We recently drew out some super simple mind maps to introduce the “universe” of Machine learning to new ML Candidates who might get overwhelmed by the plethora of terms and concepts mentioned on the great world wide web!!
Essentially when you see a machine learning related problem statement what should be your starting points?
1) What kind of supervision does this problem require? How is the data going to be ? Supervised(Labeled)? Unsupervised(unlabeled)? Semi Supervised or is this an example of Reinforcement learning?
2) How will the system need to learn? Online( incremental) or Batch processing?
3) How will the system need to work? By comparison(Instance Based) or by creating and detecting patterns(Model Based)?
The answers to these three questions will decide what you next steps are and everything you do next will depend on these answers! So be sure you get these answers right !!
This very vast universe is essentially so small
There are many many terms everywhere ( Deep learning, Non linear Regression,SVMs, Decision trees, Tensor-flow,Keras….) used all over the web but they are all just techniques to resolve problem statements that fall into one of these buckets !
Will put up some more mind maps over the next few weeks to show how all these techniques fit into these buckets making things look more complicated