Self-paced online course
Online course: KULeuvenX: ‘AI in Healthcare. Hype or Help’
In this online course, developed by KU Leuven with the support of VAIA, you will discover how Artificial Intelligence (AI) is revolutionizing healthcare. The added value of AI for healthcare is explained from the healthcare professional's perspective, while basic AI principles are clarified from the AI developer's point of view.
The following chapters are mandatory. This means that you are
expected
to have completed those three chapters before the first online class of
Part 2 (Module 1: Supervised Learning). Are you interested in learning
more? Feel free to complete the other chapters as well!
- Chapter 1: Breaking the (A)Ice (estimated time to complete: 1hr)
- Chapter 2: First AId Kit (estimated time to complete: 2hrs)
- Chapter 4: Inside the AI engine: Learning from data (estimated time to complete: 3-4hrs with exercises)
Not mandatory but recommended before attending Module 5 'Fairness and Privacy of AI':
- Chapter 9: How to coexist with AI (estimated time to complete: 2hrs)
The online course be freely accessed through edX platform, but you will be asked to create an account.
Part 2: Live online classes
Module 1: Supervised Learning
26 November, 13.00-15.00
Axel-Jan Rousseau (UHasselt, VAIA) & Christos Chatzichristos (KU Leuven, VAIA)
We'll focus into demystifying the core concepts of supervised learning. Discover how these learning paradigms underpin many AI-driven solutions in healthcare, enabling us to make predictions, classify patients, and uncover hidden patterns in medical data. In supervised learning, we make use of patterns and relations in labelled data to learn classification and regression models. We will delve into key concepts and discuss several algorithms such as SVM, KNN, and decision trees, as well as how we can measure and improve the performance of these algorithms.
Module 2: Unsupervised Learning
28 November, 13.00-15.00
Celine Vens (KU Leuven)
In
unsupervised learning, we aim to analyse and learn from unlabeled
datasets. The algorithms discover hidden patterns or data groups without
the need for human supervision. In this module, we will discuss several
tasks and techniques of unsupervised learning, including clustering,
association rule mining, anomaly detection and recommender systems. We
will also touch upon semi-supervised learning. For each task, we will
discuss some examples from the biomedical domain.
Module 3: Deep Learning and Neural Networks
3 December, 13.00-15.00
Matthias De Lange (Superlinear)
In this module, we will elaborate on neural networks and deep
learning, with a specific focus on the biomedical research community. We
will delve into the intricate workings of neural networks, various
derived architectures (CNNs, RNNs, Transformers, etc.), and more
advanced generative AI algorithms. Furthermore, we will elucidate how
these technologies are reshaping the landscape of biomedical research.
From screening and diagnosing in medical imaging to drug discovery and
protein folding, deep learning is revolutionizing the way we approach
complex problems in biomedical research.
Module 4: Reinforcement Learning
6 December, 13.00-15.00
Pieter Libin (VUB AI Lab)
Reinforcement learning concerns a distinct branch in the field of machine learning. While supervised and unsupervised machine learning techniques learn based on a set of instances, reinforcement learning concerns an agent that learns by interacting with an environment. To illustrate this setting, consider a video game, where the agent identifies as the player of the game. The environment concerns the video game that can be controlled with a gamepad and observed by looking at the screen. The agent receives feedback from the environment in terms of rewards, for example via a game score. Based on this feedback and the observations of the environment, a reinforcement learning agent aims at learning to play the game optimally, by interacting repeatedly with the environment. In this course, we will study both the reinforcement learning problem and the algorithms that can be used to approach such problems.
Over the last years, several important milestones were achieved using the reinforcement learning framework, such as learning to play ATARI 2600 video games, learning to play the board game Go to beat the number one human player and learning to reach grandmaster performance in the real-time strategy game StarCraft II. However, reinforcement learning is no longer limited to games, as many real-world problems can be formulated as a reinforcement learning problem. Recent studies used reinforcement learning to solve challenging problems such as wind farm control, epidemic mitigation, and robot control.
Module 5: Fairness and Privacy of AI
10 December, 13.00-15.00
Maarten Buyl (UGent)
Through machine learning, AI systems can learn and reproduce patterns of human intuition in datasets on a large scale. However, human behaviour in high-impact decision-making has always come with ethical codes and expectations, especially in biomedical sciences. Now that AI systems play an increasingly substantial role in such decisions, they should also aim to meet ethical standards. In this module, we particularly consider the ethical principles of fairness and privacy in AI systems, both highly active fields of research. Fairness aims to avoid statistical patterns of discrimination in decision-making, such as performing far worse for certain demographic groups. Privacy aims to avoid revealing sensitive information, such as people's medical records. We will discuss synergies between these two principles, but also their conflicts.
