EUVIMEDEUVIMED entdecken
Crossref

Meta-Normalizing Flow for Data-Limited Offline Meta-Reinforcement Learning (Student Abstract)

Lianghui Liu, Zongzhang Zhang

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (48) · 41278-41280

DOI: 10.1609/aaai.v40i48.42241 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Lianghui Liu und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Offline Meta-Reinforcement Learning (OMRL) leverages pre-collected data to adapt to new tasks. Context-based methods learn task representations from contexts. However, the context is influenced by both the task and the behavior policy. The mismatch between the behavior policy and the testing policy causes a context distribution shift problem, which results in poor task representations and degraded performance. This problem is exacerbated in settings with data limitations. To address this, we propose a novel approach called Meta-Normalizing Flow (Meta-NF). First, it employs a highly expressive and sample-efficient normalizing flow policy. Second, it incorporates a metric for testing-time task representation selection to effectively mitigate the context shift problem. Empirical results demonstrate that Meta-NF outperforms existing OMRL methods, with both components contributing to its strong performance.

Zitieren

APA

Liu, L., & Zhang, Z. (2026). Meta-Normalizing Flow for Data-Limited Offline Meta-Reinforcement Learning (Student Abstract). Proceedings of the AAAI Conference on Artificial Intelligence, 40(48), 41278-41280. https://doi.org/10.1609/aaai.v40i48.42241

Vancouver

Liu L, Zhang Z. Meta-Normalizing Flow for Data-Limited Offline Meta-Reinforcement Learning (Student Abstract). Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(48):41278-41280. doi:10.1609/aaai.v40i48.42241

BibTeX
@article{Liu2026,
  author = {Liu, Lianghui and Zhang, Zongzhang},
  title = {Meta-Normalizing Flow for Data-Limited Offline Meta-Reinforcement Learning (Student Abstract)},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {48},
  pages = {41278--41280},
  doi = {10.1609/aaai.v40i48.42241}
}
Kostenlos starten →