UNIL
le savoir vivant
You are here: UNIL > Programmes > Card-index course
Français | English   Print   

Card-index course

Hands-on introduction to machine learning for biomedical data - 32a/Série 3

Responsible Faculty: Doctoral school (FBM-DOCT)

Teacher(s): Jonas Richiardi

Validity: 2020 ->

Course Timetable (Aperiodic)

Date Location Notice Topics Lecturer(s)
02.03.2021 de 13:00 à 16:00 POL 204.2, Amphipôle, UNIL-Sorge Python intro: variables, data types, flow control, functions, classes, modules; Scientific packages: scipy, numpy, pandas, matplotlib; Python data science: environments, loading & saving data (tabular data, images), basic plots. Open dataset: Fisher Iris. Jonas Richiardi
09.03.2021 de 13:00 à 16:00 POL 204.2, Amphipôle, UNIL-Sorge Decision theory: supervised & unsupervised learning, classifiers, dimensionality; Binary classification: logistic regression, regularization. Intro to scikit-learn. Algorithm of the week: Elastic Net. Open dataset: Allen brain gene expresssion. Jonas Richiardi
16.03.2021 de 13:00 à 16:00 POL 204.2, Amphipôle, UNIL-Sorge Preprocessing; Feature selection: filters and wrappers; Model evaluation: cross-validation, metrics; Algorithm of the week: support vector machine (SVM). Open dataset of the week: Haxby functional MRI brain activation. Jonas Richiardi
23.03.2021 de 13:00 à 16:00 POL 204.2, Amphipôle, UNIL-Sorge Hyperparameter optimization: tuning SVMs and Elastic Nets, Bias/variance trade-off; Ensembling; Algorithms of the week: random forest, extreme gradient boosting. Open dataset of the week: Cleveland heart disease. Jonas Richiardi
30.03.2021 de 13:00 à 16:00 POL 204.2, Amphipôle, UNIL-Sorge Clustering: distance metrics, dendrograms, k-means, quality indices; Deep learning overview; Dataset of the week: BYOD (bring your own data).  

Course

Annual
Aperiodic
Teaching language(s): English
Public: Yes
Credits: 1.50

Objective

Machine learning approaches are gaining importance in all fields of medicine and biology. They offer multivariate predictive modelling tools focusing on individual subjects rather than group-level inference. The main goal of this tutorial is to equip students with enough knowledge and practical experience to start applying machine learning techniques to their own research project.

Content

We will first offer a brief hands-on introduction to Python programming, then we will cover basic decision theory, binary and multi-class problems, linear and non-linear discriminative models (in particular Elastic Net regularization, linear support vector machine, random forest, and extreme gradient boosting), feature selection, model evaluation. We will also briefly introduce unsupervised learning and deep learning approaches. In each session we will implement the discussed algorithms using standard libraries such as scikit-learn, and use open biomedical data from biology, genomics, medical imaging, and clinical domains. The last session will be a hackaton-style BYOD (Bring Your Own Data) event.

Evaluation

Personal work : Yes
Presentation : Yes
Final Test : No
Participation assessed by the tutor : Yes

Bibliography

1) Hastie et al, The Elements of Statistical Learning, 2nd ed, Springer (available as free PDF)

2) Duda et al., Pattern Recognition, 2nd ed, Wiley

3) Bishop, Pattern Recognition and Machine Learning, Springer (available as free PDF)

 

Programme requirements

Basic linear algebra and calculus; Basic statistical literacy; Some programming experience (any language); Personal laptop (any OS)

Grant requirements

The course will be assessed on the basis of active participation.

Access requirements

Registration: via the Doctoral School. Period 3

Use contextFaculty codeStatusCredits
Doctorate in Life Sciences (2003 ->) ›› "3e cycle" Courses of the Doctoral SchoolED-FBMOptional1.50
Doctorate in Life Sciences - Ecology and Evolution (2007 ->) ›› "3e cycle" Courses of the Doctoral SchoolED-FBMOptional1.50
Doctorate in Life Sciences - Cancer and Immunology (2008 ->) ›› "3e cycle" Courses of the Doctoral SchoolED-FBMOptional1.50
Doctorate in Life Sciences - Cardiovascular and Metabolism (2005 ->) ›› "3e cycle" Courses of the Doctoral SchoolED-FBMOptional1.50
Doctorate in Life Sciences - Integrated Experimental and Computational Biology (2010 ->) ›› "3e cycle" Courses of the Doctoral SchoolED-FBMOptional1.50
Doctorate in Life Sciences - Microbial Sciences (2010 ->) ›› "3e cycle" Courses of the Doctoral SchoolED-FBMOptional1.50
Doctorate in Life Sciences - Quantitative Biology (2018 ->) ›› "3e cycle" Courses of the Doctoral SchoolED-FBMOptional1.50
Doctorate in Medecine and Sciences (MD-PhD) (2010 ->) ›› "3e cycle" Courses of the Doctoral SchoolED-FBMOptional1.50
Canton de Vaud
Swiss University
Unicentre  -  CH-1015 Lausanne  -  Suisse  -  Tél. +41 21 692 11 11  -  Fax  +41 21 692 26 15