Author Correction: Community Detection on Networks with Ricci Flow

Chien-Chun Ni, Yu-Yao Lin, Feng Luo, Jie GaoScientific 2019 · DOI · OpenAlex

Abstract: This formal amendment addresses an omission in the original publication of "Community Detection on Networks with Ricci Flow" (Scientific Reports, 2019). It provides the official link to the open-source code repository, ensuring the reproducibility of the geometric algorithms used for community detection in complex networks.

Introduction

This document serves as an official Author Correction for the research article titled "Community Detection on Networks with Ricci Flow," which was originally published in Scientific Reports in July 2019. The primary purpose of this amendment is to rectify an omission in the "Data Availability" section by providing the necessary link to the study's open-source code repository.

In the original publication, the "Data Availability" section erroneously stated that only datasets were available upon request, failing to include a link to the computational tools developed by the authors. This correction ensures that the computational framework is accessible to the broader scientific community, adhering to modern standards of transparency and reproducibility.

The Geometric Approach to Community Detection

In the original study, the authors addressed the challenge of community detection—identifying clusters of densely connected nodes within a network—through a novel geometric lens. While traditional methods often rely on modularity optimization or random walks, this work introduced the use of Ricci flow, a concept borrowed from differential geometry and famously utilized in the proof of the Poincaré conjecture.

The authors proposed that the geometric properties of a network could reveal its underlying community structure more effectively than purely topological measures. By applying a discrete adaptation of the Ollivier-Ricci curvature, the authors developed a method to "smooth out" networks. This process involves two main computational components:

  1. Graph Ricci Curvature: This metric measures the connectivity of local neighborhoods. In a network, edges within a dense community typically exhibit positive curvature, while "bridge" edges between communities often exhibit negative curvature.
  2. Ricci Flow on Graphs: This is an iterative process that evolves edge weights based on their curvature. Under this flow, edges with negative curvature are "stretched" (their weights are decreased or the edges are removed), while edges within communities are strengthened. This process effectively "flows" the network toward a state where clusters are naturally separated and easily identifiable.

Without the source code, other researchers would face significant hurdles in implementing or verifying this sophisticated geometric technique.

Updated Data and Code Availability

The correction formally updates the "Data Availability" section to include the computational tools developed by the authors. The source code for the Graph Ricci Curvature and Ricci flow algorithms is now publicly accessible.

Code Repository: https://github.com/saibalmars/GraphRicciCurvature

Regarding the datasets used in the study:

  • Generated/Analyzed Datasets: These remain available from the corresponding author, Jie Gao ([email protected]), upon reasonable request.
  • Computational Framework: The full codebase for implementing the geometric analysis, including the Ricci curvature and Ricci flow computations, is now open-source via the GitHub link provided above.

Significance for Open Science

Reproducibility is a cornerstone of modern scientific research. For methods as mathematically specialized as Ricci flow on networks, providing the source code is as critical as the paper itself for peer validation and further development. The existence of this correction highlights the evolving standards of open science and the continued interest in using "continuous" mathematical tools (like curvature and flow) to solve "discrete" problems in network science.

The release of this software resource ensures that the scientific community can implement, verify, and build upon the geometric network analysis tools developed by the authors, fostering further innovation in the field of community detection.

Summary updated

Abstract

An amendment to this paper has been published and can be accessed via a link at the top of the paper.
BibTeX
@inproceedings{ni2019author,
  author = {Chien-Chun Ni and Yu-Yao Lin and Feng Luo and Jie Gao},
  title = {{Author Correction: Community Detection on Networks with Ricci Flow}},
  booktitle = {Scientific},
  year = {2019},
  doi = {10.1038/s41598-019-49491-5},
}
Figure