linguistlist.org>
Editor for this issue: Karolina Owczarzak <karolinalinguistlist.org>
cs.rutgers.edu>============================================================================= CALL FOR PAPERS Second International Workshop on Paraphrasing: Paraphrase Acquisition and Applications July 11, 2003, in Sapporo, Japan in conjunction with ACL-2003 (WS5) http://nlp.nagaokaut.ac.jp/IWP2003/ BACKGROUND A common characteristic of human languages is the possibility to convey the same information in several ways. Paraphrases, which in the literature have also been referred to as variants, reformulations, or inference rules, span a wide range of variation: - article / paper / publication - Oswald killed Kennedy. / Kennedy was assassinated by Oswald. - a plant in Alabama / the Alabama plant - Edison invented the light bulb. / Edison's invention of the light bulb - He plays better than everybody else in the team. / He's the best in the team. - The tree healed its wounds by growing new bark. / The tree healed its wounds. It grew new bark. - The stapler costs $10. / The price of the stapler is $10. - Where is Thimphu located? / Thimphu is the capital of what country? This diversity of expression presents a major challenge for many NLP applications. Thus, automatic paraphrase identification and generation can benefit a broad range of NLP tasks, including machine translation, summarization, information retrieval, question answering, generation, and authoring and reading assistance. Previous workshops on paraphrasing: - Workshop on Automatic Paraphrasing, November 2001 accompanying the NLPRS2001 conference, with 55 participants http://nlp.nagaokaut.ac.jp/pub/NLPRS2001WS.html - Workshop on Automatic Paraphrasing (in Japanese), March 2001 accompanying Japanese NLP conference, with 165 participants http://nlp.nagaokaut.ac.jp/pub/NLP2001WS.html TOPICS OF INTEREST The workshop will be open to any research topic related to paraphrases. More specifically, topics of interest include, but are not limited to: * definition and typology of paraphrases * representation of paraphrases * algorithms for recognizing, generating and choosing among paraphrases * construction of paraphrase resources * existing and potential applications of paraphrases: - question answering, summarization, information retrieval, machine translation, authoring and reading assistance - inferencing with paraphrases * evaluation of paraphrase algorithms and resources Special topic: Paraphrase Acquisition The increased availability of parallel corpora and comparable corpora has opened up possibilities for automatic paraphrase acquisition. As we have recently witnessed, a number of new methods for paraphrase extraction have emerged. The availability of appropriate evaluation techniques is a key part of a progress in the area. Is it possible to create a common benchmark for evaluating different paraphrase extraction approaches? On which terms should different acquisition approaches be compared? How can we define the notion of baseline? Another important objective of the workshop is to take a first step towards a standardized paraphrase resource that could be shared among a large variety of researchers. "SOMETHING_1 costs MONETARY_QUANTITY_2" :is-equivalent-to "the price of SOMETHING_1 is MONETARY_QUANTITY_2" :can-be-inferred-from "to sell SOMETHING_1 for MONETARY_QUANTITY_2" Such a resource, with possibly tens of thousands of entries such as the one above (in one format or another), can be viewed as a valuable extension of WordNet and holds great promise to advance many areas of natural language processing. SUBMISSIONS Paper submissions must be anonymous and are limited to at most 8 pages including references, figures etc. Authors are encouraged (but not required) to use the ACL style format of the main conference. Only electronic submissions will be accepted. Please email your submission in pdf (preferred), postscript, or MS Word to the following address: iwp2003-submissionMail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue
cs.rutgers.edu>CALL FOR PAPERS CoNLL-2003: Seventh Conference on Natural Language Learning Organized at HLT-NAACL-02, Edmonton, Canada May 31 - June 1 2003 http://cnts.uia.ac.be/conll2003/ CoNLL is an international forum for discussion and presentation of research on natural language learning. We invite submission of papers about natural language learning topics, including, but not limited to: - Computational models of human language acquisition - Computational models of the origins and evolution of language - Machine learning methods applied to natural language processing tasks (speech processing, phonology, morphology, syntax, semantics, discourse processing, language engineering applications) - Symbolic learning methods (Rule Induction and Decision Tree Learning, Lazy Learning, Inductive Logic Programming, Analytical Learning, Transformation-based Error-driven Learning) - Biologically-inspired methods (Neural Networks, Evolutionary Computing) - Statistical methods (Bayesian Learning, HMM, maximum entropy, SNoW, Support Vector Machines) - Reinforcement Learning - Active learning, ensemble methods, meta-learning - Computational Learning Theory analysis of language learning - Empirical and theoretical comparisons of language learning methods - Models of induction and analogy in Linguistics CoNLL is the yearly meeting organized by SIGNLL, the Association for Computational Linguistics Special Interest Group on Natural Language Learning. After previous CoNLL meetings were held in Madrid (1997), Sydney (1998), Bergen (1999) Lisbon (2000), Toulouse (2001), and Taipei (2002). See http://www.aclweb.org/signll and http://ilk.uvt.nl/~signll/conll.html for more information about SIGNLL and CoNLL. Special Theme - ----------- As in previous years, in addition to submissions on the general topics listed above, we encourage submissions on a special theme. This year's special theme is: Semi-supervised / unsupervised learning and sample selection techniques for language learning (co-training, active learning, EM, etc). Supervised Machine Learning methods suffer from a "data annotation bottleneck" which is especially harmful for language learning tasks where a lot of training data is needed (e.g. parsing). Sample selection techniques, and combination of supervised learning with semi-supervised and unsupervised techniques may provide a solution to this problem. Shared Task - --------- This year's workshop will also accept submissions for a shared task: machine learning approaches to named entity recognition. Special attention will be given to the use of multiple sources of knowledge, like training data, lists of examples and unannotated data. Interested groups will be supplied with the same training and testing material (in several languages), and will all use the same evaluation criteria, thus allowing comparison between various learning methods. More information on the shared task will be available at: http://cnts.uia.ac.be/conll2003/ner/ Invited Speaker - ------------- TBA Submissions - --------- * Main Session Submissions Submit a full paper of no more than 8 pages (Postscript, PDF or plain text ASCII) by March 16, 2003 electronically to the address below. Authors of accepted submissions will be invited to produce a final paper to be published in the proceedings of the workshop, which will be available at the workshop for participants, and distributed afterwards by ACL. Final submissions must follow the HLT-NAACL style (http://www.sims.berkeley.edu/research/conferences/hlt-naacl03/format.html). We strongly recommend the use of these style files also in the submission. Submit main session papers to: Walter Daelemans daelemMail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue