Prepping the next generation for A.I.
Today, my colleague Steve Lescault and I wrapped up our second of two classes to sixth graders at The Study school for girls. This two hour, interactive course is designed to teach the students how to recognize A.I. in applications; differentiate A.I. from other kinds of programming; appreciate how bias enters A.I.; and learn basic A.I. concepts like: feature sets, positive and negative samples selection and confidence scores.
I am genuinely impressed by the aptitude of these students; their grasp of these abstract concepts and their ability to reason and think. This is hard stuff — I don’t know many adults, some even in my field who can accurately differentiate between artificial intelligence and regular old programming. The elastic minds of these 6th graders picked it up right away.
For me the most rewarding part of the class is watching as students use the Intellogo cognitive platform to teach it concepts of their choosing. It is hopeful seeing the students struggle with the important stuff: Should they reward Intellogo or not for returning a news story about Chipotle when they’re trying to train it to recognize Starbucks? Well guess what? There is no right or wrong answer. How they train the system will necessarily introduce their bias into it. The point is to get them to realize this. In the end, they decide to include the story because it describes how Chipotle could learn a thing or two from Starbucks.
Another team debates whether they should reinforce Intellogo returning results for ice cream even though they were trying to train the system for milkshakes. A lively debate ensues with some students advocating that without ice cream there wouldn’t be a milkshake while others are adamant that milkshakes and ice cream are two different things. Again, there is no right or wrong answer — the debate is the real lesson.
By the end of the class, the students have greatly improved the accuracy of their concepts. And, they have a much more nuanced understanding of how systems like Google’s auto-complete and Siri’s natural language processing work. As a surprise, I let the students know that their concepts of milkshakes, cherries, Starbucks and sushi would be made live to the thousands of users on Intellogo.
All in all, it’s a good day at work.
If you like this post, don’t forget to recommend and share it. Check out more great articles at Code Like A Girl.