[Corpora-List] Call for papers: Natural Language Processing and Information Retrieval Workshop

Oakes, Michael Michael.Oakes at wlv.ac.uk
Tue May 9 11:40:00 CEST 2017


Call for Papers. Natural Language Processing and Information Retrieval Workshop. RANLP 2017, Varna, Bulgaria, 7-8 September, 2017 http://rgcl.wlv.ac.uk/events/NLPIR/home.html

Since the 1990s, people have mooted the probability of enriching search engine algorithms with higher levels of linguistic processing. However, this idea was one of "promise rather than substance" (Smeaton), and search engines today still work predominantly with the "bag of words" model, where single words are considered in isolation. The main exceptions, which have for some time have led to significant improvements, are the linguistically simple methods of stopwords and stemming. Higher levels of linguistic processing, such as parsing and Word Sense Disambiguation, have yielded only small improvements or even a degradation in performance, at the cost of increased storage costs and computational complexity (Brants, 2014).

Recently though, there have been breakthroughs, with the word2vec model being derived from distributional semantics, and the idea of semantic search where the documents are annotated with semantic codes corresponding to meaning. Away from core information retrieval, natural language processing has proved itself useful in many tasks associated with peripheral information retrieval: Question Answering, Text Summarisation, Information Extraction, Speech Processing, and Natural Language Generation, Word Prediction, Terminology Extraction, Named Entity Recognition, Anaphor Resolution and Cross-Language Information Retrieval. Russell-Rose and Stevenson (2009) describe how the use of Natural Language Processing in Text Analytics or Text Mining, the discovery of new knowledge from unstructured text resources has now become mainstream.

Conversely, search engines have a role to play in studies of natural language, particularly those involving corpora or large collections of real-life text stored electronically. Search engines are needed to retrieve that data which is pertinent to the focus of linguistic enquiry, especially since many people have regarded the whole WWW as a corpus. Specialised search engines are needed to retrieve text which is annotated for a specific purpose, such as all occurrences of the word "spring" when it is a noun, or when it is a grammatical error made in a corpus of learner English. Search engines are also needed in the creation of corpora, for example finding all the texts on the web in a given lesser-resourced language. This workshop will focus mainly on the use of NLP in information retrieval tasks.

The workshop is held for the first time. It encourages high quality contributions from all over the world. Accepted papers will appear in the proceedings of RANLP 2017 and will be published in the ACL anthology.

* Papers submission deadline: 30 June 2017

* Papers acceptance notification: 28 July 2017

* Papers camera-ready versions: 20 August 2017 Mireille Makary and Michael Oakes, University of Wolverhampton, England.

-------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/html Size: 8969 bytes Desc: not available URL: <https://mailman.uib.no/public/corpora/attachments/20170509/2346b30b/attachment.txt>



More information about the Corpora mailing list