Machine Learning
Deep learning in biology
Trainer: Joris Roels
Goal
The goal of this two-day workshop is to get acquainted with the rapidly evolving deep learning techniques that exist for bio informatics and bio image informatics, for both predictive and explorative analysis.
Objectives
Large amounts of data and compute resources have enabled the development of high-performance machine learning models. This is particularly due to deep learning techniques. By looking at many data samples, these models can find structure in the data that is useful for predictive and explorative analysis: e.g. classification, clustering, data generation, dimensionality reduction, etc. The most popular applications within biotechnology are concerned with image segmentation, diagnostics, sequence analysis, etc.
However, deep learning models are far from straightforward to implement correctly due to the many different hyperparameter settings, optimization procedures, architecture choices, etc.
In this course, we will make use of Jupyter Notebook and PyTorch, which are both based on Python, to apply deep learning techniques on both bio informatics and bio image informatics data. We aim to work towards applications that participants would like to study.
Required skills
Basic knowledge of Python is required (and machine learning is recommended) to fluently participate in the theoretical and practical parts of the course. If you don't meet these requirements you should follow the Python introduction course. and/or the machine learning introduction course first.
A tour of machine learning: classification
Trainer: Sven Degroeve
Goal
In the current age of data, advanced analysis offered by Machine Learning research is essential for transforming this data into new knowledge and useful tools. In this two-day workshop I explain the principles underlying Machine Learning and teach you how to build accurate prediction models from data through hands on exercises, with a focus on building classification models.
Objectives
At the end of the workshop you will be able to compute and evaluate prediction models from your data using the Scikit-Learn Python library. You will also be able to understand current advances in Machine Learning such as Deep Learning.
Required skills
The workshop is written in the Python programming language. If you have no experience with Python, you should follow the Python introduction course first. No background in Machine Learning is assumed, just a keen interest.