Earlier this month in London, Mustafa Suleyman had some news to share. Speaking at a machine learning event held by the investment firm Playfair Capital, he opted to display it in a giant red font on the screen behind him. This was really something — Google has replaced 60 handcrafted rule-based systems with deep learning networks.
Handcrafted what? Are we talking about a TV network? And who is this Mustafa guy?
If your eyes are glazing over — or you just don’t have a clue why Suleyman’s revelation was significant — you’re not alone. For better or (almost certainly) worse, most of us don’t have any idea what’s going on under the hood of our digital toys.
Although I’ve found no study looking at how many people grasp machine learning and deep learning networks, it’s sure to be depressingly low.
Of course this is complex, abstract stuff being developed by braniacs. Suleyman co-founded his start-up — which Google bought — with a guy who was a child chess prodigy. Schools generally aren’t requiring students to take computer science classes, and that damages the average Joe’s chance to be tech literate.
Thankfully, Mario is here to help. Over the weekend, YouTube publisher SethBling released a short video in which a computer teaches itself to beat a level of Super Mario World.
It’s a wonderful explanation of machine learning. It’s visual and tied into pop culture. We all can relate to and understand the basics of Mario. Although the work of guys such as Mustafa Suleyman will have a huge effect on our lives, they’ll never be household names whose speeches will be appointment viewing. The explanations of their work — using phrases such as “DNNs,” “agents with parameters” — might as well be in Greek.
But it has become increasingly clear that we shouldn’t ignore machine learning, which makes videos like this one so valuable. Google Chairman Eric Schmidt has singled out machine learning as having huge potential to reshape our world. We have to understand machine learning to evaluate claims that artificial intelligence is a threat to humanity.
Machine learning is one of the reasons self-driving cars could become a reality. It could help save 1 million lives a year worldwide from traffic fatalities. But self-driving vehicles also would be devastating for the careers of taxi, delivery and truck drivers. Understanding technology is essential to making sure we have relevant skills in a digital economy.
Ignorance about technology is bliss, until it costs you your job.