The ENRICH Workshop, Dublin

Paula Goodale

PATHS colleagues participated in the ENRICH  workshop held in August at SIGIR 2013, one of the leading conferences relating to information retrieval. The workshop was hosted by Trinity College, Dublin. 

ENRICH was organised by members of the EU-funded Cultura  project, which has been undertaking work that is complementary to PATHS, also looking at exploration of cultural heritage content, but with a stronger focus on specialist digital humanities collections, than more general materials. PATHS were also sponsors of the workshop.

The keynote speech , “When search becomes research and research becomes search”, by Jaap Kamps, of University of Amsterdam, gave a very interesting vision of the future needs for information retrieval research in cultural heritage, calling for a revolution in our approach to this area. In particular, he talked about possibilities for much more collaborative developments, actively involving both developers and humanities researchers, and using co-creation methods such as ‘living labs’  for building systems that address real-life problems and research questions.

A diverse programme  of six long and short papers then followed, addressing a wide variety of research relating to the workshop theme of Exploration, Navigation and Retrieval of Information in Cultural Heritage), incorporating aspects as diverse as user studies, natural language processing, personalisation, transaction log analysis, linked data, and more traditional IR evaluations. One of the more novel and fascinating research projects was presented by Doug Oard of University of Maryland, describing the mammoth task of digitising all of the vast amounts of available data to reconstruct the Apollo Moon Landings mission control centre. These included days of individual verbal communication lines between team members in mission control and the astronauts on the Apollo spacecraft, as well as flight and other data. Once completed, the reconstruction provides great opportunities to analyse aspects of human behaviour such as decision making under stressful conditions.

Later, a poster and demo session gave an opportunity to talk in more detail about a range of work-in-progress research, covering more NLP and linked data, social network analysis, and some live system demos, including Cultura and PATHS. It was interesting to see these two systems side-by-side, as their overlaps and differences become very apparent. Both use an element of narrative for navigating their collections, but whereas Cultura uses narratives curated by experts, in PATHS the narratives can be created by any user via simple embedded editing and publishing tools. In addition, Cultura makes much greater use of in-collection links enabled by rich metadata, whereas the links in PATHS are based on similarity of simple topic data, and PATHS content also links to external material in Wikipedia. In short, whilst both systems have similar goals, there is an overall impression that at the present time, Cultura is tailored more to digital humanities research at a more scholarly level, and PATHS has a more general audience, although this may largely be a factor of the content each has been working with.

The PATHS demo (second prototype ) was well-received, with particular interest in the new thesaurus and map exploration modes, and we explained how we had noticed that the availability of these tools had begun to change users’ information seeking behaviour compared to the first prototype. Several people also asked about how PATHS might support more collaborative approaches to working, creating and sharing paths in different ways, especially in educational and social media contexts, and amongst existing communities of researchers. These ideas have also been raised by participants in our evaluation sessions and provide interesting ideas for future research.

Share |