Senseval 3
Third International Workshop on the Evaluation of Systems
for the Semantic Analysis of Text
An ACL 2004 workshop, Barcelona, Spain, July 25-26, 2004
Evaluations: March-April 2004
Organized by ACL-SIGLEX
|
|
|
|
Proceedings
Senseval-3 proceedings
Panels
Senseval-3 panels
Program
The Senseval-3 workshop program is now available.
Program highlights
Invited speaker: Kenneth CHURCH, Microsoft Research, July 25.
Panel on "Applications of WSD" Bill Dolan (Microsoft), Phil Edmonds (Sharp Labs), Dekang Lin (U. Alberta/Google), Carlo Strapparava (IRST). More to be confirmed. July 25.
Panel on "Planning Senseval-4" Ido Dagan (Bar-Ilan U.), Adam Killgariff (U. Brighton), Lluis Marquez (U. Politecnica Catalunya), Rada Mihalcea (U. North Texas), Martha Palmer (U. Pennsylvania), Philip Resnik (U. Maryland). July 26.
|
|
|
|
|
|
Information for Authors Presenting a Paper at Senseval-3
The poster session is planned for July 26, 10:30-12:00pm. If you are
doing an oral presentation, please check the program for the time of your
presentation.
Posters
You will have a 90x120cm board to display the poster. Posters
will have to be put in place right before the poster session, and removed
after the session is finished. It is likely that we will be using the
workshop room for the poster session, where chairs will be
rearranged to make enough room for the posters. Please plan to be in the
workshop room at 10am with your poster (we will only have 30' to get the
poster session in place).
Oral Presentations
We expect to have the workshop room equipped with
the usual facilities (overhead/LCD projector). The length of each oral
presentation is 15'. We strongly encourage you to leave enough time for
questions (e.g. plan for a 11-12' presentation, and leave some 3-4' for
questions). Since this is not a regular workshop, but a series of
presentations of tasks and systems participating in an evaluation exercise
- we believe it is very important to encourage discussions around each
task/system, to gain most of these evaluations.
|
|
|
|
|
|
Call for Participation in the Senseval-3 Evaluations
The main purpose of this workshop is to analyze and discuss the results
of systems participating in the Senseval-3 evaluations, to be held in
March-April 2004. Fourteen different tasks are planned for Senseval-3,
to conduct evaluations of systems that perform automatic semantic analysis
of text, including: word sense disambiguation for various languages,
identification of semantic roles, logic forms, multilingual annotations,
subcategorization acquisition.
This is an advance notice of the evaluation exercise and workshop.
Registration for the evaluation will open in February (watch the website
for updates). Papers will be accepted from participants only.
Background
There are now many computer systems that do automatic semantic analysis of
text. The purpose of Senseval is to evaluate the strengths and weaknesses
of such systems with respect to different words, relations, types of texts,
different varieties of language, and different languages.
This workshop is a follow-up to Senseval-1 and Senseval-2.
Senseval-1 took place in the summer of 1998 for English, French, and Italian,
culminating in a workshop held at Herstmonceux Castle, Sussex, England on
September 2-4.
Senseval-2 took place in the summer of 2001, and was followed by a workshop
held in July 2001 in Toulouse, in conjunction with ACL-2001. Senseval-2
included tasks for Basque, Chinese, Czech, Danish, Dutch, English, Estonian,
Italian, Japanese, Korean, Spanish, Swedish.
Tasks
The following tasks are planned for Senseval-3 (see Tasks for a detailed description):
- English all words
- Italian all words
- Basque lexical sample
- Catalan lexical sample
- Chinese lexical sample
- English lexical sample
- Italian lexical sample
- Romanian lexical sample
- Spanish lexical sample
- Automatic subcategorization acquisition
- Multilingual lexical sample
- WSD of WordNet glosses
- Semantic Roles
- Logic Forms
This 2-day workshop will consist of several Senseval-3 task and system
presentations, including analyses of results obtained during the
evaluations, with comparisons across different systems, techniques,
and languages. We also plan for two panels on (1) the interaction
between systems for semantic analysis of text and other NLP applications,
and (2) planning Senseval-4.
|
|
|
|