Skip to main content
. Author manuscript; available in PMC: 2010 May 12.
Published in final edited form as: Proc Int Conf Mach Learn. 2008;301:256–263. doi: 10.1901/jaba.2008.301-256

Figure 1.

Figure 1

Boxplot of POMDP learning performance with a discrete set of four possible models. The medians of the policies are comparable, but the active learner (left) makes fewer mistakes than the passive learner (center). The Bayes risk action selection criterion (right) does not cause the performance to suffer.