Difference between revisions of "2010Q3 Reports: SIGNLL"

From Admin Wiki
Jump to navigation Jump to search
 
(9 intermediate revisions by the same user not shown)
Line 1: Line 1:
=== ACL SIGNLL Annual Report (2009-2010) by Joakim Nivre and Lluís Màrquez ===
+
= ACL SIGNLL Annual Report (2009-2010)=
 +
 
 +
'''by Joakim Nivre and Lluís Màrquez'''
  
 
The goals of SIGNLL, ACL's special interest group on natural language
 
The goals of SIGNLL, ACL's special interest group on natural language
Line 31: Line 33:
 
Antal van den Bosch, Claire Cardie, Walter Daelemans, Hwee Tou Ng,
 
Antal van den Bosch, Claire Cardie, Walter Daelemans, Hwee Tou Ng,
 
David Powers, and Dan Roth; and the larger SIGNLL International
 
David Powers, and Dan Roth; and the larger SIGNLL International
Advisory Board (see http://ifarm.nl/signll/about/#officers for a
+
Advisory Board (see http://ifarm.nl/signll/about/#officers, for a
 
complete description of SIGNLL officers and boards). David Yarowsky
 
complete description of SIGNLL officers and boards). David Yarowsky
 
acts as the SIGDAT Liaison Representative.
 
acts as the SIGDAT Liaison Representative.
  
== Main Activities 2009-2010 ==
+
=== CoNLL-2010 ===
 +
 
 +
The Fourteenth SIGNLL Conference on Computational Natural Language
 +
Learning, CoNLL-2010, will be held in Uppsala, Sweden, 15-16 July
 +
2010, in conjunction with ACL 2010. Program co-chairs are Anoop
 +
Sarkar and Mirella Lapata. The special topic of interest of the 2010
 +
edition is Grammar Induction, including varied perspectives (machine
 +
learning, natural language engineering and cognitive approaches) and
 +
aiming at integrating works form several areas of research
 +
(statistical methods, pattern recognition, neural networks,
 +
computational linguistics, computational learning theory, automata
 +
theory, and language acquisition).
 +
 
 +
CoNLL 2010 received 99 submissions, of which 18 were eventually
 +
withdrawn. Of the remaining 81 papers, 12 were selected to appear in
 +
the conference programme as oral presentations, and 13 were chosen
 +
as posters.  The invited speakers are Lillian Lee (Cornell
 +
University) and Zoubin Gharamani (University of Cambridge) and their
 +
talks will cover knowledge-lean information extraction and Bayesian
 +
HMMs, respectively.
  
= CoNLL 2010 =
+
As in 2009, the CoNLL-2010 best paper award is sponsored by
  The Fourteenth SIGNLL Conference on Computational Natural Language
+
Google. More information can be found at the conference website:
  Learning, CoNLL-2010, will be held in Uppsala, Sweden, 15-16 July
+
http://www.cnts.ua.ac.be/conll2010/.
  2010, in conjunction with ACL 2010. Program co-chairs are Anoop
 
  Sarkar and Mirella Lapata. The special topic of interest of the 2010
 
  edition is Grammar Induction, including varied perspectives (machine
 
  learning, natural language engineering and cognitive approaches) and
 
  aiming at integrating works form several areas of research
 
  (statistical methods, pattern recognition, neural networks,
 
  computational linguistics, computational learning theory, automata
 
  theory, and language acquisition).
 
  
  CoNLL 2010 received 99 submissions, of which 18 were eventually
+
=== CoNLL-2010 Shared Task ===
  withdrawn. Of the remaining 81 papers, 12 were selected to appear in
 
  the conference programme as oral presentations, and 13 were chosen
 
  as posters.  The invited speakers are Lillian Lee (Cornell
 
  University) and Zoubin Gharamani (University of Cambridge) and their
 
  talks will cover knowledge-lean information extraction and Bayesian
 
  HMMs, respectively.
 
