Hindsight and Sequential Rationality of Correlated Play
Authors: Dustin Morrill, Ryan D'Orazio, Reca Sarfati, Marc Lanctot, James R Wright, Amy R Greenwald, Michael Bowling5584-5594
AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Figure 3: The gap between CFR s self-play empirical distribution and an extensive-form or agent-form (C)CE (E/AF(C)CE) in the extended Shapley s game with b = 0.003. (Left) simultaneous-update CFR. (Right) alternating-update (Burch, Moravcik, and Schmid 2019) CFR. In all of our experiments, optimal causal and action deviations achieved the same value. The E/AFCE lines correspond to an optimal informed deviation while the E/AFCCE lines correspond to an optimal blind deviation. Notice that in all figures, the gap does not continue to decrease over time as we would expect if CFR were to minimize causal or action regret. See Figure A.6 for experiments with two other bonus values (0.3 and 30). |
| Researcher Affiliation | Collaboration | Dustin Morrill1, Ryan D Orazio2, Reca Sarfati3, Marc Lanctot4, James R. Wright1, Amy R. Greenwald5, Michael Bowling1, 4 1University of Alberta; Alberta Machine Intelligence Institute, Canada 2Universit e de Montr eal; Mila, Canada 3Massachusetts Institute of Technology, United States 4Deep Mind 5Brown University, United States |
| Pseudocode | No | The paper describes algorithms and theoretical concepts but does not include structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | The game can be found in Open Spiel (Lanctot et al. 2019) under the name extended bos.efg. [...] The game can be found in Open Spiel (Lanctot et al. 2019) under the name extended shapleys.efg. |
| Open Datasets | Yes | The game can be found in Open Spiel (Lanctot et al. 2019) under the name extended bos.efg. [...] The game can be found in Open Spiel (Lanctot et al. 2019) under the name extended shapleys.efg. |
| Dataset Splits | No | The paper describes experiments in game theory simulations but does not specify train, validation, or test dataset splits. |
| Hardware Specification | No | The paper does not specify any particular hardware used for running experiments. |
| Software Dependencies | No | The paper mentions using 'Open Spiel (Lanctot et al. 2019)' as a framework for games, but it does not provide specific version numbers for Open Spiel or any other software dependencies. |
| Experiment Setup | No | The paper describes the game environments and theoretical frameworks, but it does not provide specific hyperparameter values or detailed system-level training settings for algorithms like CFR. |