Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Distances Between Top-Truncated Elections of Different Sizes
Authors: Piotr Faliszewski, Jitka Mertlová, Pierre Nunn, Stanisław Szufa, Tomasz Wąs
AAAI 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In the experiments, we largely focus on synthetic elections, as such data is easier to control, but our main motivation is to visualize real-life data. Hence, in Section 5 we form a map of a large fragment of Preflib, a database of real-life elections (see Figure 4). |
| Researcher Affiliation | Academia | 1AGH University 2Czech Technical University in Prague 3Université de Rennes 4CNRS, LAMSADE, Université Paris Dauphine-PSL 5University of Oxford |
| Pseudocode | No | The paper describes methods verbally and mathematically but does not present any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Code https://github.com/Project-PRAGMA/Map Different-Sizes-AAAI-2025 |
| Open Datasets | Yes | Preflib database (Mattei and Walsh 2013) |
| Dataset Splits | Yes | We generated the size-oriented dataset in the same way as the basic one, except that for each culture we partitioned its elections into four groups, with either 8 or 16 candidates and either 96 or 192 voters. We obtain top-truncated elections from complete ones by using the following methods: [...] To form the comprehensive dataset, we took the sizeoriented one and for each group of elections (of a given size, generated using a given statistical culture) we left the first half of the elections in the group intact, we applied the top-k truncation to the next quarter of them, and we applied random cut truncation to the last quarter. We chose the truncation parameters so that, in expectation, each voter ranked half of the candidates. We generated the truncation-oriented dataset in the same way, but starting from the basic dataset. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers like Python 3.8, CPLEX 12.4) needed to replicate the experiment. |
| Experiment Setup | Yes | We chose the truncation parameters so that, in expectation, each voter ranked half of the candidates. [...] The normalized Mallows model is similar but the votes are clustered around a given central one (the strength of this clustering is controlled by parameter norm-ϕ [0, 1], the higher the value the less concentrated are the votes; see the work of Boehmer, Faliszewski, and Kraiczy (2023) for a discussion of this model). The urn model generates elections with clusters of identical votes (the larger its parameter of contagion α 0 is, the fewer clusters there are, each containing more votes (Faliszewski et al. 2023)). |