I'm a postdoc researcher at Aalto University in Finland. I work in Antti Oulasvirta's User Interfaces group.

I'm primarily interested in the application of probabilistic reasoning and techniques from machine learning to problems in HCI. In particular, I've worked extensively on a probabilistic model for touch using Gaussian Process regression.


Sketchplore: Sketch and Explore with a Layout Optimiser

Kashyap Todi, Daryl Weir and Antti Oulasvirta
In Proceedings of DIS '16
This paper proposed a tool for supporting the creativity of layout designers using an optimiser. The system automatically evaluates thousands of candidate layouts according to models of human performance and preference, and makes suggestions about how sketched layouts can be improved.

How We Type: Movement Strategies and Performance in Everyday Typing

Anna Maria Feit, Daryl Weir and Antti Oulasvirta
In Proceedings of CHI '16
This paper studies how people type in the real world. We used motion capture to study the movements of both trained touch typists and untrained, everyday users. We found surprisingly few differences between these groups: those without training were just as fast on average.

Modelling and Correcting for the Impact of the Gait Cycle on Touch Screen Typing Accuracy

Josip Musić, Daryl Weir, Roderick Murray-Smith and Simon Rogers
mUX: The Journal of Mobile User Experience
This paper extends the GP regression approach to touch by incorporating information about periodic variations in behaviour as the user walks around. We show this improves accuracy over existing touch models.

Uncertain Text Entry on Mobile Devices

Daryl Weir, Henning Pohl, Simon Rogers, Keith Vertanen, and Per Ola Kristensson
In Proceedings of CHI '14
This paper looked at combining probabilistic touch models with statistical language models to create an autocorrect system. We also looked at what happens when users are able to control the amount of uncertainty they express in their typing.

Sparse Selection of Training Data for Touch Correction Systems

Daryl Weir, Daniel Buschek, and Simon Rogers
In Proceedings of MobileHCI '13
This paper investigated the use of the Relevance Vector Machine (RVM) to train touch models with very small numbers of training examples - 10 or fewer.

A User-specific Machine Learning Approach for Improving Touch Accuracy on Mobile Devices

Daryl Weir, Simon Rogers, Roderick Murray-Smith, and Markus Löchtefeld
In Proceedings of UIST '12
This paper proposed the use of Gaussian Process regression to model touch offsets on mobile devices in a probabilistic manner.


Modelling Uncertainty in Touch Interaction

Ph.D. Thesis
University of Glasgow, 2014
My doctoral thesis focused on the application of machine learning algorithms to the touch offset problem. It shows that explicitly modelling the uncertainty in touch can have interaction benefits.

A Machine Learning Teaching Aid

M.Sc. Thesis
University of Glasgow, 2010
My Master's thesis details the implementation of a Java application designed to visually teach people about the use of a range of machine learning algorithms.