5-1372msg2.txt

运用贝叶斯网编的一个识别垃圾邮件的程序

Subject: call for paper - acl-95 corpus workshop

acl 's sigdat present the third workshop on very large corpora preliminary call for paper when : june 30 , 1995 - immediately follow acl-95 ( june 27-29 ) where : mit , cambridge , massachusett , usa workshop description : as in past year , the workshop will offer a general forum for new research in corpus-base and statistical natural language process . area of interest include ( but be not limit to ) : sense disambiguation , part-of - speech tag , robust parse , term and name identification , alignment of parallel text , machine translation , lexicography , spell correction , morphological analysis and anaphora resolution . this year , the workshop will be organize around the theme of : supervise train v . self - organize method is annotation worth the effort ? historically , annotate corpus have make a significant contribution . the tag brown corpus , for example , lead to important improvement in part-of - speech tag . but annotate corpus be expensive . very little annotate datum be currently available , especially for language other than english . self - organize method offer the hope that annotate corpus may not be necessary . do these method really work ? do we have to choose between annotate corpus and unannotate corpus ? can we use both ? the workshop will encourage contribution of innovative research along this spectrum . in particular , it will seek work in language other than english and in application where appropriately tag train corpus do not exist . it will also explore what new kind of corpus annotation ( such as discourse structure , co-reference and sense tag ) would be useful to the community , and will encourage papers on their development and use in experimental project . the theme will provide an organize structure to the workshop , and offer a focus for debate . however , we expect and will welcome a diverse set of submission in all area of statistical and corpus-base nlp . program chairs : ken church - at&t bell laboratory david yarowsky - university of pennsylvanium sponsor : sigdat ( acl 's special interest group for linguistic datum and corpus-base approach to nlp ) format for submission : author should submit a full-length paper ( 3500-8000 word ) , either electronically or in hard-copy . electronic submission must either be plain ascius text or a single latex file follow the acl-95 stylesheet ( no separate figure or . bib file ) . hard copy submission should include four ( 4 ) copy of the paper . author should consult the primary call for papers in late january for update specification . schedule : submission deadline : march 10 , 1995 notification deat : april 10 , 1995 camera ready copy due : may 10 , 1995 contact : ken church david yarowsky room 2b-421 dept . of computer and info . science at&t bell laboratory university of pennsylvanium 600 mountain ave . 200 s . 33rd st . murray hill , nj 07974 usa philadelphium , pa 19104-6389 usa e-mail : kwc @ research . att . com email : yarowsky @ unagus . ci . upenn . edu