26 January 2018
With centuries of experience in libraries, one would expect that book searching is a solved problem. This talk will provide an introduction to book information needs that are too difficult for current search systems to solve. Such information needs often cover a combination of different aspects, such as specific genres or plot elements, engagement or novelty. The talk presents a large-scale empirical investigation how current book (meta-)data in search systems cannot provide enough clues to fulfill these search requests using a test collection of over 2 million book records and over 330 real-world book information needs. An expanded analysis shows that complex information needs might be remarkably similar in related domains such as movies. The talk invites participants to discuss potential solutions and future research opportunities in interactive information retrieval.
Vivien Petras is a professor for Information Retrieval at the Berlin School for Library and Information Science at Humboldt-Universität zu Berlin, Germany. She received her PhD in Information Management and Systems from the School of Information, University of California, Berkeley. Her research interests are in the design and evaluation of cultural heritage information systems, particularly the challenges of multilinguality and semantic heterogeneity. She actively collaborates with the European digital library, archive and museum Europeana (www.europeana.eu) and coordinated and led several evaluation labs at the CLEF Conference, for example the domain-specific track, the CHIC (Cultural Heritage in CLEF) and the SBS (Social Book Search) labs. The work presented at the colloquium stems from work on the SBS research data. An additional focus is the development and evaluation of national information infrastructures: she served as a working group chair for the BMBF-funded project on Kerndatensatz Forschung (research data sets) and is now part of the expert committee for the evaluation of the DFG Specialized Information Services Programme (FID Fachinformationsdienste).
Host: Dr. Ralf Krestel