Virtual Embodied Agents
This chapter examines how agents interact within virtual environments — from generative agent architectures to virtual world simulation, NPC behavior evolution, and social behavior emergence.
Contents:
- Overview of Virtual Embodied Agents — Concepts, classification, and research challenges of virtual embodied agents
- Generative Agents Introduction — Generative Agents architecture and social simulation
- Memory Streams and Reflection Mechanisms — Memory Stream architecture and tri-factor retrieval scoring
- Virtual World Simulation Engines — Operating environments from 2D grid worlds to 3D physics simulations
- NPC Behavior Evolution — From state machines to LLM-driven NPC behavior control
- Game AI Frontiers — Breakthroughs at the intersection of large language models and game AI
- Social Behavior Emergence — Emergent social behaviors in virtual agent societies
- Digital Twins and the Metaverse — AI agent applications in digital twins and the metaverse