<<CALL FOR PAPERS>>
Special issue of ACM Transactions on Asian Language Information
"Recent Advances in Statistical Language Modeling - Beyond N-grams"
Jianfeng Gao, Microsoft Research Asia, Beijing, China
Chin-Yew Lin, Information Sciences Institute, USC, USA
Statistical language modeling (SLM) aims to estimate probability
distribution of various linguistic units, such as words, sentences,
and documents, for the purpose of many natural language applications.
Over the last two decades, many attempts have been made to improve the
state of the art. In this issue, we solicit papers showing recent
advances of SLM in both theory and applications.
It is ironical that the most popular language model (n-grams) uses
very little language knowledge. In recent years, many attempts have
been made that try to "put language back into language model". But
little improvement has been achieved so far in realistic applications
due to two major obstacles: (1) the number of parameters of the
knowledge-based models is usually too large to estimate; (2) the
construction and use of these models requires a large annotated training
corpus and a decoder that assigns linguistic structure, which are not
always available. We are seeking ideas that enhance our understanding
of these core problems in SLM. We encourage submissions that describe
principles, concepts or models on which work in SLM could be based.
SLM has been successfully applied in many applications such as speech
recognition, Asian language input, information retrieval, and machine
translation. We welcome submissions that demonstrate significant
improvement in performance using knowledge-based models, present novel
applications of SLM in new areas such as paraphrasing, question
answering, and text summarization, or how SLM techniques are used
in novel ways to improve the systems' performance.
Areas of interest include, but are not limit to:
- Theory of statistical language modeling (SLM), including
o Formal models (N-gram model, HMM, maximal entropy model,
language model, word/class model, grammar model, etc.)
o Parameter estimation (model smoothing/combination/adaptation)
o Resource (tagged training data) for SLM
- Applications of SLM, including the application of SLM in the
o Question answering
o Text summarization
o Speech recognition
o Asian language input
o Information retrieval
o Named entity recognition
o Text generation
o Machine translation
- Other statistical natural language processing methods beyond the
of SLM, e.g. statistical parsing, machine learning for NLP etc.
The tentative plan is to publish this special issue in Spring 2004.
<<Instructions for Submission>>
Papers should follow the style guidelines for the ACM Transactions on
Language Information Processing
Papers should be sent to the guest editors, by the submission date
listed below. The submission should be either:
- Electronically to firstname.lastname@example.org. The "Subject:" line should
Special Issue Submission.
The following formats are acceptable:
- Adobe PDF
If you cannot produce an electronic version in either of these formats,
if the editor informs you of a problem with your electronic submission,
please follow the instructions for hardcopy submission.
- Or, Three hardcopies to:
Microsoft Research Asia
5F, Beijing Sigma Center
No. 49, Zhichun Road, Haidian District
Beijing, 100080, P.R.C
USC/Information Sciences Institute
4676 Admiralty Way
Marina del Rey, CA 90292
Call for Papers: April 1, 2003
Submission of Papers: August 31, 2003
Notification of Acceptance: October 31, 2003
Final Version Due: January 1, 2003
Special Issue Date: Spring 2004
This archive was generated by hypermail 2b29 : Fri Mar 28 2003 - 23:56:44 MET