For some reason snippyhollow's comment was `dead`ed:
> For the use of Bayesian models in StarCraft AI, see my last 3 papers http://emotion.inrialpes.fr/people/synnaeve/index.html#publi... :)
> For the use of Bayesian models in StarCraft AI, see my last 3 papers http://emotion.inrialpes.fr/people/synnaeve/index.html#publi... :)
The thread-like programs is an interesting approach, too. I wonder if the author plans to treat stale state using Bayesian belief models instead of assuming it's still correct until a scheduled poll checks it.