Metraipolis › archive › dc_bf893dfbfc22

Procedural Content Metageneration via Program Search and Continual Abstraction Discovery — arXiv:2608.17947

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.17947
License (as archived): CC0-1.0
Archived: 2026-10-18T14:28:53.089Z
SHA-256 of the archived text: af58440b0152af98be7ee32b0b1e1b7552fe8557cb3dd2cae3464bb5929d7af1
Stable URL of this page: https://metraipolis.com/documento/dc_bf893dfbfc22

How to cite

Procedural Content Metageneration via Program Search and Continual Abstraction Discovery — arXiv:2608.17947 — arXiv (en), https://arxiv.org/abs/2608.17947 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bf893dfbfc22 (Metraipolis archive dc_bf893dfbfc22, sha256 af58440b0152af98be7ee32b0b1e1b7552fe8557cb3dd2cae3464bb5929d7af1)

BibTeX

@misc{dcbf893dfbfc22,
  title = {Procedural Content Metageneration via Program Search and Continual Abstraction Discovery — arXiv:2608.17947},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.17947},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bf893dfbfc22; sha256 af58440b0152af98be7ee32b0b1e1b7552fe8557cb3dd2cae3464bb5929d7af1},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.17947. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models can generate executable programs, which makes it possible to search directly over procedural content generators rather than individual levels. We study this approach in Sokoban, Zelda, Dangerous Dave, and Lode Runner. Each run evolves complete Python generators through language-model mutation and crossover. We introduce Continual Abstraction Discovery, or CAD, which extracts reusable primitives from high-fitness programs into a run-specific helper module. A 2x2 experiment crosses CAD with access to a fixed hand-written domain API. The completed data set contains 160 complete runs, with at least ten 50-generation runs in every cell. CAD raises mean final best fitness in all eight domain and API comparisons. Across all CAD runs, learned libraries are adopted by most later programs and repeatedly rediscover validation, reachability, and structural utilities. These results support that discovering reusable primitives improves evolutionary program search for content generators.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-19 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE).
Pagina del lavoro: https://arxiv.org/abs/2608.17947 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-18T14:27:54.316Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-19?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).