Google’s AlphaGo Zero: An AI Breakthrough

We’ve all heard about artificial intelligence for years now. It has dominated tech talks of the future and raised a lot of questions as well about what’s waiting for us in the coming years. Various companies have taken initiatives of realizing this big AI dream most people have and Google is not one to get left behind. Knowing that the company is a key player in the tech industry, it makes perfect sense why they are also interested in revolutionizing the tech world through AI. They showcased the updated AlphaGo Zero to show to the world that technology can work on its own even without human manipulation.

The AlphaGo Zero managed to beat its human opponent who happened to be the best Go player in the world there is now despite thousands of years of passed-on knowledge about this game after the machine was updated. It managed to rediscover patterns usually used by humans and come up with unique moves all on its own that resulted to its success. This achievement just goes to show that AI advancements are headed in the right direction and showing results faster than anticipated in the past.

Unlike its predecessors which trained on data from thousands of human games before practising by playing against itself, AlphaGo Zero did not learn from humans, or by playing against them, according to researchers at DeepMind, the Google-owned British artificial intelligence (AI) company developing the system.

“All previous versions of AlphaGo … were told: ‘Well, in this position the human expert played this particular move, and in this other position the human expert played here’,” Silver said in a video explaining the advance.

AlphaGo Zero skipped this step.

Instead, it was programmed to respond to reward — a positive point for a win versus a negative point for a loss.

Starting with just the rules of Go and no instructions, the system learnt the game, devised strategy and improved as it competed against itself — starting with “completely random play” to figure out how the reward is earned. This is a trial-and-error process known as “reinforcement learning”.


It is actually made up of a single machine that mimics how the brains neural network works or simply put, how humans think and process information before coming up with an answer. While many are excited about these new AI developments, many are also afraid of what else it can do that is beyond human comprehension. It is a reason for worry to some especially that AI machines are designed to function on its own and without human intervention whatsoever.

AlphaGo Zero doesn’t use any human data whatsoever, he said, but learns completely from self-play.

By taking this approach, it is able to always compete with a perfectly matched opponent, maximizing learning, creating much more effective algorithms.

“People tend to assume that machine learning is all about big data and massive amounts of computation. But actually what we saw is that algorithms matter much more than either data or computability,” said Silver.

AlphaGo Zero’s self-learned algorithms are so good that it uses over 10 times fewer computations than the previous version of AlphaGo.


It is now possible for machines to think for themselves without relying at all on prior human knowledge. Beating the best Go player worldwide is no easy feat and as such has catapulted the AlphaGo Zero to international stardom. Yet despite the worries of many, we should not yet fret about AI in general as the technology itself is quite limited even if we may think otherwise. These machines may be programmed to excel on a certain aspect but it has its limitations and they don’t really appear to be too smart to not need human intervention at all.

Little by little, we are trying to realize that AI dream and only time can tell what else is in store for us in the years to come. By then, we may no longer be constrained by common problems we encounter today like data loss and data recovery and deal with the often pricey or  that goes along with it. It’ll definitely have its pros and cons, so we better be ready for that.

Leave a Reply

Your email address will not be published. Required fields are marked *