Metraipolis › archive › dc_65199ed039ed

Exponential Hardness of Off-Policy Evaluation under History-Dependent Logging — arXiv:2609.19135

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.19135
License (as archived)
CC0-1.0
Archived
2026-10-15T15:05:02.996Z
SHA-256 of the archived text
2ca1e939638dd085d0958dd4b0098a75d93d76c4dc3d4d4babc7fe9eae62a154
Stable URL of this page
https://metraipolis.com/documento/dc_65199ed039ed

How to cite

Exponential Hardness of Off-Policy Evaluation under History-Dependent Logging — arXiv:2609.19135 — arXiv (en), https://arxiv.org/abs/2609.19135 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_65199ed039ed (Metraipolis archive dc_65199ed039ed, sha256 2ca1e939638dd085d0958dd4b0098a75d93d76c4dc3d4d4babc7fe9eae62a154)
BibTeX
@misc{dc65199ed039ed,
  title = {Exponential Hardness of Off-Policy Evaluation under History-Dependent Logging — arXiv:2609.19135},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.19135},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_65199ed039ed; sha256 2ca1e939638dd085d0958dd4b0098a75d93d76c4dc3d4d4babc7fe9eae62a154},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.19135. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Can a logged dataset visit every hidden state frequently and still be exponentially uninformative about a target policy's value? We show that it can when the logger depends on history. For every horizon $H \ge 3$, we construct two POMDPs with at most two latent states per stage, three actions, and a common logger with three memory states. Action coverage, belief coverage, and two behavior-marginal outcome-revealing conditions all have constants independent of $H$. Nevertheless, evaluating a known deterministic target policy to accuracy $1/8$ requires $\Theta((3/2)^H \log(1/\delta))$ logged episodes at confidence $1-\delta$, for $0 < \delta \le 1/4$, even when both candidate models are known. The mechanism is simple: a reset erases the unknown transition that determines the target value. We characterize the resulting statistical experiment exactly and obtain a matching optimal estimator. A directed two-lane gridworld realizes the construction, and trajectory simulations agree with its finite-sample prediction. The result establishes intractability for the history-dependent-logging, model-based case posed by Zhang and Jiang (2025, arXiv:2503.01134 ), under their behavior-marginal definition of revealing.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-17 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.19135 — 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 2026-10-15T14:08:19.183Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).