Jake Young is a 
Video games of late have gotten crazy complicated. Making life-like characters and realistic worlds is an incredible computational challenge. Popular Science lists the 10 Biggest Challenges in video game production, and number 4 struck my eye:
4. Artificial Intelligence Like teaching 1,000 kids to think for themselves overnightProblem: Once upon a time, the bad guys in videogames wandered around mindlessly, shooting at you while they waited to die. That doesn't cut it anymore. Players demand sophisticated enemies to fight and reliable in-game allies with which to fight them. Thing is, it's freaking complicated, and it eats up processor speed. "We're faking just enough smarts to make it work," says Mathieu Mazerole, lead engineer on Ubisoft's Assassin's Creed.
Status: Imbuing characters in a game with lifelike decision-making ability involves employing the kind of high-level logic theories -- learning decision trees, mobile navigation, finite-state machine models -- used by top robotics engineers. In games like Assassin's Creed, such calculations cause your pursuers to form squadrons, scale buildings, and dash across rooftops to get you. "Sometimes they find paths we've never thought of," says Mark Bresner, the game's lead AI programmer. "You really feel threatened." It takes particular skill to make the AI seem unskilled. "Missing a target believably is really hard," says game programmer Neil Johnson, of the World War II game Brothers in Arms: Hell's Highway. "They automatically shoot straight along the sight line, deadly accurate every time, so we have to adjust for dramatic spurts of fire at the player's feet, things like that."
You know there is a group of people who have some ideas about decision making: neuroscientists (probably of a psychology-sort). I wonder if these companies are hiring neuroscientists to make the AI in the characters more realistic. It might help with the problem of making realistic errors because human beings and animals make errors in relatively defined ways. (Maybe there is another career out there for me when the grants dry up.)
Actually I imagine the most troublesome bit in programming decision making is making the AI unpredictable. How do you encode the AI so that it is realistic and learns, but isn't like the villains from Megaman where you can just memorize the repeated elements in their movements? That's tough stuff.
Hat-tip: Slashdot
Comments
"You know there is a group of people who have some ideas about decision making: neuroscientists (probably of a psychology-sort). I wonder if these companies are hiring neuroscientists to make the AI in the characters more realistic"
or, just plain psychologists jake. not to mention cognitive scientists. and yeah, I know a couple in game companies.
so called "neuroscientists" are actually kind of counter productive. The digital computer is not a human brain and in gaming the idea is to make the characters act realistically. who cares whether this is accomplished by human-brain-like computation or not?
now, the cog scientists will argue that brain-like computing will get you some efficiencies but that's another discussion...
Posted by: Drugmonkey | September 21, 2007 6:25 PM