Introduction

Jesper Dramsch is a scientist for machine learning at the European Centre for Medium-Range Weather forecasts. They have a phd in applied Machine Learning to Geoscience from Technical University of Denmark. They are a Kaggle Kernals Expert and TPU star, ranking at top 81/100k worldwide. We talked about weather forecasting, things they learned from Kaggle, how to deal with missing data and ourliers, deep learning, Keras vs Pytorch, XGBoost, their struggles as a phd student, working in the EU vs US. Follow @DalianaLiu for more updates on data science and this show.

(00:01:27) how he got into in ML 

(00:09:10) how he handled missing data 

(00:28:34) Transformers are eating the world 

(00:49:36) Hoover Loss is a fantastic metric to deal with extreme values 

(00:54:48) his experience with Kaggle competition 

(01:02:59) Kaggle tricks that helped his models perform better 

(01:08:18) PyTorch vs Keras 

(01:30:30) working in different countries and cultures 

Resources shared by Jesper:

The newsletter with missing data:

https://buttondown.email/jesper/archive/towels-have-quite-a-dry-sense-of-humor/

The paper by Gael about missing data:

https://academic.oup.com/gigascience/article/doi/10.1093/gigascience/giac013/6568998

The Huber Loss:

https://en.wikipedia.org/wiki/Huber_loss

Skill Scores:

https://en.wikipedia.org/wiki/Forecast_skill

Brier Skill in Weather:

https://www.dwd.de/EN/ourservices/seasonals_forecasts/forecast_reliability.html

CRPS Continuous Ranked Probability Score

https://datascience.stackexchange.com/questions/63919/what-is-continuous-ranked-probability-score-crps

ConvNext, Convnets for the 2020s:

https://arxiv.org/abs/2201.03545

Transformers for ensemble forecasts:

https://arxiv.org/abs/2106.13924

Books I recommend:

https://www.amazon.com/shop/jesperdramsch/list/2DYS5KVR5TX0E

Blog posts I wrote about these books:

https://dramsch.net/tags/books/

Short I made about Test-Time Augmentation

https://www.youtube.com/shorts/w4sAh9lKyls

Their links: https://dramsch.net/links

Their open PhD thesis: https://dramsch.net/phd

Newsletter: https://dramsch.net/newsletter

Twitter: https://dramsch.net/twitter

Youtube: https://dramsch.net/youtube

Linkedin: https://dramsch.net/linkedin

Kaggle: https://dramsch.net/