Papers are invited on substantial, original, and unpublished research on all aspects of Named Entity Recognition (NER) for South and South East Asian (SSEA) languages. Papers can report results for other languages, but at least one of the languages considered should be an SSEA language. We also invite researchers to be contestants in a shared task (the second track of the workshop) on NER for SSEA languages.
Background and Motivation
Most of the SSEA languages are scarce in resources as well as tools and NER systems are no exception. It is very important that good systems for NER be available, because many problems in information extraction and machine translation (among others) are dependent on accurate NER. However, the issues involved are significantly different for these languages from those for European languages or even East Asian languages. For example, these languages do not have capitalization, which is a major feature for NER systems for European languages.
Another similarity among these languages is that many of them use scripts of Brahmi origin. For some languages, there are additional issues such as word segmentation (e.g. for Thai). Large gazetteers are not available for most of these languages. Lack of standardization and spelling variation add further problems. The number of frequently used common nouns which can also be used as names is very large for many languages, unlike European languages where a larger proportion of the first names are not used as common words. Lastly, and most importantly, there is a serious lack of labeled data for machine learning.
This workshop follows and builds upon this year's NLPAI Machine Learning Contest (an annual event) which focuses on application of machine learning techniques for one major NLP problem every year. This year the problem was NER. However, unlike that event, this workshop will have one track for regular research papers on NER for SSEA languages and the second track will be on the lines of a shared task.
In the shared task, the contestants having their own NER systems will be given some annotated test data. The participating systems will be ranked according to their performance on the test data. There may or may not be training data for a particular language. In either case, the contestants will have the freedom to use any technique for NER, e.g. a purely rule based technique or a purely statistical technique.
At present some data is available for Hindi, Bengali and Telugu for the shared task. Other languages can be included in the contest provided data for them becomes available. The data released for the shared task will be made accessible to all researchers, not just the participants.
If the language you are interested in has not been included in the shared task, you can also prepare the annotated test data and submit it to us. We will then include that language in the shared task.
The task in this contest will be different in one important way. The NER systems also have to identify nested named entities. For example, in the sentence The Lal Bahadur Shastri National Academy of Administration is located in Mussoorie, 'Lal Bahadur Shastri' is a Person, but 'Lal Bahadur Shastri National Academy of Administration' is an Organization. In this case, the NER systems will have to identify both 'Person' and 'Organization' in the given sentence.
Paper submission is through the centralized workshop submission page. Papers have to be written in English. Note that shared task contestants also have to submit a paper describing their method and the results etc. Long or short papers can be submitted to either of the tracks. Long papers can be up to 8 pages long, while the maximum length for short papers is 5 pages (including references, figures, tables etc.). All selected papers will be published in the workshop proceedings.
The papers should be formatted using the LaTeX styles or MS Word templates recommended for the main IJCNLP conference. These documents are available here. Reviewing will be blind. The draft papers should not contain any information that can identify the authors, as far as possible.
- Release of Training and Development Data: Aug 2 to Aug 25, 2007 (for different languages)
- Release of Test Data: Sept 13, 2007
- Annotated Test Data Submission Deadline: Sept
15 16, 2007
- Paper Submission Deadline:
Sept 21, 2007 Sept 25 (11:59 pm PST)
- Notification of Paper Acceptance:
Oct 26, 2007 Nov 5, 2007
- Camera Ready Submission Deadline: Nov 16, 2007
Note: There is no separate registration for the shared task(the contest). You will be a contestant if you submit the annotated test data by the deadline mentioned above.
Rajeev Sangal, IIIT, Hyderabad, India
Dekai Wu, The Hong Kong University of Science & Technology, Hong Kong
Ted Pedersen, University of Minnesota, USA
Dipti Misra Sharma, IIIT, Hyderabad, India
Virach Sornlertlamvanich, TCL, NICT, Thailand
Alexander Gelbukh, Center for Computing Research, National Polytechnic Institute, Mexico
M. Sasikumar, CDAC, Mumbai, India
Sudeshna Sarkar, Indian Institute of Technology, Kharagpur, India
Thierry Poibeau, CNRS, France
Sobha L., AU-KBC, Chennai, India
Tzong-Han Tsai, National Taiwan University, Taiwan
Prasad Pingali, IIIT, India
Canasai Kreungkrai, NICT, Japan
Manabu Sassano, Yahoo Japan Corporation, Japan
Kavi Narayana Murthy, University of Hyderabad, India
Sivaji Bandyopadhyay, Jadavpur University, Kolkata, India
Anil Kumar Singh, IIIT, Hyderabad, India
Doaa Samy, Universidad Autónoma de Madrid, Spain
Ratna Sanyal, Indian Inst. of Inf. Tech., Allahabad, India
V. Sriram, IIIT, Hyderabad, India
Anagha Kulkarni, Carnegie Mellon University, USA
Soma Paul, IIIT, Hyderabad, India
Sofia Galicia-Haro, National Autonomous University, Mexico
Grigori Sidorov, National Polytechnic Institute, Mexico
Dipti Misra Sharma, Rajeev Sangal, Anil Kumar Singh
Language Technologies Research Centre
International Institute of Information Technology
Gachibowli, Hyderabad, India
Phone: 91-40-23001412, 91-40-23001967/9 Extension 144
Email: firstname.lastname@example.org, email@example.com, firstname.lastname@example.org