Lyfe Agents: Generative agents for low-cost real-time social interactions

Zhao Kaiya,Michelangelo Naim,Jovana Kondic,Manuel Cortes,Jiaxin Ge,Shuying Luo,Guangyu Robert Yang,Andrew Ahn

2023-10-04

Abstract:Highly autonomous generative agents powered by large language models promise to simulate intricate social behaviors in virtual societies. However, achieving real-time interactions with humans at a low computational cost remains challenging. Here, we introduce Lyfe Agents. They combine low-cost with real-time responsiveness, all while remaining intelligent and goal-oriented. Key innovations include: (1) an option-action framework, reducing the cost of high-level decisions; (2) asynchronous self-monitoring for better self-consistency; and (3) a Summarize-and-Forget memory mechanism, prioritizing critical memory items at a low cost. We evaluate Lyfe Agents' self-motivation and sociability across several multi-agent scenarios in our custom LyfeGame 3D virtual environment platform. When equipped with our brain-inspired techniques, Lyfe Agents can exhibit human-like self-motivated social reasoning. For example, the agents can solve a crime (a murder mystery) through autonomous collaboration and information exchange. Meanwhile, our techniques enabled Lyfe Agents to operate at a computational cost 10-100 times lower than existing alternatives. Our findings underscore the transformative potential of autonomous generative agents to enrich human social experiences in virtual worlds.

Human-Computer Interaction,Artificial Intelligence,Machine Learning

What problem does this paper attempt to address?

The paper attempts to address the problem of how to create a low-cost, real-time interactive generative agent. Existing generative models, especially large language models (LLMs), although excellent at mimicking human behavior, still face challenges in maintaining low cost while achieving real-time interaction with humans. Specifically, the paper raises the following key issues:

1. **High computational cost**: Existing autonomous agents often rely on expensive large language models for decision-making, leading to high computational costs.
2. **Real-time responsiveness**: How to achieve real-time responsiveness and autonomy of the agent while keeping costs low.
3. **Adaptability and consistency**: How to maintain the agent's adaptability and consistency in multi-agent environments, especially in complex social scenarios.

To address these issues, the paper introduces Lyfe Agents, a type of autonomous generative agent that combines low cost, real-time responsiveness, and intelligence. The paper achieves these goals through the following innovative methods:

1. **Option-Action Framework**: Reduces the cost of high-level decision-making by lowering computational burden through hierarchical action selection.
2. **Asynchronous Self-Monitoring**: Enhances the agent's consistency and situational awareness by maintaining summaries of recent events.
3. **Summarize-Forget Memory Mechanism**: Prioritizes the storage of key memory items to manage memory efficiently at low cost.

The paper evaluates the self-motivation and sociality of Lyfe Agents in a custom 3D virtual environment platform called LyfeGame, demonstrating their performance in multiple multi-agent scenarios, including solving a murder case, choosing extracurricular activity clubs, and helping patients. These experimental results show that Lyfe Agents not only exhibit human-like reasoning and decision-making abilities in complex social interactions but also significantly outperform existing methods in terms of computational cost.