A research team from the Swiss Federal Institute of Technology in Lausanne won an international artificial intelligence (AI) contest at the recent International Joint Conference on Artificial Intelligence in Beijing with software that replicates a human-like way of playing the Angry Birds video game. The competition required that participants' machines had to solve the game's levels while only having access to the same information available to humans.
The winning software was created by researchers Jason Jingshi Li, Mirko Katanic, and Arnaud Jutzeler. Their goal in the contest was the creation of a perfect intelligence that can beat anyone analyzing the game while using the same tools available to people on their smartphones. "What we're trying to do here is to understand and define intelligence in order to emulate it," Li says.
The researchers say the resulting insights would be applicable to many AI disciplines, such as the processing of surveillance videos. The team used the exploitation versus exploration precept, which involves identifying which tactics to use based on the level the player is at. They first determined a way of beating every type of level the player may encounter, and then developed an algorithm tasked with choosing the strategy most likely to succeed based on previous attempts.
From Swiss Federal Institute of Technology
View Full Article
Abstracts Copyright © 2013 Information Inc., Bethesda, Maryland, USA