New Book: Machine Learning with Julia

Community Teaching & Outreach
1

If you’re interested in diving into Machine Learning and playing with Pluto notebooks for various demos, check out a new Springer book:
Machine Learning with Julia: An Algorithmic Exploration by Jeremiah D. Deng

The demo notebooks can be downloaded from JuliaHub, search for public notebooks “MLwithJulia-demos”

2

I had a look on the index, very cool, thank you!

Just a few notes:

  • you wrote in the preface that Reinforcement Learning is included, but I don’t see it in the book
  • how do you deal/organize stuff that are themselves a mix of stuff ? For example, Autoencoder as dimensionality reductions, neural networks in the context of reinforcement learning, classifiers/regressors for missingness imputation.. I think the cool part of ML is indeed when you start mixing and matching the various methods…
  • for the feature importance you have also Sobol indexes, I use them in the BetaML FeatureRanker, and they seems to provide good rankings

Let me know if you are open to share in private PDFs for genuine reviews :slight_smile: :slight_smile:

3

Thanks @sylvaticus for the quick reply!
Re reinforcement learning - to meet the publisher’s deadline I left it for the next edition (I hope there is one;0)) but I was able to include a section on Q-learning as an extension or application of the competitive learning concept.
It’s a challenge to organize the entangled ML space in some order… my approach is start on basics: metrics, clustering, dimension reduction etc., then onto classification and regression. Autoencoder etc. is briefly mentioned in the DR chapter, but its full treatment is left to later chapters related to deep learning. I tried to keep it simple for ML starters, but at the end of each chapters those “mixing and matching” methods especially in research contexts were discussed (Further Readings).
Thanks for Sobol and FeatureRanker … and imputation - I’ll definitely consider them for the revised version.
Re reviews: thanks for expressing your interest - I’ll contact the publisher so that they can arrange it :slight_smile: