Identify the Nash Equilibrium in Static Games with Random Payoffs
Authors: Yichi Zhou, Jialian Li, Jun Zhu
ICML 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We now empirically verify the sample complexity of our algorithms. We choose a simple algorithm as our baseline (denoted by ALL), which pulls all arms at each round until stopping. The stopping and recommendation rules are the same as LUCB-G, and the confidence bound for this baseline is slightly different, see Appendix G for details. We evaluate on synthetic 5 5 games, where the payoffs are all random Bernoulli variables. The first game has a NE, while the second game has no NE. The results are shown in Fig. 1, where both axes are in log-scale with base 10. The number of pulls needed for both games are shown in Fig. 1(a) and Fig. 1(c) separately and our algorithms outperform the baseline (i.e., ALL). Fig. 1(b) shows the number of pulls on s / nei(s ) in the first game and we can see that this number is bounded, agreeing with our analysis. |
| Researcher Affiliation | Academia | 1Dept. of Comp. Sci. & Tech., TNList Lab, State Key Lab for Intell. Tech. & Systems, CBICR Center, Tsinghua University. |
| Pseudocode | Yes | Algorithm 1 LUCB-G |
| Open Source Code | No | The paper does not provide any statement or link regarding the availability of open-source code for the described methodology. |
| Open Datasets | No | We evaluate on synthetic 5 5 games, where the payoffs are all random Bernoulli variables. The first game has a NE, while the second game has no NE. |
| Dataset Splits | No | The paper does not specify exact percentages, sample counts, or predefined splits for training, validation, or test datasets. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used to run the 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 | No | The paper describes the evaluation on 'synthetic 5x5 games' and compares with a 'baseline (ALL)' but does not provide specific experimental setup details such as hyperparameters or training configurations. |