NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:6251
Title:Optimizing Generalized PageRank Methods for Seed-Expansion Community Detection

This paper considers different weighting schemes for aggregating landing probabilities with applications to graph clustering. There is consensus that the submission presents some interesting and solid contributions, both theoretically and empirically. The proposed Inverse PageRank is interesting in that the weights increase for the first few landing probability walk lengths (however, this is also true for the heat kernel). A few paper clarification issues outlined by the reviewers should be addressed in a camera version.