Why Robots Can’t Learn Exactly Like Babies Do… Yet
Arguably the biggest challenge in education is how to determine the best way to teach someone so that they will not only learn the thing you are trying to teach, but they will retain it and use it. That seems like it should be the first lesson on the first day of Education School, but it isn’t. And there is a very good reason why not: no one knows exactly how to do it.
There are several schools of thought on the subject, pardon the pun, but no one philosophy on education covers every child in every situation. But if we could teach a robot to learn as a human does, we could fundamentally perfect the education system.
One research team that consisted of Deb Roy and Rupal Patel set out with exactly that goal in mind. Roy dedicated his life to working in robotics and Patel was a Speech Language Pathologist, so the thought was that between the two of them, finding the exact combination of robotics and child development should be achievable.
Roy had previously tried to build a robot that acquired and used knowledge as a human baby would, but his designs always fell a little short. The robotics weren’t the challenge; it was the algorithm.
Roy and Patel set out to discover the perfect algorithm. The first teacher a child has is its mother, so that’s who they observed. They spent countless hours recording, watching and rewatching the tiny moments that pass between a child and its mother in the first months of its life, hoping to determine how exactly the learning process develops.
What they discovered was that babies are incredible learning machines.
A baby is fascinated by the stimuli from its environment even before it is born. As soon as the senses develop, one after another, a baby is figuring out how he can use them to explore his world.
While a robot, like a computer, can only pay attention to the details about his world that he is programmed to look for, a baby can take in everything at once. Babies use all of that information to constantly form and reform their views on the world all while maintaining focus on the most important thing in their entire world- the people they love.
Another difficult part of a baby’s design is the innate need for social contact. While a robot can be trained to react to words and patterns of speech, seeking out social interaction for its own sake it difficult to program. When you tack on a baby’s ability to use these social situations to deepen their learning and understanding, you have something that learns much differently than your everyday robot.
But just because we have not yet succeeded in building a robot that can learn with the same unique variances that stimulate the brain of a human baby, human children of all ages can benefit from robots in many ways.
The fact that the brains of children are so unique is exactly the reason why Roybi Robot is exceeding all expectations in the classroom. Children thrive when they receive individualized lessons that are tailored to exactly what that particular child is ready to learn. Much like having a one-on-one teacher, Roybi Robot is like having an extra teacher that never forgets a child’s name, birthday, favorite song, or reading level and never gets tired of repeating the same book for the tenth time in a row.
Although science may not be any closer to cracking the secret algorithm behind how the human mind works, we do know that Roybi Robot is invaluable when it comes to bridging the gap between busy classrooms and individualized lesson plans.
Source:
Beard, Alex (2018, April 3). How Babies Learn and why robots can’t compete. Retrieved from https://www.theguardian.com/news/2018/apr/03/how-babies-learn-and-why-robots-cant-compete
