Analysis of Robustness of a Large Game Corpus Conference: International Conference on the Foundations of Digital Games; April 15–18, 2025; Graz, Austria International Conference on the Foundations of Digital Games (FDG ’25), April 15–18, 2025, Graz, Austria DOI: 10.1145/3723498.3723820 ISBN: /25/04
Mahsa Bazzaz, Seth Cooper · 2025
Read Analysis of Robustness of a Large Game Corpus Conference: International Conference on the Foundations of Digital Games; April 15–18, 2025; Graz, Austria International Conference on the Foundations of Digital Games (FDG ’25), April 15–18, 2025, Graz, Austria DOI: 10.1145/3723498.3723820 ISBN: /25/04 by Mahsa Bazzaz, Seth Cooper (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-03940.txt.
Description
Mahsa Bazzaz and Seth Cooper study how sturdy game level data are for machine learning, in a 2025 paper. Learning tools can craft game levels, and flat tile levels are a standard test set. Levels obey strict local and global rules, and tiny tile flips can break a level. That fragility sets game data apart from photo or text data.
The authors define sturdiness as how well output holds when input shifts a bit. They grade levels from four games against top learning data sets on that scale. They also built a large new level set from four classic-style games. The set aims to ease the data shortage that dogs the field.
Catalog record
| CREATOR | Mahsa Bazzaz, Seth Cooper |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-03940