Preference-Guided Multi-Objective UI Adaptation
summary
A system infers ranked user preferences from a few manual element moves (smoothed with a triangular moving average), maps ranks to high/mid/low priority buckets, and runs a Pareto lexicographic MOO that focuses search on higher priorities. Built for MR with 7 UI elements and 5 objective functions, it auto-selects the Pareto solution closest to the user’s adjustment. A user study shows fewer element moves, smaller objective-distance to manual edits, and ratings that most often rank this method first or second.
AI-written summary · about
chapters (8)
- 0:00Introduction to talk
- 0:24Background and challenges
- 1:30Pareto frontier in UI layout
- 3:35Research questions addressed
- 4:59Preference-guided optimization method
- 7:46User study and results
- 10:53Discussion and future work
- 11:56Q&A session
related talks
17mContext-Aware Online Adaptation of Mixed Reality Interfaces
35mApplied Psychology: Psychology based UI improvements
47mOver-Complicated? Over-Simplified? The UX Efficient Frontier
11mThe invisible craft of great UX
39mUnderstanding the Mysterious Behavior of Complex Digital Product Users
38mThe Psychology of UX
11mEvaluating the Usability of Privacy Choice Mechanisms
37mMind-Reading with adaptive UIs