[论文解读] Generative Agents: Interactive Simulacra of Human Behavior

Joon-Sung Park, Joseph C. O'Brien+4 more|arXiv (Cornell University)|Apr 7, 2023

Artificial Intelligence in Games被引用 163

一句话总结

{

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.

研究动机与目标

• 激发为交互系统和虚拟环境创建可信的人类行为代理。
• 提出一种将大型语言模型扩展为具备长期记忆和推理能力的架构，以实现连贯、持续演化的行为。
• 展示在一个小镇沙盒中涌现的社会动态与日常生活例程。
• 评估记忆、反思和规划组件对可信度的因果影响。

提出的方法

• 将 memory stream 引入，作为代理经历的综合、自然语言的长期记忆。
• 开发一个记忆检索模型，在相关性、时效性和重要性之间取得平衡，以指导行为。
• 增加反思，将记忆综合为关于自身和他人的更高层次推断。
• 实现规划，将反思与当前情境转化为高层计划和具体行动。
• 将记忆、反思和规划与大型语言模型整合，以驱动可信的行为。
• 在一个 25-agent 沙盒（Smallville）中演示代理如何通过自然语言记忆、计划、行动和互动。

实验结果

研究问题

• RQ1 在开放世界仿真中，代理架构如何维持长期的一致性和记忆？
• RQ2 记忆、反思和规划组件对可信度和行为质量的因果影响是什么？
• RQ3 由小型语言模型驱动的代理群体能否随着时间产生涌现的社会动态和协同行动？
• RQ4 在交互式环境中使用 LLMs 实现可信代理行为时的局限性和失效模式有哪些？

主要发现

• Generative agents 在开放世界沙盒中产生可信的个体行为和涌现的社会动态。
• 如消融实验所示，记忆检索、反思和规划中的每一项都对可信度有关键贡献。
• 代理能够从最小的初始提示（例如计划一个聚会）出发，协调完成复杂任务，而无需逐一手工编码所有行为。
• 端到端评估揭示了代理社会中涌现的现象，如信息扩散和关系记忆。
• 该架构揭示了如记忆检索错误和过于正式语言倾向等失效模式，为未来改进提供参考。