Course Description
Healthcare is an extremely challenging field for Machine Learning, either for its vast application (disease mechanisms understanding, diagnosis, prognosis, …) or for the variety of data available (medical-imaging, EEG, networks, biomarkers, genomics, …).
The course starts with a global overview of the specific problems raised by healthcare, and how Machine Learning try to address them. Then, it will dive into the description and pre-processing of the different data available (feature extraction) hand in hand with a practical session . Finally, the last part focuses on the transition from the raw data to their use in classifiers, handling the common challenges of overfitting, dimensionality reduction, missing values.
During the practical sessions, the students will be provided different types of data (images, EEG, genomics, biomarkers) and/or different challenges (Missing values, dimensionality reduction, unbalanced classes) and could focus on the one they are interested in.
Course tools
- Python (with numpy, scikit-learn and pandas)
- Jupyter Notebooks
Prerequisites
- Python
- Basic knowledge of some Machine Learning concepts (classifiers, overfitting, curse of dimensionality, dimensionality reduction, missing values)
Lecturer
Igor Koval
PhD student in Applied Mathematics and Machine Learning
Education curriculum with a Licence in Applied Economics, a master of Engineering at Ecole Nationale des Ponts et Chaussées, master of Data Science at Ecole Polytechnique. I’ve also performed several internships, in consulting and research, including two in the USA (GeorgiaTech and Argonne National Laboratory)
Fields of interests: Machine Learning, Riemannian Manifold, Statistical Learning, Programming
Contacts: igor.koval@icm-institute.org
linkedin.com/in/igor-koval-98767360/
github.com/Symcies/