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].
On the Structure of Synergies in Cooperative Games
Authors: Ariel Procaccia, Nisarg Shah, Max Tucker
AAAI 2014 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Finally, we complement our theoretical results with experiments dealing with two real-world WVGs: decision making in the European Union (EU) and in the International Monetary Fund (IMF). |
| Researcher Affiliation | Academia | Ariel D. Procaccia Carnegie Mellon University EMAIL Nisarg Shah Carnegie Mellon University EMAIL Max Lee Tucker Carnegie Mellon University EMAIL |
| Pseudocode | No | The paper contains mathematical derivations and proofs but does not include any pseudocode or clearly labeled algorithm blocks. |
| Open Source Code | No | The paper does not contain any statement about making its code open-source or providing a link to a code repository. |
| Open Datasets | Yes | The EU game consists of 28 agents with weights varying from 3 to 29 (total weight 352), and a quota of 260 (Edward and Lane 2013). The IMF game consists of 128 agents with weights varying from 0.03 to 16.75 (total weight 100). For most policy decisions, the IMF uses simple majority (50% quota), while some decisions require supermajority quotas of 70% and 85% (Weiss 2012). |
| Dataset Splits | No | The paper describes the 'games' (EU and IMF) and their parameters for analysis but does not specify any training/validation/test dataset splits. The experiments involve computational analysis of game theory models, not machine learning model training. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used to run the computational experiments. |
| Software Dependencies | No | The paper states: "We used the dynamic programming algorithm of (Matsui and Matsui 2000; Bachrach and Shah 2013) for computing the Shapley values in WVGs." While it mentions the algorithm used, it does not specify any software names with version numbers (e.g., programming languages, libraries, or specific computational tools with their versions) that would be needed for reproduction. |
| Experiment Setup | Yes | We used the dynamic programming algorithm of (Matsui and Matsui 2000; Bachrach and Shah 2013) for computing the Shapley values in WVGs. We use heat maps to represent synergy and antagonism in any WVG. A heat map of a WVG is a square image where on both axes agents are sorted in increasing order of their weights, from top to bottom and from left to right. Thus, the entry in row i and column j represents the synergy or antagonism between the agent with the i th lowest weight and the agent with the j th lowest weight. In the plain heat map, a cell is colored blue (dark gray in grayscale) if the corresponding pair of agents is synergistic, and colored red (light gray in grayscale) if it is antagonistic. In the gradient heat map, the colors have varying intensity, which reο¬ect the magnitude of synergy or antagonism between various pairs. ... Figures 2(a), 2(b), 2(d), and 2(c) show the plain heat maps of the EU game with quotas 20%, 40%, 60%, and 80%, respectively. |