“Machine teaching” is going to become a sunshine job description

“Machine teaching” is going to become a sunshine job description

Those of us who have been in the IT Industry for many years, have witnessed the boom in software testing. Thousands have been employed in software solution testing. It is still significant but it has plateaued off.

Over the next decade many jobs would be created in machine teaching. Not many are talking about it yet. However, this is logical given the growth of AI, Data Sciences and Machine Learning.

For the machine to learn, someone needs to teach it. Just like in a classroom, students learn from a teacher.

In algorithms which are purely statistical, human intervention may not be needed. However, in cases where the machine learns by mimicking human behavior, teaching the machine is essential for machine learning.

Human beings are aware of universal truths (the Sun rises in the East!). That, coupled with domain knowledge, makes them context aware. (Example: A credit card transaction at a retail merchant location is suspicious if the same card has been used half hour ago, in a location that is 500 miles away!).

Machines lack the context. So, when they are learning by mimicking the human behavior, a person needs to teach the machine.

The analogy to testing is striking. While teaching the machine may be a little higher order function, it has close similarities to testing a software solution.

Instead of a “test plan,” a “teach plan” needs to be used here.

Instead of “test cases,” we need to deploy “teach cases.”

The teacher needs to be familiar with the domain based process and given a situation, what are the right actions to take and steps to follow.

The teacher will perform these actions for each “teach case.” In the background, the machine will learn from this.

Like “test coverage,” “teach coverage” needs to be measured and optimized for higher quality.

It’s quite likely that in the next decade, many post-millennials would be engaged in the activity of “machine teaching.” Those who catch the trend early are likely to gain the most.

Photo by Joanna Kosinska