AlphaGo is an artificial intelligence (AI) program created by Google and got its fourth win against a master player in the final match of this year’s DeepMind Challenge which was held in Seoul this March.
It won against Lee Sedol, a nine-dan ranked professional player and one of the world’s best Go players. Sedol managed to win just one (out of a possible five) of the matches against the program, failing to take home the $1 million prize that was up for grabs. The prize was instead donated by Google to UNICEF, STEM (Science, Technology, Engineering, and Math) charities, and other Go organizations.
A complex game, even for AIs
In Go, competing players take turns strategically placing black and white stones, attempting to take control of most of the spaces on a 19×19 grid.
The strategy game was invented more than 2,500 years ago in China, and was once thought to be too complex, even for AIs.
AlphaGo
The AlphaGo AI program was developed by DeepMind, a British computer firm which was purchased by Google in 2014.
To turn it into an expert Go player, its developers used previous matches to build different patterns and styles of play into its system.
The program first displayed its capability to defeat top Go players by besting European champion Fan Hui in January. It won two consecutive games against one of the most dominant players in the world, proving it can consistently win against even the most skilled human opponents.
I think that neural networks (the technology behind AlphaGo) is really the way to go. It combines the efficiency of an adaptive network with machine learning. However the slightest amount of apprehension does develop at having an AI that intelligent. Maybe I’ve just watched Terminator too much 🙂
Thumbs up for Google and artificial intelligence. This field have great potential to improve the world as long as we don’t plug it into weapons and military technology. As someone who worked with AI I must say that it was one of the most satisfying things I have ever done and I maintain it as a personal hobby. Too bad it was only briefly covered in college, I had to study more on my own.
I also had this very briefly covered in college, like you have. But even worst was te coverage of big data analyse! They tried to teach us it using Rapidminer software in just 3 weeks. Did you covered it? Which software have you used?
Having been in Korea at the time I can tell you that this sparked a big interest in Machine Learning, the government literally dropped billions on ML and AI research as a whole. You would figure a country so involved with technology advances wouldn’t need such kind of display to invest into what is obviously meant to have investments.
Artificial intelligence’s cognitive abilities keep getting more refined.
Last year there’s one blogger who in response to Steven Hawkins claims that AI could wipe out humanity, wrote an article claiming that AI had the intelligence of a 6 year old child. He was wrong it appears.
AI is far more intelligent than some skeptics and those who fear “the rise of the machines” want to admit. Of course it will take a while for it to reach perfection but I have doubt in less than a decade AI will be more intelligent than the smartest person on earth. This Google experiment [or whatever you might call it] is proof enough of that.
This is another breakthrough that will be marked in history. Go is the last traditional game whereby humans are still superior. Now that artificial intelligence has beaten humans at it, computer scientist will need to create new games specially made to challenge computers. I really respect Google for all their innovations. I believe artificial intelligence will be mainstream within this decade. At that time, we will see many changes to our society. When smartphone first came out, it already created many ripple effects. Now, artificial intelligence will have a much bigger effect indeed.
I am excited over what this new technology will bring for us. Personally, I hope artificial intelligence can be used on autonomous cars and speed up the process. I really want to ride in those cars and I believe artificial intelligence is the key ingredient to make it happen.
For a long time it was believed that an impossible to beat A.I for this game is impossible to create, but AlphaGo is 99% there, the world champion only managing to beat it once. It’s an amazing creation for such a wide and hard to predict strategy game.
I’m curious to see if they will ever get to 100% win rate with their A.I, but I highly doubt it since there are so many strategies.
I don’t think people realize actually how big of a news this is. It’s a huge step in terms of advancing our AI research. Learning machines are both fascinating and terrifying at the same time. There’s going to be a time when they won’t need human interaction anymore and they’re going to be learning from themselves. It’s probably going to happen in our lifetime, too. I can’t wait.
It is amazing how much smarter a computer can be compared to a human. It can also be pretty scary to think about, could AIs overthrow humans? I don’t think we’ll ever get to that point, but it’s still a pretty cool thing to think about.
I know, right? I think that it was obvious that we were going to reach to a point where computers are smarter than us, but I still think that we have some abilities that computers can’t, but I still wonder how long it will take until computers defeat us completely… a really scary thought.