  
  As in 2009, the CoNLL-2010 best paper award is sponsored by
+
As in previous years, CoNLL-2010 has a shared
  Google. More information can be found at the conference website
+
task, "Learning to detect hedges and their scope in natural language
  http://www.cnts.ua.ac.be/conll2010/
+
text". The shared task is completely new and was selected from a set
 +
of 4 proposals submitted to the SIGNLL call for shared tasks in
 +
September 2009. It addresses the detection of uncertainty in two
 +
domains, using the BioScope corpus (biological scientific articles)
 +
and the public Wikipedia weasel annotations. Two uncertainty
 +
detection tasks were proposed: 1) detecting sentences containing
 +
uncertainty, and 2) resolving the in-sentence scope of hedge cues.
 +
A total of 23 teams have participated in the shared task, which can
 +
be considered a real success, given the novelty of the topic. Those
 +
who participated in both tasks were invited to write a 8 page
 +
paper. The page limit for those who participated only in the first
 +
task was 6 pages. The Shared Task papers are collected into an
 +
accompanying volume of CoNLL-2010.
  
* CoNLL-2010 Shared Task As in previous years, CoNLL-2010 has a shared
+
The task maintains its own website where all the details (including
  task, "Learning to detect hedges and their scope in natural language
+
datasets and results) can be found at: http://www.inf.u-szeged.hu/rgai/conll2010st.
  text". The shared task is completely new and was selected from a set
 
  of 4 proposals submitted to the SIGNLL call for shared tasks in
 
  September 2009.  It addresses the detection of uncertainty in two
 
  domains, using the BioScope corpus (biological scientific articles)
 
  and the public Wikipedia weasel annotations. Two uncertainty
 
  detection tasks were proposed: 1) detecting sentences containing
 
  uncertainty, and 2) resolving the in-sentence scope of hedge cues.
 
  A total of 23 teams have participated in the shared task, which can
 
  be considered a real success, given the novelty of the topic. Those
 
  who participated in both tasks were invited to write a 8 page
 
  paper. The page limit for those who participated only in the first
 
  task was 6 pages. The Shared Task papers are collected into an
 
  accompanying volume of CoNLL- 2010.
 
  
  The task maintains its own website where all the details (including
+
=== Other Activities ===
  datasets and results) can be found:
 
  http://www.inf.u-szeged.hu/rgai/conll2010st
 
  
* SIGNLL is also endorsing the Workshop "Negation and Speculation in
+
* SIGNLL is also endorsing the Workshop "Negation and Speculation in Natural Language Processing (NeSpNLP 2010)", organized by Roser Morante and Caroline Sporleder. The workshop will be held next July in Uppsala, just before the ACL 2010 events. Workshop website: http://www.clips.ua.ac.be/NeSpNLP2010/index.html.
  Natural Language Processing (NeSpNLP 2010)", organized by Roser
 
  Morante and Caroline Sporleder. The workshop will be held next July
 
  in Uppsala, just before the ACL 2010 events. Workshop website
 
  http://www.clips.ua.ac.be/NeSpNLP2010/index.html
 
  
* In July-August 2010, SIGNLL will launch a call for shared task
+
* In July-August 2010, SIGNLL will launch a call for shared task proposals to select the task to be run in CoNLL-2011. A designated committee will judge the submitted proposals and declare the selected task no later than October 2010.
  proposals to select the task to be run in CoNLL-2011. A designated
 
  committee will judge the submitted proposals and declare the
 
  selected task no later than October 2010.
 

Latest revision as of 02:44, 15 June 2010

ACL SIGNLL Annual Report (2009-2010)

by Joakim Nivre and Lluís Màrquez

The goals of SIGNLL, ACL's special interest group on natural language learning, are to promote and inform about research on computational modeling of learning in natural language. These are served by (i) the maintenance of an informative and up-to-date website and associated mailing list, and (ii) the organization of annual events (the CoNLL conference and the CoNLL shared task), and support of other related activities.

