The KNDN Problem: A Quest for Unity in Diversity

Given a query location Q in multi-dimensional space, the classical KNN problem is to return the K spatially closest answers in the database with respect to Q. The KNDN (K-Nearest Diverse Neighbor) problem is a semantic extension where the objective is to return the spatially closest result set such that each answer is sufficiently different, or diverse, from the rest of the answers. We review here the KNDN problem formulation, the associated technical challenges, and candidate solution strategies.