In our first part of Artificial intelligence, we talked about the hurdles that AI faced in its initial days and how it finally made a break through. In our part two of artificial intelligence, we talk about how the AI is slowly becoming more advanced. And no, it is not going to be the rise of machines scenario here.
The AI is not going to be some blood thirsty, shape shifting, fate changing robot that will rule the world, well that’s what we hope so. But at this moment, AI is beating us humans at the game of chess.
We talked about Deep blue in the previous article, in this post we will take a look at some of the more recent updates in the field.
The emergence of Deep learning and Big data:
Deep learning is a part of machine learning and works on algorithms that imitate the structure and functioning of the brain. These algorithms work through artificial neural networks. The AI market has grown exponentially in the past decade or so and seems like there’s no stopping it now. The rapid growth is a result of faster more advanced computers, big data and machine learning technologies.
DeepMind is one of the world’s biggest AI research. The program is a division of Alphabet, Inc. which is developing artificial intelligence for general-purposes. It is also known as Google deep mind. What DeepMind does is that it takes input from raw pixel data and learns from the experience.
In 2013 DeepMind introduced an AI that was capable of playing Atari games over the level of human players. Ok, so the Deep blue had already beat a human chess champion in his own game, what’s new in this? Basically, DeepMind had used reinforcement learning and neural networks, the AI had to self-learn all the moves. Now that’s a breakthrough isn’t it? Well, we’d say that’s just the beginning.
The same company in 2015 launched an updated version of the program and it was successful in playing almost 50 Atari games on its own.
The DeepMind researchers managed to go a step ahead with AlphaGo. Alpha go challenged the 9-Dan Go champion Lee sedol and won 4-1 against him.
The AplhaGo is the real breakthrough in AI, in 2017 the researches updated the program and released the next generation AlphaGo named AlphaGo Zero. With just three days of training the AlphaGo Zero was much more efficient than its previous version that beat Lee Sedol with a mammoth score of 100-0. It also beat its former version with 40 days training.
So where are we heading in the coming years?
At the moment it is really difficult to predict anything for sure. At the moment we don’t have Artificial general intelligence as such, though there is work in progress in the same. For now humans are and will stay dominant on this planet. Let’s hope there’s no rise of the machines scenario in the future too.