Waves to Weather

Breadcrumb Navigation


ECS Machine learning workshop (Oct. 2020)

The ECS community of W2W gathered online for a three-day machine leaning workshop on 12-14 October 2020. The workshop with around 35 participants was hosted by Stephan Rasp (W2W alumni), Sebastian Lerch (W2W) and Eva-Maria Walz (W2W).

A broad spectrum of topics was covered, from well-established machine learning techniques like multi-linear regression and random forests to state-of-the-art deep learning methods like neural networks, convolutional neural networks and transfer learning. Introductory presentations to each topic were paired with hands-on coding sessions. Addressing the participant’s background, the coding exercises focused on datasets from atmospheric science. To work on the python-based exercises, the participants gathered in teams of three in online breakout rooms.
The teams not only worked closely together learning from each other, but also competed against other teams in two machine learning competitions on the platform Kaggle. Considering the relatively short workshop, the teams mastered the two competitions on weather forecast post-processing and satellite image classification with significant success.

The online format worked out really well for everyone, not least because of the tutor’s ingenious technical preparations.

In the name of all participants, the ECS committee would like to thank the three tutors Stephan, Sebastian and Eva-Maria for all their work and passion that went into this workshop.

ml_workshop_2020_smallParticipants of the machine learning workshop