DEOS 2013 - The Third International Workshop on “Data Extraction and Object Search” (DEOS 2013)

The Third International Workshop on “Data Extraction and Object Search” (DEOS 2013) will take place as a satellite event of BNCOD 2013 in Oxford, United Kingdom, on July 7th, 2013. The goal of the workshop is to present and discuss ongoing work on data extraction and object search for products, events, reviews, and other types of structured data on the web. Nilesh Dalvi (Facebook) and Roberto Navigli (Rome, La Sapienza) are confirmed as keynote speakers.
We invite researchers and practitioners in this field to contribute with talks about recent work or to join us to get an up-to-date view of this dynamic field of research. The workshop brings together researchers from all aspects of object search, crawling and automated form filling, object identification and extraction, and integration and cleaning of the extracted objects.
This year DEOS has a particular focus on (1) the challenges posed by modern, scripted, highly visual interfaces and websites, and (2), in line with BNCOD’s big data theme, on the use of *big data* for improving data extraction. For example, preexisting “big” data bases, web services, or linked open data endpoints can guide extraction or help to enrich the extracted data.
This is the third installation of DEOS, the first held in Como in 2010 jointly with the SeCo workshop, the second in Vienna in 2011. The workshop is supported by the ERC DIADEM grant and the Oxford Martin school. There is a small amount of travel support available from the sponsors (contact: deos2013 at easychair.org).
TOPICS OF INTERESTS
The topics of interest for this workshop include (but are not limited to):
Object identification and extraction approaches in domains such as products, events, reviews, forum posts, real estate, etc. Hybrid approaches are of particular interest, i.e., approaches that make use of a variety of clues on a web site, e.g., annotations and structure, visual and structure, or visual and annotations.
Big data-supported data extraction where data extraction is guided or improved through the use of external knowledge bases such as wordnet, DBpedia or LinkedGeoData.
Automatic crawling and exploration of web interfaces with a particular focus on highly-visual, scripted web applications.
Information extraction meets data extraction including approaches that integrate information extraction, e.g., from product titles, with data extraction for mutual verification of the extracted data.
Integration and cleaning of extracted web data for object search including approaches or tools for deduplication (intra- and inter-site) and for reconciliation of differing attribute values.
Object search approaches and systems that provide a search interface to data extracted from the web.
Benchmarks for approaches in all of the above topics, but particularly for object identification and form exploration.