MOVING successfully participated with two full research papers in the 18th ACM/IEEE Joint Conference on Digital Libraries (JCDL 2018) in Fort Worth, Texas, from 3 to 6 June 2018. JCDL is the top conference for research on digital libraries and gathered 200 participants from 25 different countries, including experts from information science, computer science, and the humanities. Florian Mai presented the full paper “Using Deep Learning for Title-Based Semantic Subject Indexing to Reach Competitive Performance to Full-Text” on Tuesday, June 5, in the session on text collections.
Florian Mai presents the MOVING paper (image available from Twitter)
Florian’s talk attracted more than 50 participants and showed that for scientific digital libraries automatic subject indexing methods based on the title of a publication can be as good as or even better than the full-text-based methods by leveraging the surplus of data available to train powerful neural network models. Using titles only is important as this information is usually freely available while full-text may be subject to license.
Tobias Backes presented the full paper “Exploring and Analyzing Collections” on Tuesday, June 5, in the session on exploring and analyzing collections. Tobias’s talk attracted more than 50 participants and presented the problem of author name homonymy in the Web of Science and introduced a novel probabilistic similarity measure for author name disambiguation based on feature overlap.