We present our approach to the problem of how an agent, within an economic Multi-Agent System, can determine when it should behave strategically (i.e. learn and use models of other agents), and when it should act as a simple price-taker. We provide a framework for the incremental implementation of modeling capabilities in agents, and a description of the forms of knowledge required. The agents were implemented and different populations simulated in order to learn more about their behavior and the merits of using and learning agent models. Our results show, among other lessons, how savvy buyers can avoid being ``cheated'' by sellers, how price volatility can be used to quantitatively predict the benefits of deeper models, and how specific types of agent populations influence system behavior.
Efficient Multi-agent Epistemic Planning: Teaching Planners About Nested Belief
School of Computing Queen’s University(女王大学计算机学院);University of Edinburgh(爱丁堡大学);Alan Turing Institute(艾伦图灵研究所);Faculty of Computer Science University of Bozen-Bolzano(博岑-博尔扎诺大学计算机学院);Department of Computer Science University of Toronto(多伦多大学计算机科学系);School of Computing and Information Systems University of Melbourne(墨尔本大学计算与信息系统学院)
Neural Amortized Inference for Nested Multi-agent Reasoning
Dartmouth College(达特茅斯学院);Google Research(谷歌研究院);New York University(纽约大学);University of Virginia(弗吉尼亚大学);Massachusetts Institute of Technology(麻省理工学院);Johns Hopkins University(约翰斯·霍普金斯大学)
A Dynamic Agent Based Model of the Real Economy with Monopolistic Competition, Perfect Product Differentiation, Heterogeneous Agents, Increasing Returns to Scale and Trade in Disequilibrium
Prompt Injection Mitigation with Agentic AI, Nested Learning, and AI Sustainability via Semantic Caching
Head of AI, Tesisquare Member, Open Voice Interoperability Initiative Linux Foundation AI & Data Torino, Italy(AI负责人,Tesisquare成员,开放语音互操作性倡议Linux基金会AI与数据部门,意大利托里诺);Principal, Conversational Technologies Member, Open Voice Interoperability Initiative Linux Foundation AI & Data Plymouth Meeting, PA, USA(首席科学家,对话技术成员,开放语音互操作性倡议Linux基金会AI与数据部门,美国普利茅斯会议)
Simulating the Ridesharing Economy: The Individual Agent Metro-Washington Area Ridesharing Model
拼车经济正在经历快速增长和创新。Uber和Lyft等公司继续以相当快的速度增长,同时提供他们的平台作为拼车服务的组织媒介,增加消费者效用,并雇用数千名兼职人员。然而,在拼车服务的建模中仍然存在许多挑战,其中许多挑战目前尚未得到广泛考虑。在本文中,开发了一个基于代理的模型来模拟华盛顿特区大都市区的拼车服务。该模型用于检查通用拼车服务的乘客(客户)和司机(服务提供商)的效用水平。提供了个体代理Metro-Washington Area Ridesharing Model(IAMWARM)的描述,以及典型模拟运行的描述。我们调查了两种情况下,两种空间运动行为的司机的经济收益为24小时内。这两种空间行为是随机运动和Voronoi运动,我们描述。这两种运动行为进行了测试,在一个固定的运行条件的情况下,和一个可变的运行条件的情况下。我们发现,Voronoi运动增加了司机的效用获得,但这种系统属性的出现,只有在可变的情况下是可行的。这一结果提供了两个重要的见解:第一个是,在乘客接送之前,司机的移动决定会影响服务和司机的经济收益,从而影响乘客的成功接送率。第二个是,这种现象是唯一明显的实验条件下,乘客和司机的到达率的变化管理。
Multi-Agent AI as a Nested Principal-Agent Problem in Private Wealth Management: Mandate Representation and Evidence Control in Switzerland, Germany and Austria
Swissi Institute for AI(Swissi人工智能研究所);Hochschule für Wirtschaft und Umwelt Nürtingen-Geislingen(叙廷根-盖斯林根经济与环境应用科学大学);Faculty of Information Technology and Engineering (FITI), Belgrade(贝尔格莱德信息技术与工程学院)
NDRL: Cotton Irrigation and Nitrogen Application with Nested Dual-Agent Reinforcement Learning
School of Computer Science and Technology, Xinjiang University(新疆大学计算机科学与技术学院);Xinjiang Multimodal Information Technology Engineering Research Center(新疆多模态信息处理工程技术研究中心);Xinjiang Key Laboratory of Signal Detection and Processing(新疆信号检测与处理重点实验室);Department of Electronic Engineering, Tsinghua University(清华大学电子工程系)