User interfaces (UIs) are typically static – they do not change their organization based on a user’s actions or context. In contrast, adaptive UIs can alter their representation, style, and content based on user-specific factors such as the prior experience, interests, and specific abilities, or contextual factors such as location, time, or task at hand. As such, they offer a promising approach to dynamically improve the long-term usability of interactive systems for individual users.
In this talk, after providing a brief overview of my research background, I will discuss, in detail, some of our work in the area of adaptive user interfaces and interactions. I will first discuss how predictive and statistical models can be applied towards adapting UIs to improve their usability. I will showcase two self-adapting systems, as examples, where common UIs such as webpage layouts and graphical menus are dynamically adapted for individual users. Next, I will discuss some key challenges faced by typical adaptive systems, where suboptimal adaptations often limit their efficacy. Finally, I will present some of our latest work that addresses these challenges by developing model-based reinforcement learning methods to plan adaptations, and offer a future outlook.
Kashyap Todi is a research scientist at Meta Reality Labs (Redmond). Previously, he was a postdoctoral researcher in Human–Computer Interaction, working at Aalto University (Finland) in the User Interfaces Research Group, led by Prof. Antti Oulasvirta. He holds a PhD from Hasselt University (Belgium) where he worked primarily with Prof. Kris Luyten at the Expertise Centre for Digital Media. His research focuses on computational methods for intelligent interactive systems. More specifically, he develops and applies computational approaches (e.g. predictive/statistical models, optimisation, RL) for adaptive user interfaces and mixed-initiative tools. Kashyap’s work has been published at venues such as CHI, IUI, MobileHCI, and DIS, and demonstrated as interactive exhibits as well. He is also an active member of the international academic community – he frequently serves as an associate chair or organising committee member at venues such as CHI, MobileHCI, IUI, and ISS, and is the Vice President for SIGCHI Operations. You can find more details about his interests and academic activities on his website and reach out to him via twitter (@kashtodi).