Combining rank-based and distance-based vote smoothing

Construct and analyze an evolutionary private synthesizer that combines WASP's rank-based multi-candidate vote sharing with BF-Soft's distance-based softmax weighting, and determine whether the combination improves utility or privacy-sensitive calibration.

Background

WASP spreads voting mass across a selected group of candidates using rank-based weights, whereas BF-Soft assigns voting mass according to continuous distances through a temperature-scaled softmax. The paper treats these as complementary forms of vote relaxation: one changes how many candidates receive mass, while the other makes the response vary continuously with the magnitude of a protected substitution.

The authors state that the two approaches are compatible but do not develop or evaluate a mechanism that combines them, explicitly leaving this integration for future work.