Our research focuses on building representative user models of the long-term and evolving information needs expressed through human-computer interactions during Web search sessions. The implicit and explicit interactions are considered to represent the intentions driving a person’s information needs and their behaviours are used to model their intentions. Multiple-aspect user models are created and maintained based on these actions and are used to retrieve new information on behalf of the people they represent to attempt to assist them with the ongoing search tasks.

I recently presented a paper at the Contextual Information Access, Retrieval, and Evaluation Workshop at the 31st European Conference on Information Retrieval in Toulouse, France. It was quite a challenging experience to present my work in 15 minutes so I chose to focus on the evaluation challenges of multi-session experiments in interactive information retrieval.

We are currently investigating and experimenting with evaluation methodologies for multiple-session retrieval processes and hope to publish our progress soon.

I have been part of a team researching and developing a recommender system for the DSpace digital repository system. Our approach is to extract terms from the stored items and the metadata associated with the items and use the terms to perform a comparison between a pair of items to determine if a recommendation is appropriate. Our work has been exploratory and proof-of-concept.

On 3rd April 2008, we delivered a presentation to approximately 100 delegates at the Open Repositories 2008 DSpace User Group in Southampton, England. On 21st April 2008, we published a technical report through the HP Labs Library detailing the first phase of our work.

I’m rather enjoying my time at HP Labs, there are plenty of interesting projects and a wide variety of extra-curricular activities. For example, the HP Labs Science Lectures are a series of evening events to promote the breadth and depth of science to the public. In January, we were entertained by Mike Edmunds of Cardiff University and his presentation of the work-in-progress discovery of the Antikythera Mechanism. This week, Herbert Huppert from Cambridge University will deliver a lecture on Extreme Natural Disasters and I’m looking forward it!

Finally, I liked Apple’s Think Different corporate culture advertising campaign and was pleased to find a dusty old poster in Building 3 of a corporate culture advertising campaign from HP. I think these rules offer an interesting way to think about your new career, regardless of profession.