Hierarchical Belief Modeling for Zero-Shot Opponent Adaptation in Partially Observable Multi-Agent Navigation

2026-09-11

AI Summary

研究解决了部分可观测多智能体导航中的零样本对手适应问题，通过HORIZON模型结合多种技术，实现了对隐藏游戏参数和对手风格的推断。

Abstract

Lux AI Season 3 requires agents to act under partial observability, randomized episode level dynamics, and a best of five match structure that rewards both tactical execution and fast adaptation. We present HORIZON, a hierarchical agent that combines symmetry aware spatial perception, dual memory belief tracking, relic centric graph attention, information gain driven exploration, and an opponent conditioned policy mixture. HORIZON separates short horizon control from cross match meta reasoning, while auxiliary belief and world model objectives stabilize learning. Trained with PPO in a large scale JAX simulator, the resulting agent explicitly infers hidden game parameters and opponent style. Experiments show consistent gains in match win rate, episode win rate, adaptation gain, and league rating over strong recurrent and feed forward baselines.

Problem

• partial observability
• opponent adaptation
• multi-agent navigation

Innovation

• Hierarchical Agent
• Dual Memory Belief Tracking
• Graph Attention
• Information Gain Driven Exploration
• Opponent Conditioned Policy Mixture