The Unstoppable Power of Deep Learning – AlphaGo vs. Lee Sedol Case Study

The Unstoppable Power of Deep Learning – AlphaGo vs. Lee Sedol Case Study
20th Jun, 2019
2350 Views

Deep Learning is one of those technologies that is hurtling towards a very promising future. It can be called as the bleeding-edge of Artificial Intelligence. It is the epitome of machine learning and the truest manifestation of Artificial Intelligence. Until a few years back when we called AI it was just plain joke since software engineers programmed the machines with enough lines of code and made it do things taking into consideration all the possibilities of the outcome and preparing machines for it and sadly we called it Artificial Intelligence. When IBM Deep Blue beat the reigning champion Gary Kasparov in a game of chess the world hailed it as the Holy Grail of AI. But what Google AlphaGo did in 2016 to the reigning world Lee Sedol makes the IBM Deep Blue Vs. Gary Kasparov duel looks like child’s play. Read on to find out more.

Watch this Artificial Intelligence Tutorial for Beginners video

Machine that learns relentlessly to outsmart humans at their own game

This is the story of Machine Learning or Deep Learning to be more specific, showing its prowess which was until now hidden from the outside world. It is the triumph of machines over humans in an unsurpassed game of Go which is one of the oldest games known to mankind having originated in China some 2,500 years ago. It is considered to be most complex board game ever invented and it is like playing four chess games on a single board.

This epic game of Go took place between Google’s Artificial Intelligence machine named AlphaGo and the reigning world champion Lee Sedol in the South Korean capital city of Seoul at the Four Seasons hotel. Before the start of the first match everybody was betting big on Lee Sedol since he was unassailable as he was rightly the Einstein of the game of Go. But as the game progressed it seemed that Lee was up against a formidable enemy one which was very hard to fathom.

Just months back AlphaGo had defeated the European champion Fan Hui but now it seems that Lee is not playing against the same machine. So what has changed in just a couple of months? Well, the machine had a chance of playing a couple of million matches against itself in order to perfect itself and reach a level which was absolutely incredible.

The first game Lee loses much to the agony and anguish of the whole Go community. In the second game Lee seems to have sized his enemy but even that won’t save him from disgrace. The machine it seems plays moves that have never been played in the international tournaments since its outcome is never known. But the machine due to the brute power of computation under its hood is able to fathom what will happen if a certain move is played. It is all about statistics and probability and nobody does it better than a Deep Learning machine like AlphaGo.

The third game goes on similarly and Lee has to resign to the power of the machine yet again. So this is when the whole Go community goes into introspection. What went wrong with Lee? How did the machine which learnt to play the game of Go just a while ago become an immovable juggernaut? So therein lies the Holy Grail of Deep Learning. It can be easily called as a technology that is nothing like anything that mankind has ever seen before.

Humanity had over 2,500 years to perfect its game of Go while the machine had a span of a few months. But thanks to the brute computing power of AlphaGo it can go on playing, perfecting itself each time in a much lesser time span. It can play millions and millions of games with itself, find out the strong moves, the opponents weak points, make billions of calculations, apply analysis, come up with newer and newer ways of playing the same old game of Go and stun the enemy each time. Now that is the power of Deep Learning. We are currently seeing just the dawn of Deep Learning thanks to this incredible game between Google AlphaGo and Lee Seedol. What the future holds for humanity, we just have to wait and watch. Deep Learning is definitely a technology to watch for in the next few years!

If all these things excite you and you are yearning to make a career in Artificial Intelligence and Deep Learning in particular then check out the Intellipaat Deep Learning training!

 

 

Related Articles

Leave a Reply

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

Solve : *
34 ⁄ 17 =