arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.08835cs.AIcs.CYcs.LG
收藏

基于偏好多目标强化学习的学习社会价值观系统

Learning the Value Systems of Societies with Preference-based Multi-objective Reinforcement Learning

CETINIA, University Rey Juan Carlos(CETINIA皇家卡洛斯大学);Federation University(联邦大学);Deakin University(德肯大学)

Andrés Holgado-Sánchez, Peter Vamplew, Richard Dazeley, Sascha Ossowski, Holger Billhardt

AI总结:

本文提出基于偏好多目标强化学习的方法,用于在马尔可夫决策过程中学习社会价值系统,通过聚类和价值对齐模型实现不同用户群体的个性化价值偏好适应。

摘要:

具有价值意识的人工智能应能够识别人类价值观并适应不同用户的价值系统(基于价值的偏好)。这需要对价值观进行操作化,这可能容易出现规格不准确的问题。价值观的社会性要求其表示必须符合多个用户,尽管价值系统各不相同,但它们在群体间表现出一定的模式。在顺序决策中,已尝试通过不同目标或价值观的演示来实现个性化。然而,这些方法需要手动设计的特征或缺乏基于价值的可解释性和/或适应不同用户偏好能力。我们提出基于聚类和基于偏好的多目标强化学习(PbMORL)的算法,用于在马尔可夫决策过程(MDPs)中学习代理社会的价值对齐模型和价值系统。我们共同学习由社会衍生的价值对齐模型(基础)和一组代表不同用户群体(集群)的价值系统。每个集群包含一个代表其成员基于价值的偏好的价值系统以及一个近似帕累托最优的策略,该策略反映了与该价值系统一致的行为。我们在具有人类价值观的两个MDP上评估了我们的方法,与最先进的PbMORL算法和基线进行比较。

英文摘要:

Value-aware AI should recognise human values and adapt to the value systems (value-based preferences) of different users. This requires operationalization of values, which can be prone to misspecification. The social nature of values demands their representation to adhere to multiple users while value systems are diverse, yet exhibit patterns among groups. In sequential decision making, efforts have been made towards personalization for different goals or values from demonstrations of diverse agents. However, these approaches demand manually designed features or lack value-based interpretability and/or adaptability to diverse user preferences. We propose algorithms for learning models of value alignment and value systems for a society of agents in Markov Decision Processes (MDPs), based on clustering and preference-based multi-objective reinforcement learning (PbMORL). We jointly learn socially-derived value alignment models (groundings) and a set of value systems that concisely represent different groups of users (clusters) in a society. Each cluster consists of a value system representing the value-based preferences of its members and an approximately Pareto-optimal policy that reflects behaviours aligned with this value system. We evaluate our method against a state-of-the-art PbMORL algorithm and baselines on two MDPs with human values.

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。

↑