EMNLP-CoNLL 2007 Call for Papers Print

EMNLP-CoNLL 2007
Conference on Empirical Methods in Natural Language Processing
Conference on Computational Natural Language Learning

June 28-30, 2007
Prague, Czech Republic
http://cs.jhu.edu/EMNLP-CoNLL-2007

The annual EMNLP and CoNLL conferences will be holding an unprecedented 3-day joint meeting this year. The joint conference will be co-located with ACL 2007 in Prague. We hereby invite submissions.

The focus of EMNLP-CoNLL 2007 is learned models and data-driven systems concerning all aspects of human language. Both empirical and theoretical results are welcome.

EMNLP is the annual conference organized by SIGDAT, the ACL Special Interest Group on linguistic data and corpus-based approaches to NLP. Previous EMNLP meetings were held in Sydney (2006), Vancouver (2005), Barcelona (2004), Sapporo (2003), Philadelphia (2002), Pittsburgh (2001), Hong Kong (2000), College Park (1999), Granada (1998), Providence (1997), and Philadelphia (1996).

CoNLL is the annual conference organized by SIGNLL, the ACL Special Interest Group on natural language learning. Previous CoNLL meetings were held in New York (2006), Ann Arbor (2005), Boston (2004), Edmonton (2003), Taipei (2002), Toulouse (2001), Lisbon (2000), Bergen (1999), Sydney (1998), and Madrid (1997).

TOPICS

We solicit high-quality papers from academia, government, and industry on all areas of interest to SIGDAT and SIGNLL researchers. These areas include (among others):

  • Empirical, statistical, or learning approaches to core NLP tasks
    • Words (speech, phonology, morphology, tagging, word senses, ...)
    • Syntax (chunking, entity recognition, parsing, ...)
    • Semantics (information extraction, relation extraction, textual entailment, sentiment analysis, full semantics, ...)
    • Generation (machine translation, dialogue, summarization, speech synthesis ...)
  • Data-driven techniques in applications of human language technology
    • Information retrieval; information navigation
    • User interfaces, including dialogue systems
    • Domain-specific applications (email, medicine, education, helpdesk, ...)
  • Data mining on natural language data
    • Knowledge discovery
    • Clustering, segmentation, sense discovery, topic discovery, dimensionality reduction
    • Language acquisition (grammar induction, etc.)
    • Scalable algorithms on extremely large corpora
  • New machine learning techniques as applied to natural language data
    • Symbolic methods; statistical methods; ensemble methods
    • Architectures for structural and relational learning
    • Learning with little supervision (unsupervised or semi-supervised learning, active learning, inductive biases, ...)
    • Techniques for multilinguality or cross-domain adaptation
    • Joint modeling of language and non-language data (networks, images, structured databases, ...)
    • Theoretical or empirical comparisons of learning techniques
    • Theory (error bounds, learnability, computational complexity, ...)
    • Algorithms and efficiency
  • Computational models of human language learning
    • Human language acquisition
    • Linguistically informed models, priors, or learning methods
    • Models of analogical or generative processes
    • Learning or frequency effects in human language processing
    • Evolution of language
  • SHARED TASK

    CoNLL traditionally runs a friendly competition, the "shared task." The 2007 CoNLL shared task is dependency parsing, for the second year running. There are actually two tasks: a "multilingual track" as in 2006, and a new "domain adaptation track" for adapting to different domains of English.

    A separate call is being sent out for the shared task. Participants must register by JANUARY 20, 2007. The systems and their results will be described in short papers, which will be refereed separately and will be presented in a special session at EMNLP-CoNLL 2007.

    More information, including the call for participation, is available at the shared task website http://depparse.uvt.nl/depparse-wiki/SharedTaskWebsite . Questions can be sent to conll07st at uvt dot nl.

    SUBMISSIONS

    FORMAT: Please submit your complete paper as a PDF file. Submissions should follow the two-column format of ACL proceedings and should not exceed 8 pages, including figures. We strongly recommend using the ACL 2007 style files at http://ufal.mff.cuni.cz/acl2007/styles/. If you cannot use these style files directly, please read the sample document at that URL for a description of the required format. We reserve the right to reject submissions that do not conform to this format; note especially the restriction to 8 pages at 11-point font.

    NEW LENGTH POLICY: To encourage thorough citation of related work, the References section DOES NOT count against the 8-page submission limit for EMNLP-CoNLL 2007. Thus, your PDF file may exceed 8 pages. However, all material other than the bibliography must fall within the first 8 pages!

    ANONYMIZATION: As reviewing will be double-blind, the paper should not include the authors' names and affiliations. Furthermore, avoid self-references that reveal the authors' identities, e.g., "We previously showed (Smith, 1991) ..." Instead, use citations such as "Smith previously showed (Smith, 1991)...". Papers that do not conform to these requirements will be rejected without review.

    DOUBLE SUBMISSION POLICY: Papers presented at EMNLP-CoNLL should consist mainly of new material that has not been previously published. Submission of similar papers to EMNLP-CoNLL and another conference or workshop must be disclosed on the first page of the submission. For details, see the standard policy for ACL-affiliated conferences, http://www.cis.udel.edu/~carberry/ACL/double-submission-policy.html .

    ELECTRONIC SUBMISSION: To submit your paper, please visit http://www.softconf.com/acl07/EMNLP-CoNLL07/submit.html , fill in the online submission form, and upload your anonymous PDF file. The only accepted file format is PDF. You may revise your submission up until the deadline given below. Papers submitted after the deadline will not be reviewed.

    REVIEWING

    Reviewers will be asked to provide detailed comments, and to score the submitted paper on a number of factors such as

    • technical soundness and completeness
    • originality of the problems or solutions
    • experimental and theoretical comparison with previous work
    • clarity of presentation
    • significance to the EMNLP-CoNLL research community

    as well as a new factor, which encourages authors to release their systems so that others can replicate or build on their results:

    • likely impact of any promised new resources to be released along with the paper (code, data, ...)

    Reviewing will be conducted by an international program committee with a number of Area Chairs. Each paper will be blind-reviewed by at least 3 program committee members, who can then discuss the paper further to resolve disagreements.

    The program committee will make an Best Paper Award to the authors of the submission that makes the most original or significant impact on the field. We seek to honor scientific contributions that advance the state of the field.

    CONFERENCE FORMAT

    The majority of the accepted papers will be presented as 20-minute talks, probably in 2 parallel sessions. Additional papers will be accepted for presentation in poster sessions. We also plan additional events such as invited talks.

    MAIN ORGANIZERS

    Program Chairs:
    Jason Eisner (Johns Hopkins University)
    Taku Kudo (Google Japan)

    Shared Task Organizer:
    Joakim Nivre (Vaxjo University)

    IMPORTANT DATES

    Paper submission deadline: Mon, March 26, 2007, 11:59pm EDT (GMT -0400)
    Notification of acceptance: Thu, April 26, 2007
    Camera-ready copy due: Mon, May 14, 2007
    Conference meeting: Thu, June 28 - Sat, June 30, 2007

    NEWS AND CONTACT INFORMATION

    Updates will be posted at the conference home page: http://cs.jhu.edu/EMNLP-CoNLL-2007
    Questions may be sent to the program chairs at this address: emnlp-conll-2007 at clsp dot jhu dot edu

 
Close Window