May 23

Knowledge is commonly socially constructed, through collaborative efforts towards shared objectives or by dialogues and challenges brought about by different persons' perspectives.

-- G. Salomon (in "Distributed Cognitions: Psychological and Educational Considerations")

 
 

Featured chapter

Read the fascinating history of Wearable Computing, told by its father, Steve Mann

Read Steve's chapter !

 
 

Help us help you!

 
 

Daryl Weir

Add description
Add publication

Publications by Daryl Weir (bibliography)

 what's this?
2012
 
Edit | Del

Weir, Daryl, Rogers, Simon, Murray-Smith, Roderick and Löchtefeld, Markus (2012): A user-specific machine learning approach for improving touch accuracy on mobile devices. In: Proceedings of the 2012 ACM Symposium on User Interface Software and Technology 2012. pp. 465-476.

We present a flexible Machine Learning approach for learning user-specific touch input models to increase touch accuracy on mobile devices. The model is based on flexible, non-parametric Gaussian Process regression and is learned using recorded touch inputs. We demonstrate that significant touch accuracy improvements can be obtained when either raw sensor data is used as an input or when the device's reported touch location is used as an input, with the latter marginally outperforming the former. We show that learned offset functions are highly nonlinear and user-specific and that user-specific models outperform models trained on data pooled from several users. Crucially, significant performance improvements can be obtained with a small (≈200) number of training examples, easily obtained for a particular user through a calibration game or from keyboard entry data.

© All rights reserved Weir et al. and/or ACM Press

 
Edit | Del

Weir, Daryl (2012): Machine learning models for uncertain interaction. In: Adjunct Proceedings of the 2012 ACM Symposium on User Interface Software and Technology 2012. pp. 31-34.

As interaction methods beyond the static mouse and keyboard setup of the desktop era -- such as touch, gesture sensing, and visual tracking -- become more common, existing interaction paradigms are no longer good enough. These new modalities have high uncertainty, and conventional interfaces are not designed to reflect this. Research has shown that modelling uncertainty can improve the quality of interaction with these systems. Machine learning offers a rich set of tools to make probabilistic inferences in uncertain systems -- this is the focus of my thesis work. In particular, I'm interested in making inferences at the sensor level and propagating uncertainty forward appropriately to applications. In this paper I describe a probabilistic model for touch interaction, and discuss how I intend to use the uncertainty in this model to improve typing accuracy on a soft keyboard. The model described here lays the groundwork for a rich framework for interaction in the presence of uncertainty, incorporating data from multiple sensors to make more accurate inferences about the goals of users, and allowing systems to adapt smoothly and appropriately to their context of use.

© All rights reserved Weir and/or ACM Press

 
Add publication
Show this list on your homepage
 
 

Join the technology elite and advance:

 
1.

Your career

 
2.

Your network

 
 3.

Your skills

 
 
 
 
 
 

Changes to this page (author)

23 Nov 2012: Added
23 Nov 2012: Added

Page Information

Page maintainer: The Editorial Team
URL: http://www.interaction-design.org/references/authors/daryl_weir.html
May 23

Knowledge is commonly socially constructed, through collaborative efforts towards shared objectives or by dialogues and challenges brought about by different persons' perspectives.

-- G. Salomon (in "Distributed Cognitions: Psychological and Educational Considerations")

 
 

Featured chapter

Read the fascinating history of Wearable Computing, told by its father, Steve Mann

Read Steve's chapter !

 
 

Help us help you!