linguistlist.org>
Editor for this issue: Julie Wilson <julielinguistlist.org>
cogsci.ed.ac.uk>The Fourth CFP ===================================== <<C A L L ----- F O R ----- P A P E R S>> ===================================== The Computation of Phonological Constraints =========================================== The 4th Meeting of the ACL Special Interest Group in Phonology (http://www.cogsci.ed.ac.uk/~sigphon/98) In conjunction with The COLING-ACL'98 Joint Conference Montreal, Canada, 15th August 1998 *** Due to popular request, we have extended the deadline for *** submissions to this workshop by 3 weeks. The new deadline *** is MAY 10. The Meeting - --------- Constraint-based theories of phonology have become enormously popular in recent years. Such theories express generalizations by stating how and when a language's phonological forms are constrained, rather than relying on rules that actively modify the forms. Computational ideas have often provided an impetus for these innovations. Koskenniemi's (1983) 2-level morphophonological processor introduced parallel constraining relations as an alternative to an ordered system of rewrite rules. Declarative Phonology (c 1990) focusses on the unity of representations and rules as constraints, drawing on various computational ideas from unification to temporal logic to finite-state calculi. One current constraint-based theory is Optimality Theory (OT) (Prince&Smolensky 1993). This theory found, at least partial, inspiration in computational work by Smolensky on the relation between symbolic and subsymbolic computation. Although this link with connectionism has been left largely unexplored, OT has proved to be a computationally productive theory, giving rise to several theoretical papers on computational issues related to complexity and learnability, as well as inspiring a number of implementations. This workshop is designed to foster the link between computational work and constraint-based phonology in general. To this end, it invites submissions on topics related to the computation of any constraint-based phonological formalism, including but not limited to the three mentioned above. Here are some example topics: * the computational interpretation of phonological theories, * constraint ranking and interaction, eg. as in OT, * implementations of particular analyses, * results in the complexity of constraint application, * algorithms for learning constraints or constraint ranking, * results on the learnability of such constraints, * novel formalisms for constraint-based phonology, * representational issues raised by constraint-based approaches. In short, papers are invited which address computational issues in constraint-based theories of phonology. Submission - -------- What: original research, not published elsewhere a completed study is prefered to proposals and progress reports originality, topicality and clarity will be the assessment criteria How: submissions must be sent by email to sigphon98Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue
wlv.ac.uk>FINAL CALL FOR PAPERS THE MACHINE TRANSLATION JOURNAL SPECIAL ISSUE ON ANAPHORA RESOLUTION IN MACHINE TRANSLATION Guest editor: Ruslan Mitkov (University of Wolverhampton) The interpretation of anaphora is crucial for the successful operation of a Machine Translation system. In particular, it is essential to resolve the anaphoric relation when translating into languages which mark the gender of pronouns. Unfortunately, the majority of MT systems developed in the seventies and eighties did not adequately address the problems of identifying the antecedents of anaphors in the source language and producing the anaphoric "equivalents" in the target language. As a consequence, only a limited number of MT systems have been successful in translating discourse, rather than isolated sentences. One reason for this situation is that in addition to anaphora resolution being itself a very complicated task, translation adds a further dimension to the problem in that the reference to a discourse entity encoded by a source language anaphor by the speaker (or writer) has not only to be identified by the hearer (translator or translation system) but also re-encoded in a coreferential expression in a different language. The nineties have seen an intensification of research efforts in anaphora resolution in Machine Translation. This can be seen in the growing number of related projects which have reported promising new results (e.g.Wada 1990; Leass & Schwall 1991; Nakaiwa & Ikehara 1992; Chen 1992; Saggion & Carvalho 1994; Preuss et al. 1994; Nakaiwa et al. 1994; Nakaiwa et al. 1995; Nakaiwa & Ikehara 1995; Mitkov et al. 1995; Mitkov et al. 1997; Geldbach 1997). However, we still feel that additional work is needed to highlight and further explore the specifics of the problem in operational MT environments, including fully automatic Machine Translation and Machine- aided Translation. We are inviting high-quality, original research papers describing recent advances in anaphora resolution in Machine Translation. Topics to be addressed include (but are not limited to) - operational anaphora resolution components in Machine Translation - resolution of zero pronouns in MT environments - lexical transfer of anaphors across languages - to what extent have the latest trends towards knowledge-poor, corpus- driven and robust approaches in anaphora resolution, been called upon in Machine Translation? - what are the most scalable contributory factors /resolution strategies in MT? - what makes anaphora resolution a more complex task in Machine Translation? - multilingual anaphora resolution SUBMISSION AND FORMAT Articles should be submitted directly to the publishers, either by e-mail to Ellen.KlinkMail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue