Futurist Thomas Frey: We’re standing on the brink of an A.I. technological revolution that will fundamentally alter the way we live, work, and relate to one another.
Several early use cases for A.I. have begun to open our eyes as to how it will be used, but none quite as strikingly as when Google’s DeepMind was used to play the Atari game – Breakout.
In this exercise, DeepMind had no domain knowledge and was given very little background information. It was simply asked to maximize the score.
As a neural network, DeepMind teaches itself, learning from past mistakes.
- After 30 minutes (100 games) – Pretty terrible, but still learning
- After 2 hours – It had mastered the game, even with the balls moving very fast
- After 4 hours – It came up with an optimal strategy, to dig a tunnel round the side of the wall, and send the ball around the back in a superhuman accurate way. (The designers of the game didn’t even know that was possible.)
Over the following weeks, DeepMind learned to play another 49 Atari 2600 video games with minimal background information, mastering everything from a martial-arts game, to boxing, and 3D car-racing games, consistently outscoring some of the best human gamers.
“If we apply A.I. to teacherbots, the new game will be to find the fastest way to teach students.”
Cracking the Code for Cognifying Education
If we apply A.I. to teacherbots, the new game will be to find the fastest way to teach students.
Over time, AI’s will learn every students interests, their proclivities, idiosyncrasies, preferred tools, personal reference points, and how to stay engaged and learning even in the face of distractions.
A.I. will know when our:
- Skills are deficient
- What’s needed to bring them up to speed
- How and when to schedule our next training
- When we’ve mastered the topic
Throughout this training curve, individual learning will begin to scale far faster than anything we’ve ever dreamed possible – 4X, 6X and perhaps even 10X faster than anything today.
Completing a four-year college degree in 1-2 months is entirely possible with this form of A.I. learning systems.
Continue reading at: http://www.futuristspeaker.com/business-trends/cracking-the-code-for-the-future-of-education/#sthash.lkRKvXtQ.dpuf