“I don´t think it will be a close match. I believe it will be 5–0, or maybe 4–1. So the critical point for me will be to not lose one match.”
– Lee Se-Dol (Korean Go champion before his matches against Alphago)
Lee-Se Dol was able to predict the future; it was just the opposite he was envisioning. Alphago (computer Go program done by Google subsidiary Deepmind) slaughtered him in six games.
Machines beating humans in a game is nothing new. In chess the gap between machines and human is already tremendous. Best chess machines are even able to win joint teams of human and computers. What makes AlphaGo´s victory intriguing is that Go is much more complicated game than chess. The first move of Go can involve 361 positions (chess has only 81) and Go game generally lasts more turns than chess.
“Simple heuristics get most of what you need. For example, in chess and checkers the value of material dominates other pieces of knowledge — if I have a rook more than you in chess, then I am almost always winning.
Go has no dominant heuristics. From the human’s point of view, the knowledge is pattern-based, complex, and hard to program. Until Alphago, no one had been able to build an effective evaluation function.”
-Jonathan Schaeffer (Creator of Chinook, first program to beat humans in Checkers)
AlphaGo is based on deep learning and neural networks. So while Deep Blue beat Kasparov with sheer computing strength, Alphago has more artificial intelligence behind it. Firstly neural networks were trained on 30 million moves from games played by human experts. That resulted to ability to predict human move 57 percent of the time. But that gets you to the same level as human players not necessarily able to beat them. So secondly, AlphaGo played thousands of games between its neural networks, and adjusting connections using trial-and-error process through reinforcement learning.
How many humans are even able to comprehend what above means (lest train themselves in even somewhat similar manner)?
Machines can already replace humans in more fields than we are willing to admit. And more importantly, they are playing better job as well. Machines can crunch data to obtain experience, which is impossible for humans during their lifetime. We have to start embracing machine learning and collaborating with machines more if we want to survive. Advertising industry has been especially almost hostile to any technological improvement. That will be a road to sure destruction. Beating a Go champion is much harder task than to do a subpar brand campaign. If we don´t take more proactive and positive approach to data and artificial intelligence, we will make ourselves redundant.
Machines can either be our allies our friends. I would opt for the latter choice.