Why are AI always playing Computer Games?
The advances of artificial intelligence become obvious with its ability to compete in more and more complex games: In 1996 IBM‘s Deep Blue reached super-human ability in chess, in 2007 scientists created an unbeatably checkers AI and just recently AlphaGo bea the Grandmaster in Go, a game that is unlikely more complex than chess and checkers. However these games still are relatively simple compared to many real world challenges. Thus the next challenge would be to develop an AI that can compete in even more complex environments.
The Freeciv Challenge
Freeciv is an online game strategy game that comes very close to many real world problems. Inspired by the proprietary Sid Meier’s Civilization series, it has more possible game states than chess or Go and there exists incomplete information, which means that the player cannot observe the progress of their enemies.
Freeciv is challenging even for bright humans. eAt Arago we are taking on the challenge of mastering the Game of Freeciv with our artificial intelligence HIRO.
Because in Freeciv the player needs to build up a strong civilization and conquer the rest of the world or dominate it with technology, this turn-based strategy game where players need to consider several hundred choices for each round (choices can relate to building and developing cities or defending them) we feel it is the perfect choice for unveiling a brand new powerful AI into the world.
HIRO’s strategy against Freeciv
The common approach of many AI systems is to take into account all possible combinations of a game and evaluate their ”desirability“. HIRO was built on top of a knowledge base of some of the best FreeCiv players in the world in order to play the game as a human would. In the past, AI teams would use a brute force method in order to teach a computer an action based off a pattern — imagine a piano player who practices past the point of skill, where it becomes habit. Now, much larger investments and resources have advanced machines to deduce the reasoning behind what to do, or the “action.” This has led to great advancements for larger players such as Google’s DeepMind, as well as innovative and nimble AI’s like our own HIRO.
What it means to win
Winning against a human FreeCiv player is a major milestone because it will show that unlike outdated experts systems, HIRO can reuse its knowledge and can be easily trained by many different people at the same time. A common misconception with regards to AI can stem from popular culture and fantasy…that we are looking for cognisance or a “singularity”. In fact, we are looking for machines with memory and adaptation for long term planning. These properties give HIRO the capabilities to excel in very complex situations such as IT system management and also in Freeciv.
Through the FreeCiv challenge Arago is unveiling this Dec, we’ve already designed an AI player that is stronger than a human, while being bound by the same limitations. It’s a massive leap for artificial intelligence since it was accomplished in much less time and with much less computing power and resources than DeepMind’s team has applied to Go.
I’d like to talk with each of you about this so that I understand for pitching and further editing before we rewrite this.
Maybe we add something about the announcement at BlizzCon?