Personalized Help for Optimizing Low-Skilled Users' Strategy

2024-11-14

arXiv.org

AI Summary

This paper addresses the challenge that, despite AI’s superhuman performance in strategic games, it remains ineffective at assisting low-skill human players. To bridge this gap, we propose an intent-driven, dual-modal (action + diplomatic message) personalized assistance framework built upon CICERO. Methodologically, it integrates natural language inference with multi-agent policy modeling to enable player intent recognition, context-aware recommendation generation, and human-AI collaborative evaluation. Our contributions are threefold: (1) the first empirical demonstration that mere *existence* of advice—regardless of its content—significantly improves human performance; (2) a non-intrusive, real-time adaptive guidance paradigm; and (3) a principled approach reconciling AI’s superhuman capabilities with tangible human benefit. Experiments show that novice players receiving such assistance achieve substantially higher win rates and decision quality—matching or even surpassing those of experienced players.

Technology Category

• Humans and AI: Planning and Decision Support for Human-Machine Teams
• Multiagent Systems: Adversarial Agents
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

AIs can beat humans in game environments; however, how helpful those agents are to human remains understudied. We augment CICERO, a natural language agent that demonstrates superhuman performance in Diplomacy, to generate both move and message advice based on player intentions. A dozen Diplomacy games with novice and experienced players, with varying advice settings, show that some of the generated advice is beneficial. It helps novices compete with experienced players and in some instances even surpass them. The mere presence of advice can be advantageous, even if players do not follow it.

Problem

• AI's impact on human game strategies
• Personalized advice for novice players
• Effectiveness of AI-generated guidance in Diplomacy

Innovation

• AI generates move advice
• AI provides message guidance
• Advice enhances novice performance