Metraipolis › archive › dc_ac562a2fcddf

Training and Evaluating Ethical Reinforcement Learning Agents on Per-Episode Distributions — arXiv:2608.14642

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.14642
License (as archived): CC0-1.0
Archived: 2026-10-18T19:39:43.018Z
SHA-256 of the archived text: bf552ae13253a601f73d72eb69e279bf8e0cbba60ba9df2499ab2df6821751d3
Stable URL of this page: https://metraipolis.com/documento/dc_ac562a2fcddf

How to cite

Training and Evaluating Ethical Reinforcement Learning Agents on Per-Episode Distributions — arXiv:2608.14642 — arXiv (en), https://arxiv.org/abs/2608.14642 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ac562a2fcddf (Metraipolis archive dc_ac562a2fcddf, sha256 bf552ae13253a601f73d72eb69e279bf8e0cbba60ba9df2499ab2df6821751d3)

BibTeX

@misc{dcac562a2fcddf,
  title = {Training and Evaluating Ethical Reinforcement Learning Agents on Per-Episode Distributions — arXiv:2608.14642},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.14642},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ac562a2fcddf; sha256 bf552ae13253a601f73d72eb69e279bf8e0cbba60ba9df2499ab2df6821751d3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.14642. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Reinforcement Learning (RL) agents trained on a single reward signal exploit the gap between the designed reward and the intended behavior. This is particularly a problem when we are trying to imbue ethical behavior into RL agents. An agent can look ethical on average while concentrating its violations in a few bad episodes, and a creature in the environment harmed in one episode is not restored by good conduct in another. We compare four ways of training ethical behavior in Craftax, an open-ended survival benchmark. The four are: scalar penalties with termination, a linear multi-objective weight sweep, an adaptive Lagrangian constraint, and a non-compensatory utility optimized per episode under the Expected Scalarized Returns (ESR) criterion. All are evaluated under a single detector-based protocol that counts every violation in every episode without censoring. On the frontier of mean return against mean violation rate, the four methods are indistinguishable; per episode they separate sharply. At matched mean return, the ESR agent holds its stated budget of one violation in effectively every episode (worst-decile 1.04 +/- 0.07 violations), the Lagrangian leaks past the same budget (1.14 +/- 0.03), and the weight sweep's worst episodes double it (2.20 +/- 0.20). An observation-augmentation control attributes the separation to the training objective rather than to what the agent observes, and the per-episode guarantee costs nothing on the mean frontier. When ethical violations do not average away across episodes, we argue both training and evaluation must target the per-episode distribution rather than the mean.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-18 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.14642 — 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 2026-10-18T19:02:44.025Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).