/----------------------------------------------------------------------------\
|                                                                            |
|                               DELiC4MT                                     |
| Diagnostic Evaluation using Linguistic Checkpoints For Machine Translation |
|                                                                            |
\----------------------------------------------------------------------------/

VERSION 20130612


DELiC4MT is a piece of software that allows to perform diagnostic evaluation of Machine Translation systems over linguistic checkpoints, i.e. source-language lexical elements and grammatical constructions specified by the user. For more details see our paper in the Credits section.



CHANGELOG
---------
20130612 html output added to Java program (for web application)
20120809 bugfix in FileNGramMatcher.java
20120223 Penalty based on brevity added to Java program
20120207 Improved treetagger_preserving_tokens_and_lines.pl
20121215 Added tutorial
20111213 Added script for filtering checkpoints
20111129 Added scripts for statistical significance, script that wraps all the stages (delic4mt.sh) and a test folder with 3 use cases (English to Dutch, German and Italian)
20110724 First release



CONTENTS
--------

doc/		Documentation, including a tutorial and related papers
evaluate/	Java program that carries out the evaluation on linguistic checkpoints
scripts/	Support scripts
test/		Sample test data



CREDITS
-------

Code by Sudip Kumar Naskar and Antonio Toral
Logo by Pawel Plesniak


If you use DELiC4MT in your research please cite the following paper:

Antonio Toral, Sudip Kumar Naskar, Federico Gaspari, Declan Groves. DELiC4MT: A Tool for Diagnostic MT Evaluation over User-defined Linguistic Phenomena. The Prague Bulletin of Mathematical Linguistics No 98, 2012, pp. 121-131., ISSN 0032-6585, DOI: 10.2478/v10108-012-0014-9.

@Article{pbml-2012-delic4mt,
  author    = {Antonio Toral and Sudip Kumar Naskar and Federico Gaspari and Declan Groves},
  title     = {{DELiC4MT: A Tool for Diagnostic MT Evaluation over User-defined Linguistic Phenomena}},
  journal   = {The Prague Bulletin of Mathematical Linguistics},
  issue     = {98},
  publisher = {Springer Berlin/Heidelberg},
  issn      = {0032-6585},
  pages     = {121-132},
  year      = {2012}
}


CONTACT
-------

For any question or suggestion please contact:

Antonio Toral, Sudip Naskar
atoral, snaskar #-at-# computing dot dcu dot ie