A Product Manager's 600-Word Guide to Machine Learning

Machine learning (ML) is a technology or field of computer science that learns from historical data to make accurate predictions or decisions.

This could be looking at the purchasing patterns of a buyer and recommending products to a similar buyer. Or, this could be analyzing weather data to predict what the weather would be like on a particular day.

There is no magic here. It’s just a wonderful combination of statistics and mathematics.

To make things easier, imagine you’re at a coffee shop with your closest friend. The barista asks your friend what kind of coffee they want. Chances are pretty high you can guess your friend’s answer with high accuracy.

That’s not because you’re a magician. It’s because you know your friend. You have seen, over the course of your friendship, what your friend typically orders. You have the (mental) data to make that prediction. You also know how your friend’s mood, coffee shop, the time of the day, and more, can affect the coffee decision.

A machine learning model is no different. It analyzes huge volumes of data, or in other words, learns all the nuances of the dataset you feed to it, to the extent that it can make correlations between different data points.

One more example. Person A and Person B have similar purchase histories. If A purchases an item C, then it’s likely that B will also be interested in C. This is a simple representation of recommendation systems, which is a widely used application of machine learning.

Machine learning isn’t a fad.

Chances are high that you benefit from the applications of machine learning every day.

Chances are even higher that a machine learning algorithm was responsible for recommending this article to you or bringing it up on your feed.

Machine learning isn’t a complex topic. It could be complex for ML engineers who build ML applications, but you as a Product Manager (or a non-technical person) don’t have to be overwhelmed by this subset of artificial intelligence.

But, have we achieved artificial intelligence? No.

So are they the same thing? Let’s dig deeper.

Long story short, yes, artificial intelligence is just a fancy phrase used to describe machine learning - at least for now, until we create an artificially intelligent agent.

“People are getting confused about the meaning of AI in discussions of technology trends—that there is some kind of intelligent thought in computers that is responsible for the progress and which is competing with humans," says Michael I. Jordan, a leading researcher in machine learning. “We don't have that, but people are talking as if we do."

Probably, it has to do with marketers who felt machine learning wasn’t interesting enough. And I agree with them.

The phrase “machine learning” gives me the mental picture of some boring computer, learning to do something. Whereas, “artificial intelligence” inspires me. It gives me the mental image of a futuristic robot making intelligent conversations with me and telling things no other human has ever told me.

Machine learning can be broadly divided into three categories: supervised, unsupervised, and reinforcement learning.

Supervised learning is a machine learning method in which labeled data is used to train algorithms.

Unsupervised learning is another method in which there is no labeled data and the algorithm has to learn by identifying patterns.

Finally, reinforcement learning is a machine learning training technique in which the learning agent is rewarded for desired behaviors and punished for undesired ones.

Some applications of machine learning include:

That’s 600ish words and a quick guide to machine learning technology. :D