Since I expect the growth of AI to occur at an exponential
rate it only makes sense that if we look away too long it will have made
incredible strides without being noticed.
Remember when deep blue made a big show about beating Kasparov? It was almost spectator sport. The whole world was aware of it and there was
a lot of coverage. The chess champ and
the computer split the series but the computer ended up winning the match. It was a big deal, widely talked about at the
time.
Chess has a lot of moves and is considered a very logical,
thinking man’s game. But the game of Go
has even more complexity and is considered an intuitive man’s game. And so without much fanfare, Google’s
DeepMind acquisition beat the European champion 5-0 as documented here: https://www.youtube.com/watch?v=SUbqykXVx0A.
As the power of AI increases at an exponential rate, the
shock and awe of its achievements will decrease at the same rate. That is why the deep blue-Kasparov match was
such a big deal for AI even though the computer could not trounce him on the
simpler game while pretty much nobody has ever heard of AlphaGo even though it
wiped the board with the European champ.
In the foreword to his novel Aldous Huxley states, “The theme of Brave New World is not the advancement of science as such; it is the advancement of science as it affects individuals.”
Brave New World, here we come.
Between Deep Blue, a brute force approach, and Deep Minds there's been a pivotal development in neural networks: folded neural networks. It's not my field, but it seems that, though their training is like of regular neural networks, they can make decisions based on non linear factors, unlike their predecessors. The results seem to mimic the human decision process more closely. I don't think that we've but the beginnings of the development and applications of this technology.
ReplyDelete