T he design and complexity of a software system's user interface largely determines the ease with which users can effectively operate that system. We have designed a dynamic, seamlessly personalized adaptive user interface: it reacts to different situations and requirements, and it learns individual users' behavior patterns as well as styles. Once the AUI records the events of humancomputer interaction and discovers patterns of user behavior, it can provide just-in-time assistance by predicting a user's most likely plan and then performing part of the plan on the user's behalf. It also manipulates the software system semiautonomously, thus reducing the intervention required.
The adaptive user interfaceOur approach is based on episodes identification and association. (See the "Related Work" sidebar for information on others' research.) EIA enables the interface to recognize user action plans by tracing and analyzing a user's action sequences. Implementing this approach involves five important issues:
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.