Metraipolis › archive › dc_95752f142ccd

Imagining Goals with Geometric Horizon Models for Policy Generalization — arXiv:2610.19129

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.19129
License (as archived): CC0-1.0
Archived: 2026-10-15T14:15:06.460Z
SHA-256 of the archived text: 59d20f83a1dc23a74c7cea68080e012914b658dd2c7eebcf5bde0ff496e9d2ab
Stable URL of this page: https://metraipolis.com/documento/dc_95752f142ccd

How to cite

Imagining Goals with Geometric Horizon Models for Policy Generalization — arXiv:2610.19129 — arXiv (en), https://arxiv.org/abs/2610.19129 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_95752f142ccd (Metraipolis archive dc_95752f142ccd, sha256 59d20f83a1dc23a74c7cea68080e012914b658dd2c7eebcf5bde0ff496e9d2ab)

BibTeX

@misc{dc95752f142ccd,
  title = {Imagining Goals with Geometric Horizon Models for Policy Generalization — arXiv:2610.19129},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19129},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_95752f142ccd; sha256 59d20f83a1dc23a74c7cea68080e012914b658dd2c7eebcf5bde0ff496e9d2ab},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19129. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Training from offline data has allowed for substantial progress in domains such as robotics, leading to general-purpose policies that can be easily applied zero-shot or efficiently finetuned for downstream tasks. However, these policies can still have poor generalization, both due to the choice of modeling objective and from learning from a static dataset. In this work, we focus on the challenging task of zero-shot goal generalization, where a policy is evaluated on unseen tasks that require combining its existing knowledge (compositional generalization). An avenue for improving a policy's generalization is by generating new experience with world models; however, such generation has proven difficult for longer horizons. Thus, to alleviate this issue, we propose TD-Aug, which samples from a geometric horizon model and allows for directly imagining novel outcomes that can be achieved by composing existing knowledge. We demonstrate that training on these future outcomes as goals for goal-conditioned BC policies significantly improves generalization in stitching-based OGBench tasks.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.19129 — 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 2026-10-15T13:53:35.531Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).