arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2407.17482cs.CYcs.AIcs.CLcs.HC
收藏

从人类反馈中强化学习:谁的文化,谁的价值观,谁的视角?

Reinforcement Learning from Human Feedback: Whose Culture, Whose Values, Whose Perspectives?

Ghent University(根特大学);Radboud University(拉德堡德大学)

Kristian González Barman, Simon Lohse, Henk de Regt

AI总结:

本文从社会认识论和多元主义哲学出发,论证RLHF中多元主义的认识论与伦理优势,并提出改进LLM开发的具体行动议程。

摘要:

我们主张在大型语言模型(LLM)背景下,从人类反馈中强化学习(RLHF)采用多元主义具有认识论和伦理上的优势。借鉴社会认识论和多元主义科学哲学,我们提出了使RLHF更能响应人类需求的方法,以及如何在此过程中应对挑战。论文最后提出了变革议程,即改善LLM开发的具体、可操作的步骤。

英文摘要:

We argue for the epistemic and ethical advantages of pluralism in Reinforcement Learning from Human Feedback (RLHF) in the context of Large Language Models (LLM). Drawing on social epistemology and pluralist philosophy of science, we suggest ways in which RHLF can be made more responsive to human needs and how we can address challenges along the way. The paper concludes with an agenda for change, i.e. concrete, actionable steps to improve LLM development.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