Task Knowledge

Task knowledge research focuses on how to effectively represent, transfer, and utilize knowledge about specific tasks within intelligent systems, particularly in the context of multi-agent systems and human-AI collaboration. Current research explores methods for integrating modular expertise from multiple models, leveraging attention mechanisms to focus on crucial task aspects, and employing techniques like automated excuse generation to guide human demonstrations and optimize knowledge transfer. This work is significant for improving the efficiency and effectiveness of training AI agents, enhancing human-AI teamwork, and mitigating the loss of expert knowledge in various domains.