EUVIMEDEUVIMED entdecken
Crossref

On Sample-Efficient Generalized Planning via Learned Transition Models

Nitin Gupta, Vishal Pallagani, John A. Aydin, Biplav Srivastava

Proceedings of the International Conference on Automated Planning and Scheduling · 08.06.2026 · 36 (1) · 698-702

DOI: 10.1609/icaps.v36i1.42888 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the International Conference on Automated Planning and Scheduling, online erschienen am 08.06.2026. Verfasst von Nitin Gupta und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In this work, we formulate generalized planning as transition-model learning: a neural model approximates the successor-state function and generates plans by rolling out symbolic state trajectories. To achieve size-invariant generalization, we evaluate multiple state representations, including graph embeddings. Our results show that learning explicit transition models yields higher out-of-distribution success than action-sequence prediction in multiple domains, despite significantly fewer training instances and smaller models.

Zitieren

APA

Gupta, N., Pallagani, V., Aydin, J. A., & Srivastava, B. (2026). On Sample-Efficient Generalized Planning via Learned Transition Models. Proceedings of the International Conference on Automated Planning and Scheduling, 36(1), 698-702. https://doi.org/10.1609/icaps.v36i1.42888

Vancouver

Gupta N, Pallagani V, Aydin JA, Srivastava B. On Sample-Efficient Generalized Planning via Learned Transition Models. Proceedings of the International Conference on Automated Planning and Scheduling. 2026;36(1):698-702. doi:10.1609/icaps.v36i1.42888

BibTeX
@article{Gupta2026,
  author = {Gupta, Nitin and Pallagani, Vishal and Aydin, John A. and Srivastava, Biplav},
  title = {On Sample-Efficient Generalized Planning via Learned Transition Models},
  journal = {Proceedings of the International Conference on Automated Planning and Scheduling},
  year = {2026},
  volume = {36},
  number = {1},
  pages = {698--702},
  doi = {10.1609/icaps.v36i1.42888}
}
Kostenlos starten →