Id: arxiv/2606.26750 Url: https://arxiv.org/abs/2606.26750 Tags: math.CA, math.AP Score: 0 Title: Evolving edge weights via local entropy flow and cohesion flow on graphs Domain: math.CA Source: arxiv Authors: Juan Zhao, Jicheng Ma, Yunyan Yang, Liang Zhao Reviews: 0 Subject: math.CA Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-06-25 Page: https://delphi.science/r/arxiv/2606.26750 In this paper, we first propose two different quantities on graphs, namely local entropy and cohesion, then design two corresponding flows for edge weights: the local entropy flow and the cohesion flow. We establish the global existence and uniqueness of solutions for both flows and investigate their asymptotic behaviors, including the case that the limit goes to positive infinity. Moreover, they can be applied to fundamental network analysis tasks, including community detection and node classification. Empirical evaluations demonstrate that our method achieves performance competitive with…