Thank you for attending my presentation or sneaking up on this page before the presentation.

Please find the mentioned resources and references below. I am a robot and robots forget things so please let me know if I missed anything by using the buttons on the side.

Too Long; Didn’t Watch: I am arguing that when building machine learning systems we can easily end up following patterns that give us the wrong results and interpretations. These can happen when choosing our data, implementing models or when testing those in a production environment. We go in detail in some of those and see the importance of bringing the science back in data science.

RESOURCES

Presentation slides.

You can find those awesome cat icons here (priorities ¯\_(ツ)_/¯)

NYT digitalised article “On landing like a cat, it is a fact“

Basics of machine learning design, from the classic Andrew Ng. course.

What is an antipattern? Check the official book website here.

Hidden technical debt of machine learning systems paper.

Talk on technical debt in machine learning. (Includes advice and better practices on testing all parts of a machine learning system and taking care of the data/feature-related dependencies)

More advanced automated testing for neural networks using Python.