Über uns

Let's meet every other week or so to work on ML problems and get really good fast! Machine Learning involves many disciplines (Calculus, Probability, Linear Algebra, Coding, Mining, Deployment, Data Vis) and some of them lend themselves well for group study. It's not about discussing AI papers.

Although I love the beauty of mathematical proofs and their intricacies... I want to focus on cementing insights with code.

That means working with data and getting results (predicitve models). The goal here is to become solid at doing and eventually deploying. (pair-coding and peer-testing might be a good option let's see)

If you're game you should have at least completed 1+ ML course, implemented some algorithms or have experience with one of the disciplines above so you can contribute to the group. (~~ beginner/intermediate skill level)

We will use Python.

Hope to see you soon,
Markus

Good Resources – all you need is Jeremy Howard (and Love):

Fast.ai (ML for coders) is by orders of magnitude!! better and more up to date than anything else. Jeremy Howard is the most practical and legit guy in the field. He was Chief Scientist at Kaggle, founded a few other startups and is a great teacher.

This one just came out in September 2018: http://www.fast.ai/2018/09/26/ml-launch/ and it is amazing. Or do the Deep Learning 2018 course for images and so on.

Refresher / Lookup on python / pandas

https://github.com/jakevdp/PythonDataScienc...
Only use as reference, not as homework.

Meh/Solala Resources

for Mobile or Ubahn boredom: Brilliant.org (math, lin alg and ML quizzes) – most paid quizzes are sh%t, but 1/3 are well done.

Bad Resources. DON'T WASTE YOUR TIME THERE:

BAD: datacamp.com ('Interactive copy paste with 100% irrelevance to any real world problem taught by people after at least one stroke')

BAD: Andrew Ng's Coursera course. Dated, pedagogically inferior and not pragmatic. You'll finish it and won't be able to do anything after...