This seems to be related to ICM as it relates to propagating beliefs to be more consistent. This may be an issue with learning the prior in general where the human prior is actually not stable to the pressure to be consistent (but maybe CEV is stable to this?)
This seems to be related to ICM as it relates to propagating beliefs to be more consistent. This may be an issue with learning the prior in general where the human prior is actually not stable to the pressure to be consistent (but maybe CEV is stable to this?)