Texas A&M University hosted the Association for Computing Machinery’s (ACM) 2021 Conference on Intelligent User Interfaces (IUI) virtually from April 13-17.
This annual conference serves as a premier international forum for reporting outstanding research and development on IUIs. ACM IUI is where the human-computer interaction (HCI) community meets the artificial intelligence (AI) community.
The sponsors for this year’s conference were Texas A&M, the Texas A&M College of Engineering, the Institute for Engineering Education and Innovation (IEEI), Sketch Recognition Lab, Google, Microsoft, Multimodal Technologies and Interaction, and Science Robotics.
Intelligent user interfaces for AI technologies are becoming increasingly important with the greater adoption of machine learning and other AI applications. Assimilating AI into user interfaces and merging refined user-centered design with the representation, perception and recognition that AI has to offer makes the future of AI seem brighter as well.
The IUI 2021 general chair was Dr. Tracy Hammond, director of the IEEI and professor in the Department of Computer Science and Engineering at Texas A&M. Dr. Katrien Verbert, KU Leuven and Dr. Dennis Parra from the Pontifical Catholic University of Chile, served as the co-chairs.
“All of our planning committee members did a stellar job in making sure the conference was both scientifically productive and an enjoyable time for everyone,” said Hammond.
This year’s conference was attended by about 545 students, researchers, industry professionals and academics from at least 32 countries. There were guided yoga lessons every morning, before the conference’s main sessions, and online games in the evening. This year’s poster and demonstration session was hosted on the Gather virtual platform, a video-calling space that uses a virtual space reminiscent of eight-bit arcade games to allow participants to explore interactive environments. This was just one of the ways the organizers tried to make the virtual conference feel more welcoming. Despite the big numbers, it was easy enough for participants to meet and talk to each other through the several Zoom breakout rooms and social/fun sessions.
There were four keynote speakers, including:
- Meredith Morris, director of the People + AI Research Google research team, whose talk was titled “Accessibility as an Opportunity and Challenge for Intelligent User Interfaces.”
- Juan Gilbert, computer and information science and engineering department chair at the University of Florida, whose talk was titled “Equitable AI: Using AI to Achieve Diversity in Admissions.”
- Andrea G. Parker, associate professor at Georgia Tech, whose talk was titled “Achieving Health Equity: The Power & Pitfalls of Intelligent Interfaces.”
- Kwan-Liu Ma, Distinguished Professor at the University of California, Davis, whose talk was titled “Intelligent Visualization Interfaces.”
Each year, the IUI organizing committee acknowledges one past IUI paper that had the most impact and highest visibility since being published. This year, the committee selected "Principles of Explanatory Debugging to Personalize Interactive Machine Learning’’ by Todd Kulesza, Margaret Burnett, Wengkeen Wong and Simone Stumpf, for the best impact paper award.
This year there were 241 paper submissions, and 64 papers were accepted with a 26.6% acceptance rate. The best paper award went to “ProtoAI: Model-Informed Prototyping for AI-Powered Interfaces” by Eytan Adar, Colleen M. Seifert and Hariharan Subramonyam from the University of Michigan. The nine posters, 12 demonstrations and nine student consortium posters were presented on the space station themed Gather interactive meeting space.
A diversity and inclusion panel spoke about what the IUI community could do to improve inclusivity and diversity, and the IUI Town Hall consisted of detailed suggestions and ideas on how to move forward.
The first day of the conference consisted of tutorials and workshops, including a three-hour tutorial on ‘’Human-Centered AI: Reliable, Safe and Trustworthy’’ by Ben Schneiderman that was live streamed with over 1,000 attendees. The tutorial proposed a new synthesis method in which AI algorithms are combined with human-centered thinking to make human-centered AI. The other tutorials were “ModelGenGUIs – High-Level Interaction Design with Discourse Models for Automated GUI Generation,” “Conversational Recommendation Systems” and “VisRec: A Hands-on Tutorial on Deep Learning for Visual Recommender Systems.”
Seven workshops were available for free to all attendees, including: “Theoretical and Methodological Challenges in Intelligent Conversational User Interface Interactions,” “Fourth Workshop on Exploratory Search and Interactive Data Analytics,” “Second Workshop on Human-AI Co-Creation with Generative Models,” “Healthy Interfaces,” “Transparency and Explainability in Adaptive Systems through User Modeling Grounded in Psychological Theory,” “SOcial and Cultural IntegrAtion with PersonaLIZEd Interfaces (SOCIALIZE)” and “Transparency and Explanations in Smart Systems.”
IUI 2022 is scheduled to take place in Helsinki, Finland, from March 23-26, 2022.