Summer research projects for undergraduates

Undergraduate Research Fellowships are an exciting new scheme to offer you the opportunity to work on and experience a research project. Across the College, fellowships will be available for up to 80 students with a stipend of £1000 each. After the summer projects are completed, outcomes will be presented at a celebration event, with prizes awarded to the most successful projects. While you can apply for a project in any department, the three below will be hosted in Department of Informatics.

1. Understanding the importance of exposure vs quality in driving content consumption

We all consume content online, from news to BBC iPlayer. This project attempts to unravel the question: Is content consumption driven more by what the website decides to highlight (eg on their home page), or is it driven by the quality of the item itself? You will build alternate interfaces to iPlayer, and conduct a user study to understand this issue in a controlled experimental setting. You work will complement data analysis of researchers looking into iPlayer accesses from nearly a quarter of the UK population.

Further details on this project:

https://internal.kcl.ac.uk/staff/kli/ke/ugres/Project-docs-dept/By-School/NMS-I-14-14.pdf

2. Visually mapping critical borderlands

Wherever conflicts are in the world, charities, NGOs and governmental departments need up to date and reliable information to be able to offer humanitarian support to local populations. This is especially difficult to obtain where conflicts cross national boundaries. The data that is commonly available, such as databases held by international organisations, may be reliable but updated fairly infrequently. Alternative sources, such as information posted on Twitter and other social networks by people living in the conflict zone, are updated frequently, but are harder to mine for the general facts needed by policymakers. In either case, the raw data is not what policymakers require to do their jobs. Visual map-based infographics describing at a high level what is happening across the region over time would be much more valuable. The fellow will join efforts to take initial steps in solving this problem, and will take the lead on a well-defined part of this research programme.

Further details on this project:

https://internal.kcl.ac.uk/staff/kli/ke/ugres/Project-docs-dept/By-School/NMS-I-14-08.pdf

3. Automated planning for argument dialogues
 
Argument dialogue theory provides a principled way of structuring rational interactions between participants (be they machine or human) who argue about the validity of certain claims in order to try to achieve particular dialogue goals. For example, participants may aim reach a joint agreement about what to do and each may have a preferred outcome they want to be agreed on; they would thus need to present convincing arguments in order to influence the outcome in their favour. For machines to participate effectively in such dialogues, they must be able to select particular arguments to put forward that will help them achieve their goals, taking into account what they believe about the other dialogue participants. The aim of this project is to investigate the feasibility of using automated planning techniques to determine which arguments to put forward during argument dialogues.

Further details on this project:

https://internal.kcl.ac.uk/staff/kli/ke/ugres/Project-docs-dept/By-School/NMS-I-14-03.pdf