“Technology is the heartbeat that will power the future that the whole world will wake up to tomorrow”
That’s the words of my physics professor in one of his classes back then in Secondary school. These words of his vibrated into my ear so many years ago, I am still living to see a little bit of it come to fruition.
Today, the whole world is going berserk with mind-blowing technological innovations and the Artificial intelligence is just the tip of the iceberg.
What’s Artificial Intelligence?
Abbreviated to the word A.I. Artificial intelligence or machine intelligence is the name giving to the intelligence demonstrated by machines. Modeled from the neural intelligence displayed by humans and other animals, A.I machines learn by perceiving its environment and takes actions that maximize its chance of successfully achieving its goals.
Today we are going to talk about AlphaGo; it’s general purpose A.I system that was giving birth to by Google’s Deepmind division.
In 2016, Deepmind made a lot of headlines when AlphaGo; its A.I programme trashed a human professional (Lee Sedof) in the ancient Chinese board game called “GO”. The game is believed to be the world’s hardest game and the victory was heavily celebrated as the world’s biggest breakthrough in the A.I field.
Playing and winning the best human brain in the game showed that machine can truly learn and think in a human way.
Instead of using brute force to calculate all the moves the machine can make just the previous A.Is before AlphaGo was designed to do, AlphaGo uses reinforcement learning and mirror networks to mimic the learning process of the human brain.
The Ancient Chinese game has so many moves more than the well-known chess, the number of moves we are talking about is almost the number that’s in somebody’s mind when he/she is thinking about the number of atoms in the whole entire universe.
Deepmind creators stated that the A.I could learn from time and just improve on it.
It’s a British artificial intelligence company that was founded in September 2010 as deepmind technologies. It was a startup founded by Demis Hassabis alongside Mustafa Suleyman and Shane Legg in 2010. In 2014, the programme was acquired by Google for 500 million dollars.
According to their website, the goal of deepmind technologies is to solve intelligence; using it to make the world a better place. To do this, they are trying to combine the best techniques in machine learning and systems neuroscience to build powerful general-purpose learning algorithms.
In a layman’s sentence, they want to formalize intelligence, find out what it truly is and how it works not only so they could be able to implement it into machines but to understand the human brain and how it works.
The AlphaGo A.I created by deepmind uses a slightly different technique from what other A.I technologies uses, it’s called reinforcement learning. This makes AlphaGo very different and more efficient from other A.Is like IBM’s Watson, Google’s own digital assistance, Google Now or Apple’s Siri. Thee ones where developed for a pre-defined function and can only perform on a limited prowess that encircles the platform to which they were created on.
AlphaGo is not pre-programmed and it learns from experience using raw pixels as data input.
To train AlphaGo, the experts behind it had to show it hundred thousand Go games that the best players have already played, downloaded from the internet. AlphaGo initially is pushed to mimic a human player and later on it is made to play against itself 30 million times on deepmind’s servers using reinforcement learning.
While playing with itself, 30 million times, the system learns and improves itself incrementally by avoiding the errors and mistakes and improving its possible to win against older versions of itself. And the result after each section of play is that you end up with a new version of AlphaGo that can be able to compete with the old version of it and win effectively.
The reinforcement learning deepmind uses on AlphaGo is model-free, meaning that it doesn’t need a structure or set of rules to learn. The neuro-network is like that of the human brain. AlphGo learns through experience.
Another cool thing about AlphaGO is that after playing a game and getting its experience, it can be able to transfer what it learned and use it to play a different game.
Deepmind hopes to use AlphaGo to upgrade the different sectors in life that are of immense importance, like the Healthcare sector where it promises to help human experts achieve more. Deepmind has already created a healthcare app.
Other fields that deepmind is going to use their A.I on is the online customer service, computer vision, finance, general computer science and news publishing and writing.
AlphaGo is a general purpose A.I meaning that it is quite huge and vast, it can be modeled into nearly any field and used effectively there. It also has the ability to create a different field of science which lays the foundation for the creation of other imaginable things.
But wait, should we be worried?
While the idea of an A.I will seem so interesting, there has always been these question that will always disturb man whenever issues about Artificial intelligence and machines are mentioned. It is quite scary to think about what might be the outcome, you know, those Hollywood sci-fi movies where machines evolve into what man cannot control any longer and the thought that they might take over the whole world and even enslave us.
A.I technology is the only thing that man has created that is quite unpredictable. So, we should all thread this path with caution.