Part 3: Presentation of use cases (online)
Module 6: Use cases from the (bio)medical sector
12 December, 13.00-15.00
- Pieter Libin (VUB AI Lab): 'Deep Reinforcement Learning for Epidemic Policy Control'
- Alexander Koch (BioLizard): 'Designing anti-microbial peptides'
- Joseph Miano (Superlinear): 'Anonymizing Personally identifiable Information in textual medical data: from classical NLP to LLMs'
- Melvin Geubbelmans (UHasselt): 'AI to extract cell locations and types from digital pathology slides and use of spatial statistics to determine tumour heterogeneity in lung cancer tissue'
- Bert Hoorne (Microsoft): 'From Vision to Implementation: Safe and Compliant Generative AI for Patient Care and Clinical Experiences'
Teachers / speakers
Axel-Jan Rousseau
Dr. Axel-Jan Rousseau is an AI Expert for VAIA and a Postdoctoral Researcher at Hasselt University. Specializing in trustworthy computer vision, his work focuses on model calibration, ensuring AI systems provide reliable, statistically sound predictions.
He is also one of the coordinators for the Data Science track for FLAMES and develops AI-driven decision support systems for biomedical imaging and MS diagnosis. He holds a PhD in AI (2024) and Master's degrees in both AI and Industrial Engineering.
Christos Chatzichristos
I am Christos Chatzichristos, currently a post-doctoral researcher at KU Leuven. My educational background revolves around electrical and computer engineering, with a specialization in Biomedical Applications and an emphasis on signal processing during both my Master's and Ph.D. studies. During my doctoral journey, I witnessed the profound impact of neural networks on the field of signal processing, marking the inception of my foray into the realm of AI applications. I hold a strong belief in fostering broad interdisciplinary collaborations, as I believe that research today cannot thrive in isolation within a single domain. Artificial intelligence stands as a potent tool to expedite healthcare research. However, to truly harness its potential, we must bridge the gap by facilitating healthcare professionals' understanding of fundamental AI concepts, just as they aid biomedical engineers in unraveling the mysteries of the human body. So, here's to using AI to accelerate healthcare research while ensuring that we all speak the same language – whether it's the language of algorithms or the language of anatomy!
Celine Vens
Celine Vens is professor at KU Leuven campus Kulak and leading the machine learning and AI subgroup at itec. Her main research interest is the development of machine learning algorithms for applications in health and education. She focuses on non-standard supervised and semi-supervised learning tasks, such as multi-output prediction, time-to-event prediction and interaction prediction. Besides research, she is teaching several courses at the Faculty of Medicine.
Matthias De Lange
I'm a Senior AI Researcher at TechWolf, a leading AI-product scale-up known for unlocking 'skills' as a crucial data source for companies. My focus is to drive the state-of-the-art in explainable and unbiased AI in extreme multilabel classification.
Pieter Libin
Prof. Pieter Libin graduated in 2014 at Vrije Universiteit Brussel in Informatics and obtained his PhD in Computer Science at VUB in 2020. After his PhD, he held a postdoctoral position funded by the Flemish science foundation at Hasselt university, at the department of data science. Since October 2021, Pieter is active as an assistant professor at the AI lab of the Vrije Universiteit Brussel. His research involves the use of machine learning to support decision makers by combining machine learning techniques with realistic simulation models and concerns theoretical and applicational AI with a focus on reinforcement learning and Bayesian modeling. He has a broad experience in modeling and analyzing real-world systems ranging from virus diversity, epidemic emergencies, and renewable energy providers. Pieter is a member of the Jonge Academie, an association of young top researchers and artists with an engagement to policy, society, research, and the arts. Pieter is a board member of the Benelux Association for Artificial Intelligence.
Maarten Buyl
Postdoctoral researcher at Ghent University
Alexander Koch
I'm Alexander and I'm a data science team lead at BioLizard. Our goal is to transform the biomedical data of our clients into actionable insights. Using data mining, data visualisation, bioinformatics pipeline development, software development, and the construction of ML and AI models, we support biotech and pharmaceutical companies and research institutes in their (pre)clinical research and development efforts.
Melvin Geubbelmans
PhD researcher at UHasselt
Expertise: Biostatistics, Data visualisation and high-throughput image analysis
Bert Hoorne
Trying to transform healthcare with the help of people and AI
Bert Hoorne is Senior Technical Program Manager (Health AI) at Microsoft.
Related courses
Fundamental concepts of multi-omics data integration
Opleiding - Leuven - Genomics Core Leuven
Transforming Healthcare with AI
Two-day program - Cambridge - MIT Sloan School, MIT Jameel Clinic