Machine Learning Resources

Hi,

I am looking for resources to learn more about machine learning. I have looked at the google AI space and also, I have a course on LinkedIn that I am working through.

Are there any resources out there that helped you on your ML journey? Thank you in advance :slight_smile:

3 Likes

Are you interested in learning machine learning or testing it?

Developers tester their code and we test their final product. I believe that with data scientist who create machine learning models it’s the same thing. They write & test their code with a special training-data-set. And we should do the exact same thing.

I’ve tested several ML-models so far using Probability theory & Metamorphic Testing:

Kaggle & UCI is a really nice archive to find datasets & so much more

If you want courses, Coursera & Udemy have some nice courses on it

1 Like

@kristof A foundation of knowledge first and then go from there. It would be quite cool to create something… Thanks for the resources. I’ll have a look through.

1 Like

It’s a big subject you might need to narrow it down a little! :slight_smile: some machine learning topics include:-

  • Big data
  • Algorithms
  • Assisted leaning (nerual.networks)
  • Machine vision
  • Anomaly detection
  • Fraud detection

You can then think about particularly frameworks such as Apache spark, tensors flow, torch.

Amazon have a number of services and associated learning resources.

To be honest you could be looking at an entire degree just for the foundation. I have a compsci background and I only have a vague understanding. So most the resources of will be academic ones from universities.

If you just want a light hearted, vague concept then I love code bullet A.I. LEARNS to Play Hill Climb Racing - YouTube .
I’m not sure how education it is but he covers some of the learning concepts and shows the networks as his AI learns to play games. He is also entertaining.

Testing wise none of the background really matters you just need data set or scenarios for testing. The same as you would for any testing anything.

2 Likes