Dynamic Thresholding and Pruning for Regret Minimization
Authors: Noam Brown, Christian Kroer, Tuomas Sandholm
AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments demonstrate a substantial improvement in performance for Hedge as well as the excessive gap technique. |
| Researcher Affiliation | Academia | Noam Brown, Christian Kroer, Tuomas Sandholm Carnegie Mellon University, Computer Science Department, noamb@cmu.edu, ckroer@cs.cmu.edu, sandholm@cs.cmu.edu |
| Pseudocode | No | The paper provides mathematical equations and descriptions of algorithms but no explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any statement or link indicating that its source code is publicly available. |
| Open Datasets | Yes | We tested dynamic thresholding with and without RBP on a standard benchmark game called Leduc Hold em (Southey et al. 2005) and an enlarged variant of Leduc Hold em featuring more actions, called Leduc-5. |
| Dataset Splits | No | The paper describes using Leduc Hold'em and Leduc-5 datasets but does not provide specific information regarding training, validation, or test splits (e.g., percentages, sample counts, or explicit split methodologies). |
| Hardware Specification | No | The paper discusses general performance characteristics related to hardware ('multi-core implementations,' 'memory access is the bottleneck') and acknowledges computing resources ('XSEDE computing resources provided by the Pittsburgh Supercomputing Center'), but it does not specify exact hardware components (e.g., CPU/GPU models, memory sizes) used for running its experiments. |
| Software Dependencies | No | The paper discusses algorithms and techniques (e.g., Hedge, Regret Matching, EGT, CFR) but does not list any specific software dependencies with version numbers (e.g., programming languages, libraries, frameworks, or solvers with their corresponding versions) that would be needed for replication. |
| Experiment Setup | Yes | In our experiments, we set ηt = t, where VAR(I)t is the observed variance of v(I) up to iteration t, based on a heuristic by Chaudhuri et al. (2009). In our experiments with EGT, we stop traversing a branch in the game tree when the probability (over nature and the opposing player) of the branch falls below c T for various values of c. For EGT, we threshold by c T , where the number shown in the legend is c. For Hedge, we threshold by d T , where d is shown in the legend. |