From: ICON-faculty-list <icon-faculty-list-bounces@ecn.purdue.edu<mailto:icon-faculty-list-bounces@ecn.purdue.edu>> on behalf of Nak-seung Patrick Hyun <nhyun@purdue.edu<mailto:nhyun@purdue.edu>> Date: Wednesday, October 16, 2024 at 17:53 To: ICON-faculty-list@ecn.purdue.edu<mailto:ICON-faculty-list@ecn.purdue.edu> <icon-faculty-list@ecn.purdue.edu<mailto:icon-faculty-list@ecn.purdue.edu>>, icon-students-list@ecn.purdue.edu<mailto:icon-students-list@ecn.purdue.edu> <icon-students-list@ecn.purdue.edu<mailto:icon-students-list@ecn.purdue.edu>>, icon-friends@ecn.purdue.edu<mailto:icon-friends@ecn.purdue.edu> <icon-friends@ecn.purdue.edu<mailto:icon-friends@ecn.purdue.edu>> Subject: [ICON] [ICON Seminar] This Friday 3:00 pm-4:00 pm (MSEE 112): Dr. Ehsan Moradi-Pari (Honda Research Institute) Hello ICON faculty, students, and friends, This week, we will have our second industry ICON seminar on Friday (10/18), presented by Dr. Ehsan Moradi-Pari, a division director, and Network Intelligence chief scientist at Honda Research Institute USA, Inc. For students who are interested in internship and full-time positions at Honda Research Institutde should come to the talk and see any opportunities! The flyer has been attached to this email. Title: "Mastering Human-AI Teamwork: Dynamic Role Assignment and Knowledge Representation for Adaptive Multi-agent Cooperation" Speaker: Ehsan Moradi-Pari (division director, and Network Intelligence chief scientist at Honda Research Institute USA, Inc.) Date: 10/18 (Friday) 3-4 pm (MSEE 112) Abstract: Teamwork presents the challenge of designing agents capable of seamlessly collaborating with others, even in the absence of prior coordination or established protocols. Effective teamwork relies on key concepts such as decentralization, role assignment, adaptability, and ad-hoc cooperation. When bringing humans into the fold, it becomes crucial not only to adapt AI systems to human behaviors and preferences but also to implement communication mechanisms that ensure team knowledge is conveyed in a manner easily understood by humans. The creation of cooperative AI agents also involves recognizing the human sense of agency, making it necessary for agents to adjust dynamically to human teammates. In this talk, we will introduce our view of teamwork, with a focus on role assignment, human-interpretable knowledge representation, and decentralized decision-making. These elements are essential for fully realizing the potential of human-AI teaming and ensuring that both humans and AI agents can effectively collaborate toward shared goals. Best, ICON seminar organizers Ziran Wang, Yunyue Li, Nak-seung Patrick Hyun -- Nak-seung Patrick Hyun Assistant Professor Elmore Family School of Electrical and Computer Engineering Purdue University nhyun@purdue.edu<mailto:nhyun@purdue.edu> (765)496-0725 Website: nphyun.com<https://nam04.safelinks.protection.outlook.com/?url=http%3A%2F%2Fnphyun.com%2F&data=05%7C02%7Ccce-graduates-list%40ecn.purdue.edu%7Cfec83b8c8d86408cf98a08dceeb65af4%7C4130bd397c53419cb1e58758d6d63f21%7C0%7C0%7C638647714494670981%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=HM%2BBtXUhpw5YQPAP33VK0qPKFnIjafdthXeKa%2BbTaOw%3D&reserved=0>