Crossing The Chasm - Machine Learning Predictions to Business Decisions

 

At the outset, let me state unequivocally that 'I love Machine Learning' and its associated pipeline of taking raw data all the way to create reliable predictions. There are already fantastic applications / use cases of ML usage in enterprises and this is bound to grow. Having said that, there are still considerable gaps (chasm, if you will) in applying ML predictions for business decision making and this post is an attempt to articulate some of those gaps and its possible solutions.

Here is a picture that illustrates the top five gaps:

Here are a bunch of resources to get started on these topics:

1) Machine Learning Interpretability (MLI) 

2) Probabilistic Programming

3) Causal Machine Learning

4) Optimization

5) Reinforcement Learning & Simulations

And here is a video that talks about the top 3 limitations of Artificial Intelligence.

Check your understanding by answering this question