Finally, AI bots pass the Turing Test
Right on the 100th birth anniversary of artificial intelligence (AI) "father" Alan Turing, an AI gamer has won the BotPrize after convincing judges it was human-like.
The AI created by computer scientists at The University of Texas at Austin was even deemed more human-like than half the humans it competed against, Eurekalert.org reported.
"The idea is to evaluate how we can make game bots, which are nonplayer characters (NPCs) controlled by AI algorithms, appear as human as possible," said Risto Miikkulainen, a computer science professor in the College of Natural Sciences.
Miikkulainen created the bot, dubbed the UT^2 game bot, with doctoral students Jacob Schrum and Igor Karpov.
UT^2, which won a warm-up competition last month, shared the honors with MirrorBot, which was programmed by Romanian computer scientist Mihai Polceanu.
The winners were announced this month at the IEEE Conference on Computational Intelligence and Games.
Both winning bots both achieved a humanness rating of 52 percent, while human players got an average humanness rating of only 40 percent.
The two winning teams will split the $7,000 first prize, Eurekalert.org said.
UT^2 took part in a competition sponsored by 2K Games, set inside the virtual world of "Unreal Tournament 2004," a first-person shooter video game.
In the tournament, the bots face off against one another and about an equal number of humans, with each player trying to score points by eliminating its opponents.
Each player also has a "judging gun" that can tag opponents as human or bot. The bot scored as most human-like by the human judges is named the winner.
The complex gameplay and 3-D environments in the game require bots to mimic humans, such as moving around in 3-D space, engaging in combat against multiple opponents and reasoning out the best strategy.
100 years after Turing's birth
Eurekalert.org noted the feat came 100 years after the birth of mathematician and computer scientist Alan Turing.
Turing had argued that since we will never be able to see inside a machine's hypothetical consciousness, the best measure of machine sentience is whether it can fool us into believing it is human.
His "Turing test" stands as one of the foundational definitions of what constitutes true machine intelligence.
"When this 'Turing test for game bots' competition was started, the goal was 50 percent humanness. It took us five years to get there, but that level was finally reached last week, and it's not a fluke," said Miikkulainen.
To mimic as much of the range of human behavior as possible, the team used a two-pronged approach.
Some behavior is modeled directly on previously observed human behavior, while the central battle behaviors are developed through a process of neuroevolution.
Neuroevolution runs AI neural networks through a survival-of-the-fittest gauntlet modeled on the biological process of evolution.
Networks that thrive in a given environment are kept, and the less fit are thrown away, with the holes being filled by copies of the fit ones and their "offspring."
Miikkulainen said that methods developed for the BotPrize competition should be useful not just in developing games but also in creating virtual training environments that are more realistic.
The skills learned may even be used in building robots that interact with humans in more pleasant and effective ways. — TJD, GMA News
Talk of the web