Ivan A. Sag (sag@skol.stanford.edu)
Mon, 6 Jun 94 15:46:54 PDT
All, Stuart Shieber recently suggested to me that we all use The Computation and Language E-Print Archive for electronic distribution of HPSG papers. This is a great way to get research results widely distributed (well before journals catch up with research) and also to get fast reactions from colleagues. I intend to send a couple of recently finished papers there soon, instead of making them available via anonymous ftp from Stanford. This archive makes papers available as latex source files, .dvi files or postscript files. I enclose a message telling how to get more information about subscribing to their list. Subscription means you will be notified every time new papers become available. Best, Ivan --------------- Date: Mon, 6 Jun 1994 02:03:31 -0600 Precedence: bulk Note: e-print archive software written by PG at LANL (8/91,12/92,11/93) GTDA From: no-reply@xxx.lanl.gov (send mail ONLY to cmp-lg) Reply-To: cmp-lg@xxx.lanl.gov To: rabble@xxx.lanl.gov (cmp-lg daily title/abstract distribution) Subject: cmp-lg daily 9406011 -- 9406013 received (634 served) ****************************************************************************** The Computation and Language E-Print Archive CMP-LG@XXX.LANL.GOV TODAY'S LISTINGS AND NEWS ****************************************************************************** TITLE/AUTHOR/ABSTRACT LISTINGS received from Fri Jun 3 02:13:25 MDT 1994 to Mon Jun 6 01:54:08 MDT 1994 ------------------------------------------------------------------------------ \\ Paper: cmp-lg/9406011 From: Lance Ramshaw <ramshaw@linc.cis.upenn.edu> Date: Fri, 3 Jun 1994 14:02:41 -0400 Date (revised): Fri, 3 Jun 1994 14:46:15 -0400 Title: Exploring the Statistical Derivation of Transformational Rule Sequences for Part-of-Speech Tagging Author: Lance A. Ramshaw (Univ. of Pennsylvania and Bowdoin College), and Mitchell P. Marcus (Univ. of Pennsylvania) Comments: 10 pages, in proceedings of the ACL Balancing Act workshop \\ Eric Brill has recently proposed a simple and powerful corpus-based language modeling approach that can be applied to various tasks including part-of-speech tagging and building phrase structure trees. The method learns a series of symbolic transformational rules, which can then be applied in sequence to a test corpus to produce predictions. The learning process only requires counting matches for a given set of rule templates, allowing the method to survey a very large space of possible contextual factors. This paper analyses Brill's approach as an interesting variation on existing decision tree methods, based on experiments involving part-of-speech tagging for both English and ancient Greek corpora. In particular, the analysis throws light on why the new mechanism seems surprisingly resistant to overtraining. A fast, incremental implementation and a mechanism for recording the dependencies that underlie the resulting rule sequence are also described. \\ (cmp-lg/9406011 , 150kb) ------------------------------------------------------------------------------ \\ Paper: cmp-lg/9406012 From: Patrick Juola <juola@suod.cs.colorado.edu> Date: Fri, 3 Jun 1994 14:55:37 -0600 Title: Self-Organizing Machine Translation: Example-Driven Induction of Transfer Functions Author: Patrick Juola (University of Colorado) Comments: PostScript, 30 pages, contact author for LaTeX version Report-no: CU-CS-722-94 \\ With the advent of faster computers, the notion of doing machine translation from a huge stored database of translation examples is no longer unreasonable. This paper describes an attempt to merge the Example-Based Machine Translation (EBMT) approach with psycholinguistic principles. A new formalism for context- free grammars, called *marker-normal form*, is demonstrated and used to describe language data in a way compatible with psycholinguistic theories. By embedding this formalism in a standard multivariate optimization framework, a system can be built that infers correct transfer functions for a set of bilingual sentence pairs and then uses those functions to translate novel sentences. The validity of this line of reasoning has been tested in the development of a system called METLA-1. This system has been used to infer English->French and English->Urdu transfer functions from small corpora. The results of those experiments are examined, both in engineering terms as well as in more linguistic terms. In general, the results of these experiments were psycho- logically and linguistically well-grounded while still achieving a respectable level of success when compared against a similar prototype using Hidden Markov Models. \\ (cmp-lg/9406012 , 155kb) ------------------------------------------------------------------------------ \\ Paper: cmp-lg/9406013 From: alkim@unagi.cis.upenn.edu (Albert E Kim) Date: Sun, 5 Jun 1994 13:00:57 -0400 (EDT) Title: Graded Unification: A Framework for Interactive Processing Author: Albert Kim (University of Pennsylvania) Comments: 3 pages, To appear in ACL-94 Student Session, Postscript file; extract with Unix uudecode and uncompress \\ Abstract: An extension to classical unification, called {\em graded unification} is presented. It is capable of combining contradictory information. An interactive processing paradigm and parser based on this new operator are also presented. \\ (cmp-lg/9406013 , 79kb) %-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%-%- %%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%%--%% ****************************************************************************** CMP-LG NEWS Documents of potential interest to cmp-lg subscribers: cmp-lg http://xxx.lanl.gov/cmp-lg/ cmp-lg FAQ "get faq" command or http://xxx.lanl.gov/cmp-lg/faq.html cmp-lg "get statistics" command or statistics http://xxx.lanl.gov/ftp/cmp-lg/papers/macros/statistics ACL http://www.cs.columbia.edu/~acl/home.html ACL-94 http://crl.nmsu.edu/acl94/Home.html JAIR http://www.cs.washington.edu/research/jair/home.html Please help SPREAD THE WORD by forwarding this message to your colleagues, so they, like the 600 current subscribers, can get the benefit of early notification of research results in computation and language. (Forwarded recipients: See below on how to get further information on subscribing.) For FURTHER INFORMATION about (among other topics) submission of papers to the server, subscribing or canceling your subscription, requesting full text of any of the papers below, retrieving macro files for these papers, searching past listings, or submitting comments to the server operators, send a message: To: CMP-LG@XXX.LANL.GOV Subject: help [Revision of Tue May 31 14:44:27 1994.] ******************************************************************************
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:31:39 PST