GDL-III: A Description Language for Epistemic General Game Playing
Authors: Michael Thielscher
IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We also show that it significantly enhances the expressiveness of its predecessor GDL-II by formally proving that termination of games becomes undecidable, and we present experimental results with a reasoner for GDL-III applied to general epistemic puzzles. ... To test the scalability of a GDL-III legal reasoner, we ran experiments on a 2.8 GHz processor with 8 GB of RAM with http://potassco.sourceforge.net, an offthe-shelf answer set solver, for interpreting GDL-III rules. Times are reported in seconds (CPU time). |
| Researcher Affiliation | Academia | Michael Thielscher School of Computer Science and Engineering University of New South Wales, Australia mit@unsw.edu.au |
| Pseudocode | No | The paper includes figures (e.g., Figure 1, Figure 2, Figure 3) showing GDL-III game rules and encoding examples, which are code-like. However, these are descriptions of the language itself, not pseudocode for an algorithm implemented by the authors, nor are they explicitly labeled as 'Pseudocode' or 'Algorithm'. |
| Open Source Code | No | The paper mentions using 'http://potassco.sourceforge.net, an offthe-shelf answer set solver, for interpreting GDL-III rules'. This refers to a third-party tool, not open-source code released by the authors for their specific methodology. No explicit statement or link to the authors' own source code is provided. |
| Open Datasets | No | The paper uses examples like 'Muddy Children Puzzle' and 'Cheryl's Birthday' and describes generating 'random problem instances' for experiments. While these are common problem domains, the paper does not provide concrete access information (link, DOI, formal citation with authors/year) to a public or open dataset in the required format for reproducibility. |
| Dataset Splits | No | The paper describes generating '1,000 random problem instances' and varying problem sizes for the Cheryl's Birthday puzzle. However, it does not specify explicit training, validation, or test dataset splits in terms of percentages, sample counts, or references to predefined splits, as is common in machine learning contexts. |
| Hardware Specification | Yes | To test the scalability of a GDL-III legal reasoner, we ran experiments on a 2.8 GHz processor with 8 GB of RAM |
| Software Dependencies | No | The paper states it used 'http://potassco.sourceforge.net, an offthe-shelf answer set solver, for interpreting GDL-III rules.' While it names the software, it does not provide a specific version number for Potassco or any other ancillary software. |
| Experiment Setup | Yes | The original problem consists of 10 dates across 4 different months and 6 different days. We kept a similar ratio of different months and days as we increased the problem size (# of dates) in order to ensure that the randomly chosen instances are equally difficult in that the number of solutions averages to approximately one.7 The results are summarised in the table below for each size and averaged over 1,000 random problem instances. |