The web-pages, located at URL http://www.aclweb.org/signll/ and maintained by Erik Tjong Kim Sang, remain an important source of information, complemented by an email list for announcements for SIGNLL-related events. On the web-site, links can be found to relevant associations, networks, research cooperations, research departments, groups, institutes, mailing lists, archives, journals, bulletins, conference reports, online papers (including all papers of all CoNLL proceedings), online courses and slides, bibliographies, software, corpora, companies, meta-information sources, etc.

Currently, SIGNLL has 326 members, with ~22 new registrations in 2009. An election for SIGNLL officers was organized after the summer of 2009. New president is Joakim Nivre (Uppsala University, Sweden) and secretary Lluís Màrquez (Technical University of Catalonia, Spain). The current office period runs from November 1, 2009 to October 31, 2011. Our Information Officer remains Erik Tjong Kim Sang.

Apart from the officers, SIGNLL also has two consultative committees. The SIGNLL Steering Committee, composed by all past SIGNLL officers: Antal van den Bosch, Claire Cardie, Walter Daelemans, Hwee Tou Ng, David Powers, and Dan Roth; and the larger SIGNLL International Advisory Board (see http://ifarm.nl/signll/about/#officers, for a complete description of SIGNLL officers and boards). David Yarowsky acts as the SIGDAT Liaison Representative.

CoNLL-2010

The Fourteenth SIGNLL Conference on Computational Natural Language Learning, CoNLL-2010, will be held in Uppsala, Sweden, 15-16 July 2010, in conjunction with ACL 2010. Program co-chairs are Anoop Sarkar and Mirella Lapata. The special topic of interest of the 2010 edition is Grammar Induction, including varied perspectives (machine learning, natural language engineering and cognitive approaches) and aiming at integrating works form several areas of research (statistical methods, pattern recognition, neural networks, computational linguistics, computational learning theory, automata theory, and language acquisition).

CoNLL 2010 received 99 submissions, of which 18 were eventually withdrawn. Of the remaining 81 papers, 12 were selected to appear in the conference programme as oral presentations, and 13 were chosen as posters. The invited speakers are Lillian Lee (Cornell University) and Zoubin Gharamani (University of Cambridge) and their talks will cover knowledge-lean information extraction and Bayesian HMMs, respectively.

As in 2009, the CoNLL-2010 best paper award is sponsored by Google. More information can be found at the conference website: http://www.cnts.ua.ac.be/conll2010/.

CoNLL-2010 Shared Task

As in previous years, CoNLL-2010 has a shared task, "Learning to detect hedges and their scope in natural language text". The shared task is completely new and was selected from a set of 4 proposals submitted to the SIGNLL call for shared tasks in September 2009. It addresses the detection of uncertainty in two domains, using the BioScope corpus (biological scientific articles) and the public Wikipedia weasel annotations. Two uncertainty detection tasks were proposed: 1) detecting sentences containing uncertainty, and 2) resolving the in-sentence scope of hedge cues. A total of 23 teams have participated in the shared task, which can be considered a real success, given the novelty of the topic. Those who participated in both tasks were invited to write a 8 page paper. The page limit for those who participated only in the first task was 6 pages. The Shared Task papers are collected into an accompanying volume of CoNLL-2010.

The task maintains its own website where all the details (including datasets and results) can be found at: http://www.inf.u-szeged.hu/rgai/conll2010st.

Other Activities

  • SIGNLL is also endorsing the Workshop "Negation and Speculation in Natural Language Processing (NeSpNLP 2010)", organized by Roser Morante and Caroline Sporleder. The workshop will be held next July in Uppsala, just before the ACL 2010 events. Workshop website: http://www.clips.ua.ac.be/NeSpNLP2010/index.html.
  • In July-August 2010, SIGNLL will launch a call for shared task proposals to select the task to be run in CoNLL-2011. A designated committee will judge the submitted proposals and declare the selected task no later than October 2010.