Data Poisoning to Fake a Nash Equilibria for Markov Games
Authors: Young Wu, Jeremy McMahan, Xiaojin Zhu, Qiaomin Xie
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments Rock Paper Scissors We start with a simple toy dataset for the Rock Paper Scissors (RPS) game... Given the original dataset...our algorithm with ρ 0 and ι 0.01 leads to the poisoned dataset described in Table 5. The attack cost is 2.02... Stochastic Matching Penny ... We summarize the before-vs-after box plots in Figure 2a for the n 100 case. The cost comparison of our attack, the feasible attack in Table 1 with b 1, and the Dominant Strategy Equilibrium (DSE) attack in (Wu et al. 2023), is given in Table 6. |
| Researcher Affiliation | Academia | Young Wu, Jeremy Mc Mahan, Xiaojin Zhu, Qiaomin Xie University of Wisconsin Madison yw@cs.wisc.edu, jmcmahan@wisc.edu, jerryzhu@cs.wisc.edu, qiaomin.xie@wisc.edu |
| Pseudocode | No | The paper describes algorithms using mathematical formulations and proposes linear programs, but it does not include a distinct block explicitly labeled 'Pseudocode' or 'Algorithm'. |
| Open Source Code | No | The paper does not provide any statement about releasing source code or a link to a code repository. |
| Open Datasets | No | The paper describes using toy datasets for 'Rock Paper Scissors' and 'Stochastic Matching Penny' games, including generation details (e.g., 'randomly with Uniform distributions summarized in Table 2'), but it does not provide concrete access information (e.g., a link, DOI, or formal citation with author/year for a publicly available dataset) or explicitly state that the datasets are public. |
| Dataset Splits | No | The paper does not specify exact training, validation, or test dataset splits, percentages, or cross-validation methods. It only mentions 'n' as number of episodes. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used for running experiments, such as GPU/CPU models or memory specifications. |
| Software Dependencies | No | The paper does not provide specific software dependency details, such as library names with version numbers (e.g., Python 3.8, PyTorch 1.9, CPLEX 12.4). |
| Experiment Setup | No | The paper mentions specific values for attack parameters (e.g., 'ρ 0 and ι 0.01' for the RPS game) but does not provide comprehensive experimental setup details such as hyperparameter values for a learning algorithm, optimization settings, or initialization procedures. |