Margin of Victory in Tournaments: Structural and Experimental Results

Authors: Markus Brill, Ulrike Schmidt-Kraepelin, Warut Suksompong5228-5235

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

Reproducibility Variable Result LLM Response
Research Type Experimental Furthermore, we provide experimental evidence on the extent to which the Mo V notion refines winner sets in tournaments generated according to various stochastic models.
Researcher Affiliation Academia Markus Brill,1 Ulrike Schmidt-Kraepelin,1 Warut Suksompong2 1 Research Group Efficient Algorithms, TU Berlin, Germany 2 School of Computing, National University of Singapore, Singapore {brill,u.schmidt-kraepelin}@tu-berlin.de, warut@comp.nus.edu.sg
Pseudocode No The paper contains theoretical results and experimental descriptions, but no structured pseudocode or algorithm blocks are present.
Open Source Code Yes The code for our implementation can be found at http://github.com/uschmidtk/MoV.
Open Datasets Yes We used six stochastic models to generate preferences: the uniform random model (which was used in Section 3.3), two variants of the Condorcet noise model (with and without voters), the impartial culture model, the P olya Eggenberger urn model, and the Mallows model. Detailed descriptions of these models can be found in the full version of our paper (Brill et al. 2020a). For implementing the Mallows and urn models, we utilized implementations contributed by Mattei and Walsh (2013).
Dataset Splits No The paper states 'For each stochastic model and each number of alternatives n {5, 10, 15, 20, 25, 30}, we sampled 100 tournaments,' which describes the number of instances for evaluation, but does not provide details on training, validation, or test splits.
Hardware Specification Yes The experiments were carried out on a system with 1.4 GHz Quad-Core Intel Core i5 CPU, 8GB RAM, and mac OS 10.15.2 operating system.
Software Dependencies Yes The software was implemented in Python 3.7.7 and the libraries networkx 2.4, matplotlib 3.2.1, numpy 1.18.2, and pandas 1.0.3 were used.
Experiment Setup Yes For each stochastic model and each number of alternatives n {5, 10, 15, 20, 25, 30}, we sampled 100 tournaments. Using the methods described by Brill et al. (2020b), we implemented algorithms to calculate the Mo V values for CO, UC, 3-kings, and TC. Due to their computational intractability, we did not implement procedures to calculate the Mo V values for BA and k-kings for k 4.