Machine Learning in Healthcare
Healthcare is an extremely challenging field for Machine Learning as it wraps all form of applications (classification of disease stage, regression of age at disease onset, clustering of diseases or patients) with incredibly large variety of data (biomarkers, imaging, time-series, genomics, graphs, text).
The course is designed to use Healthcare as an application of classic ML pipelines, with a focus on the specific problems raised by medecine. At the end of the course, students will have a strong understanding of the following problems:
- How to bring real-life issues down to ML problems (classification, regression, clustering)
- How to extract and engineer features (missing values, unbalanced classes) from various data.
- How to use ML tools (SVM, CNN, RNN, Random Forests, Logistic regression), while avoiding typical pitfalls (overfitting, curse of dimensionality).
During practical sessions, which include tumor extraction from images or Alzheimer’s Disease staging, the students will be provided different types of data and/or different challenges and could focus on the one they are interested in.
The course is particularly well-suited for students that want to get a wide overview of ML pipelines, to deal with different types of data, and, to solve exciting healthcare problems. Projects will be proposed for the third stage of the Summer School.
- Python (with numpy, scikit-learn and pandas)
- Jupyter Notebooks
- Basic knowledge of some Machine Learning concepts (classifiers, overfitting, curse of dimensionality, dimensionality reduction, missing values)
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. Find more about me on symcies.github.io
Fields of interests: Machine Learning, Riemannian Manifold, Statistical Learning, Programming