Day 14252 - Answers define the rules

/

As long time fan of everything having to do with Artificial Intelligence and as a student of Machine Learning, I’ve found the general concept compared to traditional programming to be surprisingly simple.

In Traditional Programming you:

  1. define and write the rules for the data

  2. send in the data

  3. expected answers come out

In Machine Learning you:

  1. define answers for the data that we expect

  2. send in the sample data

  3. the program learns the rules for understanding how to detect the answers in the data

Or to understand it visually, there is the common primer:

traditional vs ml primer

traditional vs ml primer

Thanks to today’s easy access to Cloud Computing and machine learning platforms, the “heavy lifting” programming work is done in writing the code which helps the program understand how to define the answers in the data so that it can learn the patterns that make up the rules.

Just like learning as we ourselves do, it’s learning by example. Once you show enough examples of what something is, we learn how to distinguish what we’re looking at. Which is the first step in real “learning”.