Machine Learning and Why my Wife does not need to find a New Hobby

In my last Blog, The coming of the Fourth Industrial Revolution and Why Google was Getting it Right, I wrote about why prior technical breakthroughs in industrialization e.g. steam to electricity, vacuum tubes to integrated circuits, process efficiency to Machine Learning and Artificial Intelligence had been sea changes not because of their technical breakthroughs but because of the societal changes that were made possible because of these giants steps in tech.

I got a lot of questions coming off this last blog, primarily “What is machine learning and why is it all the rage?”


Let me start by saying human to human interactions in the modern world are ambiguous and messy, Human brains have been conditioned throughout time to make sense of this messiness and come to in most cases, sensible conclusions.  Take the very human ability to detect sarcasm in a conversation. We can do this because we pick up on the subtle clues like: prior knowledge of the speaker, inflections, situation or subtle body cues like: eye rolling and smirking.

One would think that the ability to detect sarcasm is a learned behavior that machines would never be able to perceive.  Think again.  By the way, the above image of Gene Wilder was found by searching for facial sarcasm.

Blame your grandparents, but even when people age, your facial bone geometry remains the same.  The angles and distances between key bone structures on all us homo sapiens can be broken down into math.  Wait, math? Billions of images, subtle mathematical facial changes, when combined with speech patterns? That spells sarcasm? Sounds like a job that super-fast computational algorithms accessing billions of data points spread over a virtually unlimited farm of servers would be good at.

So I can auto-tag my selfies.  Who cares?  Well, how about an algorithm that can provide early warning detection to parents and mental health professionals that subtle changes in a person’s facial expressions combined with sentiment in their postings, on-line behavior and inflections in their speech from billions of data points derived a propensity for suicide.   What if we could detect that propensity years in advance of an event?

School counselors are hopelessly outnumbered and overwhelmed and parents are simply not mental health professionals. The saddest thing I’ve heard from a parent whose child has taken their own life was “We never saw it coming”.

Technology has always elevated mankind.  Masses of people were once relegated to manual labor.  From building the pyramids to coal mining and factory work.  A modest technical insertion from something as simple as a fulcrum to a steam engine to manufacturing automation was introduced and mankind is elevated.

One of my favorite examples is from right around the turn of the century when elevators had an attendant who pushed the buttons for you when you entered the elevator.  Well along came automatic elevators which did not require an attendant. People would enter the elevator and quickly exit because they thought it was not safe. 

According to the US Department of Labor Statistics, 22,360 people work in the US fruit canning and preserving industry. Many of these workers are hand sorting fruit based on size, color and quality.  Human kind is better than that.

Which brings me to my wife. She loves to do jigsaw puzzles.  She can sit for hours doing these things.  Being the student of behavioral and computer science I am, although I could never have the patience to do one myself, I can spend hours watching her do these puzzles. Like all good jigsaw puzzle solvers, she sorts them by color, geometry and then proceeds with edges and common colors.  She then works her way to the middle and eventually solves the puzzle, sans the illusive peace that the cat may have made off with.

But when she’s done, she lovingly admires her finished masterpiece for a short while before breaking it apart and putting it back into the box.  That’s generally when I sit in amazement thinking that a machine learning, image and pattern recognition algorithm could have done that puzzle as quickly as you could read this sentence.


But then I realize I’m totally missing the point.  That some experiences are better left for the human condition. The mental rush and warm feeling of accomplishment is what makes us human.   Technology frees us to from having to perform menial rote tasks as acts of survival.  How we choose to spend this gift of freedom is up to us.