Community detection in sparse time-evolving graphs with a dynamical Bethe-Hessian
Authors: Lorenzo Dall'Amico, Romain Couillet, Nicolas Tremblay
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Under the dynamical degree-corrected stochastic block model, in the case of two classes of equal size, we demonstrate and support with extensive simulations that our proposed algorithm is capable of making non-trivial community reconstruction as soon as theoretically possible, thereby reaching the optimal detectability threshold and provably outperforming competing spectral methods. |
| Researcher Affiliation | Academia | Lorenzo Dall Amico GIPSA-lab, UGA, CNRS, Grenoble INP lorenzo.dall-amico@gipsa-lab.fr Romain Couillet GIPSA-lab, UGA, CNRS, Grenoble INP L2S, Centrale Supélec, University of Paris Saclay Nicolas Tremblay GIPSA-lab, UGA, CNRS, Grenoble INP |
| Pseudocode | Yes | Algorithm 1 Community detection in sparse, heterogeneous and dynamical graphs |
| Open Source Code | Yes | On top of Python codes to reproduce most of the figures of this paper (available in the supplementary material), we provide an efficient Julia implementation, part of the Co De Bet He package (community detection with the Bethe-Hessian), available at github.com/lorenzodallamico. |
| Open Datasets | Yes | This section shows the results of our experiments on the Primary school network [42, 43] of the Socio Patterns project. The dataset contains a temporal series of contacts between children and teachers of ten classes of a primary school. |
| Dataset Splits | No | The paper describes using synthetic datasets and the Sociopatterns Primary school network for evaluation, but it does not specify explicit training, validation, or test dataset splits. Clustering algorithms often do not have a clear 'training' phase in the same way as supervised learning models. |
| Hardware Specification | Yes | The laptop s RAM is 7.7 Gb with Intel Core i7-6600U CPU @ 2.6GHz x 4. |
| Software Dependencies | No | The paper mentions 'Python codes' and 'Julia implementation' as part of the 'Co De Bet He package' but does not specify version numbers for these languages or any specific libraries/dependencies used. |
| Experiment Setup | No | The paper describes a spectral clustering algorithm and does not detail specific experimental setup parameters such as learning rates, batch sizes, epochs, or optimizer settings typically found in machine learning experiments. Inputs like 'label persistence, η' and 'number of clusters k' are mentioned, but these are model parameters rather than training hyperparameters. |