Objective: hands-on experience on real problems.
The Paris-Saclay Center for Data Science will start a new series of data science training sessions. The objective of the events is to give novice data scientists (and domain scientist interested in data science) hands-on experience on real problems. The participants will learn about practical issues such as casting and formalizing a problem into a model, data normalization, diagnosing and debugging a model, model and algorithm selection, parameter tuning, and postprocessing. Another important objective of the training is to introduce a wide range of applications in various domains (astrophysics, brain imaging, economics, etc.).
Format and content
Each session will consist in a one-day bootcamp. We will provide a scientific paper or a high-level documentation (e.g., a Kaggle site or a standalone document) describing the problem together with a corresponding data set. The participants will make an attempt to reproduce the results reported in the paper with the help of the coaches. Depending on the particular theme, a crash course may be given by the organisers to introduce the background of the exercise. The participants will than tackle the problem in small teams consisting of people with different expertise.
- The participants are expected to come prepared for each session (i.e., having read the suggested papers).
- Most sessions will be based on python, using tools such as scikit-learn. Participants will be expected to have a basic notion of programming. Python novices should train themselves using online tutorials such as
Three upcoming sessions have been programmed so far:
- January 15: High-energy physics: learning to discover, based on the HiggsML Kaggle challenge.
- February 9 (rescheduled from February 2): Brain imaging / Neuroscience / Medicine, TBD, but probably on electrophysiological data and/or medical records.
- March 31: Astrophysics: TBD, but probably on galaxy classification for sky surveys.
Participants should apply by filling out the form available at
Priority will be given to researchers (PhD students, postdoctoral fellows, permanent researchers) belonging to one of the Université Paris-Saclay partner institutes. In exceptional and well-argued cases, we may also accept applications from M.Sc. or engineering students and from researchers outside the Université Paris-Saclay cluster.