GRADE Initiative "Statistical Learning & Machine Learning Applications"

Focus and general goals

Statistical learning and its application in the computer science field (machine learning) is of growing importance for all life sciences in general and for neuroscience in particular. Size and complexity of datasets has increased dramatically and the ability of a scientist to parse their data and extract useful features depends more and more on their understanding of  statistical learning methods. The aim of these methods is to approach large amounts of data in supervised or unsupervised manners and estimate underlying rules and connections to ultimately make testable predictions. Unfortunately, advanced statistical methods are only rarely covered in undergraduate and graduate courses and if they are, they do not cover the field in the breadth necessary and appropriate for a topic of such complexity and depth.

In our Initiative, we want to gain a deep understanding of the most important algorithms of the statistical learning field and their corresponding applications in machine learning. We will thus cover three thematic main points:

1) Supervised and unsupervised statistical learning methods. We will mainly focus on the mathematical ideas behind the concepts which we will analyze together.

2) Implementation of the algorithms in the programming language python. As the most commonly used programming language in the field of machine learning, we will implement our knowledge of the statistical learning methods as freely available functions.

3) Application on real datasets. Whenever possible, we will elucidate how we can apply the newly learned functions to our own data in order to get an understanding of when and how these methods are applicable and useful.

Altogether, we will approach the wide field of statistical learning on the level of mathematical underpinnings, algorithmic implementation, and integration into our own work.

Meeting and activities

The members of the initiative have regular meetings every third Monday at 7 pm. All meetings focus on a particular kind of statistical learning method that is read up on by all members beforehand. We work through the Elements of Statistical Learning and the Introduction to Statistical Learning. Both books are freely available online. The syllabus for the next meetings can be found here. Normally, meetings take place in one of the seminar rooms at the MPI for Brain Research.



Contact

New members are always welcome! If you are interested, please contact Lukas Anneser.



Members

Elisabeth Abs - FB15 Biological Sciences
Lukas Anneser - FB15 Biological Sciences
Tim Herfurth - FB15 Biological Sciences
Hua-Peng Liaw
- FB15 Biological Sciences
Amber Longo - FB15 Biological Sciences
Luis Riquelme - FB15 Biological Sciences
Theodosia Woo - FB15 Biological Sciences