Uncertainty reduction: we need to identify model inputs that will cause significant uncertainty in the output. An example is Uncertain Volatility Models (UVM)

in option pricing problems in which all we can say about the volatility is that its values are constrained to be in a given bounded interval.