University of Brighton home page

     University of Trento home page

Copenhagen Business School Building   
"ITRI" in a range of writing systems "ITRI" in a range of writing systems "ITRI" in a range of writing systems

  UCNLG+MT at MT Summit XI




Call for papers

Key facts

Important dates

MT Summit XI

Shared Task



Using Corpora for NLG: Language Generation and Machine Translation (UCNLG+MT) is a pre-conference workshop at MT Summit XI. It will be held in Copenhagen on 11 September 2007.

Workshop aims

There are many branches of NLP research where language is generated (MT, summarisation, human-computer dialogue, data-to-text generation, document authoring, etc.). However, for the most part, these form separate research communities and the field called "Natural Language Generation (NLG)" is to some degree isolated from many of them.

The UCNLG workshops have the general aims

  1. to provide a forum for reporting and discussing data-based methods for generating language;
  2. to open up the NLG research field to neighbouring areas in NLP and to foster cross-fertilisation between these research fields; and
  3. to promote the sharing of data in NLG and the use of data-based methods of evaluation.

The special theme of this second UCNLG workshop is Language Generation and MT. Target language generation is typically given less attention than other aspects of MT, such as evaluation, source language analysis and transfer. Recently there has been increasing interest in combining linguistic knowledge with statistical MT. Conversely, NLG -- where symbolic approaches have long dominated -- has seen increasing interest in probabilistic methods. With this special theme, would like to explore crossover points and potential for cross-fertilisation between MT and NLG.

The workshop also continues from UCNLG'05 in its general focus on data-based methods for language generation and evaluation. Please see the Call for Papers for further details.

Workshop programme

We have now added links to the presentation slides to the programme below.

08:45 Welcome and Introduction
Language Generation Session
09:00 Towards Broad Coverage Surface Realization with CCG [Slides]
Michael White, Rajakrishnan Rajkumar, Scott Martin
09:30 Declarative Syntactic Processing of Natural Language Using Concurrent Constraint Programming and Probabilistic Dependency Modeling [Slides]
Irene Langkilde-Geary
09:50 Generalizing Syntactic Collocates for Creative Language Generation [Slides]
David Hardcastle
10:10 Induction and Evaluation of Word Order Rules Using Corpora Based on the Two Concepts of Topological Models [Slides]
Bernd Bohnet
10:30 Coffee/Tea
Shared Task Session
10:50 Introduction to the Attribute Selection for GRE Challenge [Slides]
Anja Belz
11:05 Presentations of Systems submitted to Challenge
Bernd Bohnet [Slides]
Raquel Herv↓ Pablo Gerv¢[Slides]
John Kelleher [Slides]
Advaith Siddharthan, Ann Copestake
Mariet Theune, Pascal Touset, Jette Viethen, Emiel Krahmer [Slides]
Philipp Spanger, Takahiro Kurosawa, Takenobu Tokunaga [Slides]
11:35 Anja Belz: Evaluation Results [Slides]
Machine Translation Session
12:00 Method of Selecting Training Sets to Build Compact and Efficient Statistical Language Model
Keiji Yasuda, Hirofumi Yamamoto, Eiichiro Sumita
12:30 Lunch
Machine Translation Session (cont.)
13:30 Automatic Evaluation of Generation and Parsing for Machine Translation with Automatically Acquired Transfer Rules [Slides]
Yvette Graham, Deirdre Hogan, Josef van Genabith
14:00 A Probabilistic Approach to Linguistic Analysis in Machine Translation Output Evaluation [Slides]
Olivier Gouirand
14:20 "Automatic Language Translation Generation Help Needs Badly" [Slides]
Kevin Knight
15:10 Coffee/Tea Break
Discussion Session on Language Generation and Machine Translation
15:30 Introduction and Aims
15:40 Discussion Starter Presentations
Andrei Popescu-Belis [Slides]
Nizar Habash [Slides]
Gregor Thurmair
Sebastian Varges [Slides]
17:00 Wrapping up

Key workshop facts

Invited speaker:

Kevin Knight, ISI, University of Southern California, USA

Discussion Session on NLG+MT:

Submission of extended abstracts invited, for details see Call for Papers.

Programme committee:

Srinivas Bangalore, AT&T, USA
Stephan Busemann, DFKI, Germany
Robert Dale, Macquarie University, Australia
Kevin Knight, ISI, University of Southern California, USA
Irene Langkilde-Geary, University of Brighton, UK
Chris Mellish, University of Aberdeen, UK
Jon Oberlander, University of Edinburgh, UK
Richard Power, Open University, UK
Flo Reeder, The Mitre Organisation, USA
Ehud Reiter, University of Aberdeen, UK
Amanda Stent, SUNY, USA
Michael Strube, EML Research, Germany
Kees van Deemter, University of Aberdeen, UK
Bonnie Webber, University of Edinburgh, UK
Mike White, Ohio State University, USA

Important dates

15 June: Extended submission deadline for workshop papers
29 June: Notification of acceptance to authors of workshop papers
01 July: Submission deadline for extended abstracts for
Discussion Session
09 July: Notification of acceptance to authors of extended
abstracts for Discussion Session
20 July: Camera-ready copy from all authors
11 September: UCNLG+MT in Copenhagen

Site News

19 September 2007
Links to presentation slides have been added to the workshop programme.

5 September 2007
Workshop programme and proceedings available.

12 July 2007
List of accepted contributions for discussion session available.

2 July 2007
List of accepted workshop papers available.

4 June 2007
Submission deadline for workshop papers extended from 8 June to 15 June.

1 May 2007
Attribute Selection for GRE Challenge announced.

March 2007
Call for papers available.

UCNLG+MT is organised by Anja Belz, NLTG, University of Brighton, and Sebastian Varges, DIT, University of Trento, and is endorsed by SIGGEN, the special interest group on generation of the Association for Computational Linguistics (ACL).

back to top

Last modified: Wed Oct 3 09:51:44 BST 2007
Maintained by: