Selfish Creation of Social Networks

Authors: Davide Bilò, Tobias Friedrich, Pascal Lenzner, Stefanie Lowski, Anna Melnichenko5185-5193

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Moreover, as a proof-of-concept we show via experiments that the created equilibrium networks of our model indeed closely mimics real-world social networks. We observe degree distributions that seem to follow a power-law, high clustering, and low diameters.
Researcher Affiliation Academia 1 Department of Humanities and Social Sciences, University of Sassari, Italy 2 Hasso Plattner Institute, University of Potsdam, Germany 3 Department of Computer Science, Humboldt-University Berlin, Germany
Pseudocode No The paper describes the dynamics of the SNCG verbally in paragraph form but does not include any structured pseudocode or algorithm blocks.
Open Source Code Yes The source code we used can be found at https://github.com/melnan/distNCG.git.
Open Datasets No The paper describes generating initial sparse networks (cycle, random spanning tree, Erdős-Renyi) for simulations rather than using a pre-existing, publicly available dataset that would typically involve a "train" split.
Dataset Splits No The paper does not mention any training, validation, or test dataset splits, as it focuses on simulations starting from generated initial network structures.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory) used to run its experiments.
Software Dependencies No The paper mentions using a function σ(x) = 2log2(n)xα and refers to source code on GitHub, but it does not specify any software dependencies with version numbers (e.g., Python version, specific libraries).
Experiment Setup Yes For all experiments we choose σ(x) = 2 log2(n) xα, where n N (the number of agents) and α R (the exponent) are input parameters. ... In each step of our simulations one agent is activated uniformly at random and this agent then performs the best possible edge addition (jointly with the other endpoint if the respective agent agrees) or edge deletion. ... Figure 6 shows the box-and-whiskers plot for the average clustering coefficient of the pairwise stable networks obtained by the algorithm for n = 1000 with respect to the value of the power coefficient α. Figure 7 shows a degree distribution for the resulting pairwise stable networks for n = 3000. ... All our experiments show that the power-law exponent γ is between 2 and 3.