Langley: Machine Learning for Adaptive User Interfaces
by Carson Reynolds
Langley argues for the application of machine learning to solve the problems of automatic interface personalization. The author begins by arguing for the need for automatic personalization. A definition of machine learning is then provided, along with a brief description of some of the results surrounding induction. Following this we find a definition for Adaptive User Interfaces:
“An adaptive user interface is an interactive software system that improves its ability to interact with a user based on partial experience with that user.”
Langley discusses the development of advisory systems and the need to prefer online learning. He also suggests that the machine learning needs to be rapid. He then splits adaptive user interfaces into “informative” interfaces (filters) and “generative” interfaces. He then provides representative instances of informative and generative interfaces.
The remainder of the paper discusses ongoing work to make adaptive systems, mostly in the context of the Daimler-Benz research and technology center. The paper concludes that there are many open issues and chances to “explore the space” of adaptive user interfaces.