A Novel Approach to Solving Goal-Achieving Problems for Board Games
Authors: Chung-Chin Shih, Ti-Rong Wu, Ting Han Wei, I-Chen Wu10362-10369
AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In our experiments, among a collection of 20 7x7 L&D problems, all are solved with RZS integrated into Alpha Zero with FTL; with RZS without FTL, 5 can be solved; without RZS nor FTL, none can be solved. Results are shown in Table 1. |
| Researcher Affiliation | Academia | 1Department of Computer Science, National Yang Ming Chiao Tung University, Hsinchu, Taiwan 2Research Center for Information Technology Innovation, Academia Sinica, Taipei, Taiwan 3Department of Computing Science, University of Alberta, Edmonton, Canada |
| Pseudocode | No | The paper describes the algorithm steps in a structured manner but does not provide a formal pseudocode block or algorithm environment. |
| Open Source Code | Yes | The above L&D problems, statistics, and training details can be accessed via the Github repository (Shih et al. 2021). |
| Open Datasets | Yes | We select problems from a well-known L&D problems book, aptly named the Life and Death Dictionary , written by Cho Chikun [1987], a Go grandmaster; problem difficulties range from beginner to professional levels. Second, we trained a 19x19 Alpha Zero Go program with the FTL method by using the problems from the tsumego book The Training of Life and Death Problems in Go (Shao, Ding, and Liu 1991). |
| Dataset Splits | No | The paper does not explicitly provide specific percentages or counts for training, validation, and test dataset splits for model training. |
| Hardware Specification | No | The paper mentions 'computing resources are partially supported by National Center for Highperformance Computing (NCHC) of Taiwan' but does not specify exact hardware details such as GPU/CPU models or memory. |
| Software Dependencies | No | The paper mentions software like 'Alpha Zero' and 'ELF Open Go' but does not provide specific version numbers for these or other libraries. |
| Experiment Setup | Yes | A problem is marked as proven if a program can prove it within 500,000 simulations. The time budget for each move is limited to 5 minutes, the same as the setting in (Kishimoto and M uller 2005b). We also added two additional feature planes to represent the crucial stones of each problem for both players. |