Success in the Laidlaw Undergraduate Internship Programme in Research and Leadership

Congratulations to Patrick Schrempf and Billy Brown who have been successful in their applications for a Laidlaw Undergraduate Internship in Research and Leadership for 2017. You can read further details about Billy and Patrick below.

Billy Brown:

I’m a fourth year Computer Science student from Belgium with too much interest for the subject. I play and referee korfball for the university, and I am fascinated by Old English and Norse history and mythology. I plan on using the Laidlaw Internship programme to get into the field of Computer Science research.

Project summary:

The Essence Domain Inference project aims to improve automated decision making by optimising the understanding of the statements used to define a problem specification. As part of the compilation of the high level Essence specification language, this project would tighten the domains to which a specified problem applies, with a domain inference algorithm.

The work is very much in the context of the recently-announced EPSRC grant working on automated constraint modelling in an attempt to advance the state of the art in solving complex combinatorial search problems. The modelling pipeline is akin to a compiler in that we refine a specification in the Essence language Billy mentions down to a number of powerful solving formalisms. The work Billy plan is to improve the refinement process and therefore the performance of the solvers, leading to higher quality solutions more quickly.

Patrick Schrempf:
I am currently a third year Computer Science student from Vienna. After enjoying doing research with the St Andrews Computer Human Interaction (SACHI) group last year, I am looking forward to the Laidlaw Internship Programme. Apart from research and studying, I enjoy training and competing with the Triathlon Club and the Pool Society.


Project summary:

This project will explore Explainable Artificial Intelligence (XAI) which aims to create artificial intelligence and machine learning models that are combined with effective explanations. Currently most models of artificial intelligence are very intricate and complex, on the contrary using XAI models will enable their users to build a better understanding of the AI components and the system as a whole.

This project builds on Patrick’s work in RadarCat which employed machine learning techniques for material and object classification which enables new forms of everyday proximate interaction with digital devices. RadarCat (Radar Categorization for Input & Interaction) was published at ACM UIST 2016 in Tokyo, Japan, Yeo, H.-S., Flamich, G., Schrempf, P., Harris-Birtill, D., and Quigley, A. (2016) RadarCat: Radar Categorization for Input & Interaction. In Proceedings of the 29th Annual ACM Symposium on User Interface Software and Technology New York, NY, USA: ACM UIST ’16.