NeurIPS 2020

Community detection in sparse time-evolving graphs with a dynamical Bethe-Hessian

Meta Review

This paper provides some extensions on detectability of community structure in dynamic networks. There was substantial discussion of the contributions with respect to Reference [14] (Ghasemanian et al., PRX 2016 / arXiv:1506.06179). In the end, there was consensus that the contributions from this paper (both theoretical and algorithmic) are still meaningful, and so the recommendation is to accept. However, the referees and AC strongly encourage the paper to make the following changes for a camera version: 1. Give proper credit to the ideas in the arXiv version of Reference [14] and provide a fair comparison to the existing ideas in Figure 3 (see Reviewer 1’s suggestions). 2. Include the results on real-world data that were summarized in the author response.