Beating humans at their own game
Artificial Intelligence (AI) has been a hot topic in pop-science for a while and there are valid reasons behind such popularity. But what's made AI more compelling to follow has been the triumphs of the computer over humans and mankind's silly games. Today, we take a look at a few instances where AI overpowered humans and pawned the noobs.
Deep Blue Beats Garry Kasparov
In 1997, IBM's Deep Blue beat Garry Kasparov, the world champion chess player. While there are speculations about the move that gave the game to Deep Blue, there have been speculations that it was either a glitch or that human intervention was involved. Regardless, by the early to mid-2000s, technology had improved to the point where machines had beat a chess grandmaster with more triumphs to follow.
AlphaGo Beats Ke Jie
In May of 2017, AlphaGo "Master" took on Ke Jie, the world's highest-ranked Go player. Over the course of three games the machine comprehensively dominated the world champion. While AlphaGo had proven to be the superior Go player on the night, a new challenger emerged courtesy of Google in October of last year named Alpha Go Zero, a more sophisticated version of AlphaGo Master.
AI Masters Poker
While Go is a complex game in itself, poker requires an entirely different set of skills which includes deception and trickery. Carnegie Mellon University's Libratus AI system spent 20 days playing 120,000 hands of No Limit Texas Hold'em Poker against four poker professionals. The human pros were no match for the machine and after nearly a month of full-time gameplay, Libratus was up by US$1.7 million, with every one of the four professionals having lost thousands of fictional dollars.
OpenAI's Bot Beats World's Best DOTA Players
In 2017, a bot developed by OpenAI beat DOTA 2 pro Danylo "Dendi" Ishutin in a game where the demonstration was highly limited to a few variables of gameplay. In 2018, two teams of pro gamers overcame a squad of AI bots that were created by the Elon Musk-founded research lab OpenAI. After that OpenAI's DOTA 2 AI crushed world champion e-sports team with back-to-back victories.
Deepmind's AI Agents Beat Human Pros at Starcraft II
AI agents developed by Google's DeepMind subsidiary have beaten human pros at StarCraft II. AI players beat the humans 10 games in a row. However, humans were able to secure one victory against the AI in the final match.
While most of AI's dominance has happened in games with fixed unit values with no chances for random encounters, it has brought forth a significant research challenge. Games like StarCraft IIare harder for computers to play than board games like chess or Go. In video games, AI agents cannot watch the movement of every piece and calculate their next move while also having to react in real time. The only limitation to AI's growth was the limitation of how much a human could teach. But with self-learning systems being implemented, theoretically there are no limitations to how much AI can learn. So, what does the future hold? Only time will tell.
Comments