We show that the algorithm to extract diverse M -solutions from a Conditional
Random Field (called divMbest ) takes exactly the form of a Herding
procedure , i.e. a deterministic dynamical system that produces a sequence
of hypotheses that respect a set of observed moment constraints. This
generalization enables us to invoke properties of Herding that show that
divMbest enforces implausible constraints which may yield wrong assumptions for
some problem settings...
Our experiments in semantic segmentation demonstrate
that seeing divMbest as an instance of Herding leads to better alternatives for
the implausible constraints of divMbest.