The paper investigates the creation of multi-agent strategy explanations to improve human-robot collaboration. It emphasizes the increasing importance of robots’ ability to collaborate fluently with humans, especially in social, home, and manufacturing settings. The study explores how this can be achieved through a blend of proactive and reactive processes that communicate coordination strategies beforehand, but adapt to human strategies as they evolve. The authors propose a method for generating visual strategy explanations for human-AI collaboration in games with multiple Nash equilibria and demonstrate its effectiveness through a user study.
Publication date: 21 Nov 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2311.11955
