With the rise of different electronic devices such as smartphones, tablets and smartwatches, and the emergence of "smart things" forming part of the Internet of Things (IoT), user interfaces (UIs) have to be adapted in order to cope with the different input and output methods as well as device characteristics. In addition to the challenges dealing with these rapidly changing technologies, user interface designers struggle to adapt their UIs to evolving user needs and preferences, resulting in bad user experience. Therefore, the field of end-user development (EUD) strives to make systems easy to develop rather than only easy to use. However, the involvement of end users in the UI design process represents a major challenge, since end users do normally not have the same skills as professional designers. The main question is how to support end users during the UI design process. In order to answer this question, we investigate how to recommend end users alternative and potentially better designs by extending the eSPACE end-user authoring tool. In order to further optimise the recommended alternative user interface designs, we apply some machine learning techniques. The resulting human-AI interaction during the UI design process should lead to better user interfaces and we investigate the potential of human-AI interaction in end-user user interface development based on extensive usability and user experience studies.
Fig. 1: Alternative design recommendation and selection process
We plan to investigate the potential of human-AI interaction in end-user user interface development. While there exist various AI-infused solutions for the semi- or fully-automatic generation of UIs, not many focus on keeping the designer fully in control by only providing recommendations during their design task. Moreover, to the best of our knowledge, there exist no AI-infused authoring solutions to support end users during the design of user interfaces that ultimately help them to get better in designing UIs over time. Our main goal is therefore to use AI algorithms and a set of UID guidelines in an end-user user interface authoring tool in order to help end users during their UI design process and improve their design skills over time.
Towards User-defined Cross-Device Interaction,
Audrey Sanctorum and Beat Signer,
Proceedings of DUI 2016, 5th Workshop on Distributed User Interfaces,
Lugano, Switzerland, June 2016
