Let’s discuss a hot topic that has been going on for the past couple of decades, the Human brain versus Artificial intelligence. Human intelligence with complex behavior versus a machine that mimics the cognitive function associated with the human mind. We have AI in our day-to-day life; a few of the tasks done by it are not that exciting to be the talk of the day. Yet there are few that are the words in the street. And today I am going to let you spark with that intriguing topic. Let me take you through a journey from the beginning where it all started, the journey through the games.
It started on February 10th, 1996, when Deep Blue played against Gary Kasparov and defeated him in chess, this wasn’t the end. They made a few software updates and had a second match on May 3rd, where Deep Blue opened with a loss. But on round two, Gary Kasparov resigned from the match rather than to be dragged out by a loss.
Deep Blue made a random move as it was wasting too much time deciding, and Gary Kasparov thought this was the move that would be defeating him as AI had the capability of looking into the near future, he resigned. And then in 2011, IBM Watson triumphs in Jeopardy(an American television show game). After Deep Blue’s success, IBM created Watson.
A question-answer system desired to beat Ken Jennings in Jeopardy. Ken Jennings, a programmer who had prior knowledge of Artificial intelligence, accepted the challenge confidently since there was no way an AI could beat him. And it also went against one of the two best players Brad Rutter. Unforeseeably, Watson defeated both ken Jennings and Brad Rutter in Jeopardy, winning a whopping amount of 77,147 dollars. Watson didn’t stop with this. It went on to play with Rush D. Hold Jr. and Bill Cassidy and defeated them successfully.
Then the revolution of AI in games began with Deepmind, a company that wanted to create an AI system that doesn’t have a constraint. They wanted to move towards the definition of general AI. One of their big goals was to beat as many Atari games using a general AI model. And in 2013, Deepmind achieved victory. It beats every Atari game ever exited using a general model.
Till now, the AI was using brute force or was used against a machine. But now, Deepmind has developed a well-defined AI to defeat humans. Deepmind has made a huge contribution to this field. If Deepmind wasn’t present, this would be a topic in vague. And from then onwards the world started to travel with the AI model of Deepmind technologies.
Before this, AI used brute force that was generating as many moves as possible and predicting the near future, but now it has changed, AlphaGo, AlphaZero, and AlphaStar are different. The change has occurred, the AI started to grow, it started to learn, to mimic human intelligence. In 2016, Deepmind’s AlphaGo learned to attach value up to a few periods in the future and learned to make decisions based on those values. AlphaGo went against the European Go champion Fan Hui. After defeating him, AlphaGo played against the highest-ranking player Lee Sedol. This loss prompted him to retire, saying, “Even if I become the number one, there is an entity that cannot be defeated.” Now we have taken the major step of the revolution.
AlphaGo was the first perfect AI model to defeat a player in a real-time match using a new move that humans never witnessed in history. This was the move that led to the great controversy. This made people think about whether AI would emerge above all human beings. Deepmind wanted to continue this for other games and believed that they could develop an AI that was so generalized that it could defeat any game by learning its moves. And then in 2017, there comes AlphaZero mastering chess, go and shogi (Japanese version of chess). AlphaZero first defeated AlphaGo. This was done only by stimulating games against themselves rather than learning from the expertise game.
It mastered Chess after 9 hours of training, Shogi after 2, and Go after 34.In 2019 AlphaStar a transformed version of AlphaZero was created. It was to beat Starcraft, a real-time strategy game. AlphaStar achieved a rank in the top 0.2 percent of humans, and this was the first time that an AI topped an e-sport game. Even though AI didn’t secure the first, it was one of the greatest milestones.AI has won against humans in Atari, chess, go, shogi and Starcraft, but it has a long way to go.
We still have games like starcraft II and Dota II that are putting complex obstacles up against the AI. Complex game theory, operating with incomplete information, undertaking multi-variable and long-term planning, real-time decision-making, and making a multitude of possible decisions at every point in time is just the tip of the iceberg. The AIs’performance in both games was massive, but also highlighted some of the areas where they could be said to struggle.
Now we have come to the present, but not the end. AI is just a baby now and we have many more years to go. Even though, many great figures like Elon Musk warned us about the dangers of AI-powered robots, even as scary as the terminator. It will remain an interesting field that people will dig into forever.