Liquid Democracy with Ranked Delegations

Authors: Markus Brill, Théo Delemazure, Anne-Marie George, Martin Lackner, Ulrike Schmidt-Kraepelin4884-4891

AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental By performing an extensive experimental analysis on synthetic as well as real-world data, we compare delegation rules with respect to several quantitative criteria relating to the chosen paths and the resulting distribution of voting power. Our experiments reveal that delegation rules can be aligned on a spectrum reflecting an inherent trade-off between competing objectives.
Researcher Affiliation Academia 1 Research Group Efficient Algorithms, TU Berlin, Germany 2 LAMSADE, Universit e Paris-Dauphine, France 3 Department of Informatics, University of Oslo, Norway 4 Databases and Artificial Intelligence Group, TU Wien, Austria
Pseudocode Yes Diffusion: Initialize the set of assigned voters: A C. While (A = V \ I), repeat the following steps: 1. F arg min{r(e) | e \delta G(A)}, where \delta G(A) is the set of edges in G having their head in A and tail in V \A. 2. A A \cup {v | (v, w) \in F} 3. f(v) = ((v, w), f(w)) for all (v, w) \in F
Open Source Code Yes Our code can be found at https://github.com/Theo Dlmz/rankeddelegation.
Open Datasets Yes The real-world network we used for this method is a subgraph of the Facebook network (Viswanath et al. 2009).
Dataset Splits No The paper evaluates algorithmic rules and does not involve training machine learning models, therefore, it does not specify training, validation, or test dataset splits.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU, GPU models, or memory) used for running the experiments.
Software Dependencies No The paper provides a link to its code but does not explicitly list any software dependencies with their specific version numbers in the text.
Experiment Setup Yes For synthetic networks, the input network H is generated by the standard Erd os R enyi model G(n, p), where we chose the edge probability p [0, 1] such that the average number of edges per voter is equal to . We let n (respectively, m) denote the number of nodes (respectively, edges) of the graph G. Our experimental setup is described in detail in (Brill et al. 2021), where we also present the results for different generation methods, datasets, and parameters. (n = 1000, = 4, pc = 0.2, α = 2)