Towards Effective Human-AI Collaboration in GUI-Based Interactive Task Learning Agents
This addresses the problem of making interactive task learning more usable and useful for users, but it is incremental as it builds on past work without presenting new results.
The paper tackles the challenge of enabling effective human-AI collaboration in GUI-based interactive task learning agents, reflecting on 5 years of experience with the SUGILITE system to discuss issues and lessons learned.
We argue that a key challenge in enabling usable and useful interactive task learning for intelligent agents is to facilitate effective Human-AI collaboration. We reflect on our past 5 years of efforts on designing, developing and studying the SUGILITE system, discuss the issues on incorporating recent advances in AI with HCI principles in mixed-initiative interactions and multi-modal interactions, and summarize the lessons we learned. Lastly, we identify several challenges and opportunities, and describe our ongoing work