Selecting the Most Conflicting Pair of Candidates

Authors: Théo Delemazure, Łukasz Janeczko, Andrzej Kaczmarczyk, Stanisław Szufa

IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We support our theoretical study with experiments on both real-life and synthetic data.
Researcher Affiliation Academia 1CNRS, LAMSADE, Université Paris Dauphine PSL 2AGH University
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes The experiments source code is freely available at https://github.com/Project-PRAGMA/conflictual-rules--IJCAI-24.
Open Datasets Yes We study 4 datasets: (i) preferences over 11 candidates gathered for experiments during French presidential elections in 2017 and 2022 [Bouveret et al., 2018; Delemazure and Bouveret, 2022], [...] (ii) preferences over 10 sushi types [Kamishima, 2003] and (iii) juries ranking of contestant performances in figure skating competitions, from Preflib [Mattei and Walsh, 2013]
Dataset Splits No The paper describes generating profiles and simulating rules on them, but it does not specify explicit train/validation/test dataset splits with percentages, sample counts, or predefined citations in the traditional machine learning sense.
Hardware Specification No The paper does not explicitly describe the specific hardware (e.g., GPU models, CPU types, memory amounts) used to run its experiments.
Software Dependencies No The paper does not provide specific software dependencies, such as library names with version numbers, needed to replicate the experiment.
Experiment Setup Yes For each instance, we marked on the plane the two selected candidates, highlighting the pairs for three random instances for reference. [...] This way, we always generated the same number n = 100 of voters and m = 10 of candidates. For each generated profile, we simluated our rules. Then, we compared the values of the select pairs polarization metrics... We repeated this experiment for 1000 random profiles. [...] We sampled the positions on [0, 1]2 of all voters and candidates using: (1) the uniform distribution and (2) the normal distribution centered in (0.5, 0.5).