There are many great courses, tutorials, and solutions, freely available online. After reading a hundred implementation of SOTA Deep Neural Network, everyone can easily notice, that they lack many essential issues. It’s like teaching to drive a car by explaining how the pedals work. Daily machine learning gives more broad problems than fitting the MNIST dataset.
- How to gather relevant data? How much data do I need?
- When is the system good enough for the production?
- What are the weaknesses of the solution?
There are many important and complex questions containing daily decisions, that are not answered online.
The website you’re reading comes to the rescue.
Machine Learning Simulators is the idea I found in Coursera – “Deep Learning: Structuring Machine Learning Projects” by Andrew Ng. It was a great discovery for me because I realized there’s no better way to gain practical knowledge, necessary to be able to work in the industry.
I’d like Machine Learning Stories to be a common space to share the ideas about real, practical ML problems, that will eventually let us be better engineers.
Let’s try the first story!