Multi-Armed Bandits with Fairness Constraints for Distributing Resources to Human Teammates
Multi-Armed Bandits with Fairness Constraints for Distributing Resources to Human Teammates
复制标题
具有公平约束的多臂强盗将资源分配给人类队友
DOI：
10.1145/3319502.3374806
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
S. Nikolaidis
中科院分区：
文献类型：
--
作者：
Houston Claure;Yifang Chen;Jignesh Modi;Malte F. Jung;S. Nikolaidis
一个与多人合作的机器人应该如何决定资源的分配（例如，社会关注，或组装所需的零件）？人们对资源的分配方式有着独特的适应性。将更多的资源分配给一个团队成员可能会被认为是不公平的，对信任有潜在的有害影响。我们引入了一种具有公平性约束的多臂强盗算法，其中机器人将资源分配给不同技能水平的人类队友。在这个问题中，机器人不知道每个人类队友的技能水平，而是通过观察他们的表现来学习。我们将公平性定义为在整个任务中选择每个人类队友的最小比率的约束。我们提供了性能的理论保证，并进行了大规模的用户研究，在那里我们调整了算法中的公平水平。结果表明，资源分配的公平性对用户对系统的信任有显著影响。ACM参考格式：Houston Claure， Yifang Chen, Jignesh Modi， Malte Jung和Stefanos Nikolaidis。2020。具有公平约束的多武装强盗向人类队友分配资源。2020ACM/IEEE人机交互国际会议论文集（HRI ' 20）， 2020年3月23-26日，英国剑桥。ACM，纽约，美国，10页。https://doi.org/10.1145/3319502.3374806
How should a robot that collaborates with multiple people decide upon the distribution of resources (e.g. social attention, or parts needed for an assembly)? People are uniquely attuned to how resources are distributed. A decision to distribute more resources to one team member than another might be perceived as unfair with potentially detrimental effects for trust. We introduce a multi-armed bandit algorithm with fairness constraints, where a robot distributes resources to human teammates of different skill levels. In this problem, the robot does not know the skill level of each human teammate, but learns it by observing their performance over time. We define fairness as a constraint on the minimum rate that each human teammate is selected throughout the task. We provide theoretical guarantees on performance and perform a large-scale user study, where we adjust the level of fairness in our algorithm. Results show that fairness in resource distribution has a significant effect on users’ trust in the system. ACM Reference Format: Houston Claure, Yifang Chen, Jignesh Modi, Malte Jung, and Stefanos Nikolaidis. 2020. Multi-Armed Bandits with Fairness Constraints for Distributing Resources to Human Teammates. In Proceedings of the 2020ACM/IEEE International Conference on Human-Robot Interaction (HRI’20), March 23-26, 2020, Cambridge, United Kingdom. ACM, New York, NY, USA, 10 pages. https://doi.org/10.1145/3319502.3374806
多臂强盗中的人类人工智能学习表现
DOI：
10.1145/3306618.3314245
发表时间：
2019
期刊：
Ethics and Society (AIES
影响因子：
--
作者：
Pandya, Ravi;Huang, Sandy H.;Hadfield-Menell, Dylan;Dragan, Anca D.
通讯作者：
Dragan, Anca D.
DOI：
10.23919/acc.2018.8431265
发表时间：
2018-02
期刊：
2018 Annual American Control Conference (ACC)
影响因子：
--
作者：
Lai Wei;Vaibhav Srivastava
通讯作者：
Lai Wei;Vaibhav Srivastava
DOI：
10.1109/hri.2019.8673234
发表时间：
2019-01
期刊：
2019 14th ACM/IEEE International Conference on Human-Robot Interaction (HRI)
影响因子：
--
作者：
Lawrence Chan;Dylan Hadfield-Menell;S. Srinivasa;A. Dragan
通讯作者：
Lawrence Chan;Dylan Hadfield-Menell;S. Srinivasa;A. Dragan