The proliferation of the Internet has revolutionised the way information is disseminated and presented. Blogs no longer only relay and comment on news stories but also influence what is talked about in the news. Such changes have not gone unnoticed by the computational linguistics research community, who are increasingly processing or exploiting blogs in an attempt to keep track of what is going on and mine information. This workshop will focus on how events can be identified and how information related to event processing (e.g. NP coreference, temporal processing) can be extracted from blogs. Emphasis will be on how existing methods for event processing need to be adapted in order to process this medium, and on linguistic differences in the reporting of events in blogs and more traditional news texts. Extensions of this research to text types from other collaborative environments such as wikis, fora, chats and social networks are also welcomed.
Event detection and processing is not a new topic in computational linguistics, but until now it has focused mainly on processing of newswire. The TimeBank corpus (Pustejovsky et. al. 2003), the AQUAINT TimeML corpus, and the NP4E corpus (Hasler, Orasan and Naumann 2006) exclusively contain newswire, which may make them inappropriate for the development of methods which need to process other text types. This workshop will give researchers the opportunity to present efforts to develop resources related to event identification and processing using blog entries. In addition, papers describing annotation guidelines and linguistic analyses of such resources are encouraged, including comparisons with annotations of more traditional text types.
The informal style and structure of most blog entries makes event detection in these documents a difficult task. This workshop will encourage submission of papers that develop new methods for event identification or test existing ones on blog entries. The fact that the same event is usually reported in several entries, in many instances containing contradictory information, makes cross-document event coreference identification a more challenging task. Researchers are encouraged to submit papers presenting such methods. Of particular interest are papers detailing methods to identify subjective or objective ways of presenting an event or to detect contradictory reporting of the same event.
The workshop is not restricted only to event processing. Papers on topics related to event processing, such as the identification of entities and temporal processing and reasoning are also encouraged. Submissions about NLP applications that exploit event identification such as summarisation of events in blogs, question answering about events described in collaborative environments (e.g. Wikipedia) or opinion mining related to events are also invited.
Original papers in three main areas are expected:
Linguistics
- linguistic representation and function of different kinds of information (e.g. opinion/subjectivity/objectivity)
- discourse structure related to the presentation of events within and across blog entries
- annotation of events in corpora of blogs
- extension/adaptation of 'traditional' annotation of news texts for blog entries
- description of resources encoding information relevant to this workshop
Methods for event processing or necessary for event processing
- event identification in blogs
- contradictory information regarding the same events (cross-document NP and event coreference)
- entity tracking across blog entries
- identification, representation and reasoning about time in blogs
- evaluation of event identification methods trained on newswire vs blogs
NLP methods and applications that use information about events
- summarisation of events from blogs
- opinion mining and analysis in blogs vs newswire
- identification of objective and subjective information in blogs vs newswire
- commercial systems which process events in blogs
Prof. Mike Thelwall, University of Wolverhampton
Format
Authors are invited to submit three types of papers: full papers which describe original and unpublished work in the topic area of this workshop, posters which describe work in progress, and demos which describe full working systems, and which, if accepted, will be presented at a special session accompanied by live demo. Papers should be submitted as a PDF file, formatted according to the RANLP 2009 stylefiles and not should not exceed 8 pages for full papers and 4 pages for posters and demos. The RANLP 2009 stylefiles are available at: http://lml.bas.bg/ranlp2009/submissions.htm.
As reviewing will be blind, the papers should not include the authors' names and affiliations. Furthermore, self-references that reveal the authors' identities should be avoided. Papers that do not conform to these requirements will be rejected without review.
Submission procedure
Submission of papers will be handled using the START system.Reviewing
Each submission will be reviewed at least by two members of the Program Committee. Reviewers will be asked to provide detailed comments, and to score submitted papers on the following factors:- Relevance to the workshop
- Significance and originality
- Technical/methodological accuracy
- References to related work
- Presentation (clarity, organisation, English)
Accepted papers policy
Accepted papers will be published in the workshop proceedings. By submitting a paper at the workshop the authors agree that, in case the paper is accepted for publication, at least one of the authors will attend the workshop; all workshop participants are expected to pay the RANLP-2009 workshop registration fee.
Paper submission deadline: June 20, 2009
Paper acceptance notification: July 27, 2009
Camera-ready papers due: August 24, 2009
Workshop date: September 18, 2009
Constantin Orasan, University of Wolverhampton, UK
Laura Hasler, University of Wolverhampton, UK
Corina Forascu, Al.I. Cuza University of Iasi, Romania