Multi-Agent Strategy Explanations for Human-Robot Collaboration

Multi-agent explanations improve human-robot collaboration efficiency: The paper's user study found that when people received explanations of a robot's collaborative strategy, they were able to "collaborate more efficiently with new robot partners," suggesting that understanding how robots will act leads to better real-world teamwork outcomes.

Visual and textual explanations help users understand robot decision-making: The framework generates explanations through two complementary methods—visual representations of key decision points (called "landmark states") and text descriptions via language models—enabling people to grasp robot strategies through multiple formats rather than relying on a single explanation type.

Current AI explainability research has a gap in multi-agent scenarios: The paper identifies that "much prior work in explainable AI and RL focuses on generating explanations for single-agent policies, but little has been explored in generating explanations for collaborative policies," highlighting an important blind spot in responsible AI as robots increasingly work alongside humans.

Users can better explore different robot strategies with proper explanations: The study found that explanations helped users "better explore the full space of strategies," meaning people weren't locked into one understanding of how a robot would behave and could anticipate different possible robot actions—critical for safe human-robot interaction.