admin管理员组

文章数量:1655506

AGI之Agent:《Generative Agents: Interactive Simulacra of Human Behavior生成代理:人类行为的交互模拟》翻译与解读

目录

《Generative Agents: Interactive Simulacra of Human Behavior》翻译与解读

Figure 1: Generative agents are believable simulacra of human behavior for interactive applications. In this work, we demonstrate generative agents by populating a sandbox environment, reminiscent of The Sims, with twenty-five agents. Users can observe and intervene as agents plan their days, share news, form relationships, and coordinate group activities.图1:生成代理是交互式应用程序中可信的人类行为模拟。在这项工作中,我们通过填充沙盒环境来展示生成代理,让人想起模拟人生,有25个代理。用户可以观察和干预代理计划他们的一天、分享新闻、建立关系和协调小组活动。

Abstract

9 CONCLUSION


《Generative Agents: Interactive Simulacra of Human Behavior》翻译与解读

地址

论文地址:https://arxiv/abs/2304.03442

时间

2023年4月7日

作者

Joon Sung Park, Joseph C. O'Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang, Michael S. Bernstein

斯坦福大学

总结

这篇论文提出了一个生成型代理人(Generative Agents),用于创建可交互的人类行为模拟代理人,实现可信赖度的模拟人类行为模式。

背景痛点:过去几十年,研究人员和从业者一直追求创建可信赖的人类行为代理的目标。但是,人类行为的复杂性使得这个问题十分困难。大规模语言模型提供了一种新的视角来解决这个问题。

解决方案:论文提出了“生成代理”的概念。生成代理依靠生成模型来模拟可信赖的人类行为。它使用一个新的代理人体系结构,该体系结构包含三个主要组成部分:记忆流、反思和规划。

>> 记忆流长期记录代理人的所有经历。它根据相关性、最近性和重要性提取需要的记忆来指导行为。

>> 反思通过合成记忆来推导出更高级的推论,帮助代理人指导自己和他人的行为。

>> 规划将结论和当前环境转化为高级行动计划,并递归地详细指定行动和反应。

核心特点

>> 依靠大规模语言模型实现,充分利用模型自身模拟人类行为的能力

>> 以动态且整体的方式管理代理人不断累积和演变的记忆

>> 能够长期保持代理人行为的一致性

>> 允许代理人通过学习和反思来不断改进和演化自身行为

应用和优势

>> 论文实际实现了25个代理人在一个类似The Sims的沙盒世界中的日常生活。这些代理人能够展现出可信赖的个体行为和自动產生的社交动态,例如会传播信息、结交新朋友和协同完成任务。相比经典方法,该方案能够以更灵活和覆盖范围更广的方式实现可信赖代理人。

总之,这篇论文提出了一个新的可交互代理人体系结构,实现了利用大规模语言模型模拟长期一致且自动產生的人类社交行为,解决了传统方法难以处理的挑战。

Figure 1: Generative agents are believable simulacra of human behavior for interactive applications. In this work, we demonstrate generative agents by populating a sandbox environment, reminiscent of The Sims, with twenty-five agents. Users can observe and intervene as agents plan their days, share news, form relationships, and coordinate group activities.图1:生成代理是交互式应用程序中可信的人类行为模拟。在这项工作中,我们通过填充沙盒环境来展示生成代理,让人想起模拟人生,有25个代理。用户可以观察和干预代理计划他们的一天、分享新闻、建立关系和协调小组活动。

Abstract

Believable proxies of human behavior can empower interactive applications ranging from immersive environments to rehearsal spaces for interpersonal communication to prototyping tools. In this paper, we introduce generative agents--computational software agents that simulate believable human behavior. Generative agents wake up, cook breakfast, and head to work; artists paint, while authors write; they form opinions, notice each other, and initiate conversations; they remember and reflect on days past as they plan the next day. To enable generative agents, we describe an architecture that extends a large language model to store a complete record of the agent's experiences using natural language, synthesize those memories over time into higher-level reflections, and retrieve them dynamically to plan behavior. We instantiate generative agents to populate an interactive sandbox environment inspired by The Sims, where end users can interact with a small town of twenty five agents using natural language. In an evaluation, these generative agents produce believable individual and emergent social behaviors: for example, starting with only a single user-specified notion that one agent wants to throw a Valentine's Day party, the agents autonomously spread invitations to the party over the next two days, make new acquaintances, ask each other out on dates to the party, and coordinate to show up for the party together at the right time. We demonstrate through ablation that the components of our agent architecture--observation, planning, and reflection--each contribute critically to the believability of agent behavior. By fusing large language models with computational, interactive agents, this work introduces architectural and interaction patterns for enabling believable simulations of human behavior.

可信的人类行为代理可以增强交互式应用程序的功能,从沉浸式环境到人际交流的排练空间,再到原型工具。在本文中,我们介绍了生成代理(Generative Agents)——模拟可信人类行为的计算软件代理。生成代理起床、做早餐,然后上班;艺术家绘画,作家写作;他们形成观点,注意彼此,并开始对话;他们记得并反思过去的日子,同时计划下一天。

为了实现生成代理,我们描述了一种架构,该架构扩展了一个大型语言模型,以使用自然语言存储代理的完整经验记录,随着时间的推移将这些记忆合成为更高层次的反思,并动态检索它们以规划行为。我们实例化生成代理以填充一个受《模拟人生The Sims》启发的交互式沙盒环境,最终用户可以使用自然语言与其中的25个代理进行交互

在评估中,这些生成代理产生可信的个体紧急的社交行为:例如,从仅有一个用户指定的一个代理想要举办情人节派对的概念开始,代理们在接下来的两天里自主传播派对的邀请,结识新朋友,相互约会参加派对,并协调在适当的时间一起出现在派对上。我们通过消融实验证明了我们代理架构的组件——观察、规划和反思——各自对代理行为的可信度做出了重要贡献。通过将大型语言模型与计算机交互代理融合在一起,这项工作介绍了用于实现可信人类行为模拟的架构和交互模式

9 CONCLUSION

This paper introduces generative agents, interactive computational agents that simulate human behavior. We describe an architec-ture for generative agents that provides a mechanism for storing a comprehensive record of an agent’s experiences, deepening its understanding of itself and the environment through reflection, and retrieving a compact subset of that information to inform the agent’s actions. We then demonstrate the potential of generative agents by manifesting them as non-player characters in a Sims-style game world and simulating their lives within it. Evaluations suggest that our architecture creates believable behavior. Looking ahead, we suggest that generative agents can play roles in many interac-tive applications, ranging from design tools to social computing systems to immersive environments.

本文介绍了生成代理,即模拟人类行为的交互式计算代理。我们描述了生成代理的架构,为生成代理提供了存储代理经验全面记录的机制,通过反思加深了对自身和环境的理解,并检索该信息的紧凑子集以指导代理的行动。然后,我们通过将它们显现为《模拟人生》风格游戏世界中的非玩家角色,并在其中模拟他们的生活,展示了生成代理的潜力。评估表明我们的架构创造了可信的行为。展望未来,我们认为生成代理可以在许多交互应用中发挥作用,从设计工具到社交计算系统再到沉浸式环境,涵盖范围广泛。

本文标签: 人类GenerativeAgentsAGIAgent