Date: 7th July, 2003
Meeting room 101
Sapporo Convention Center (where the ACL conference takes place)
How to get to the Convention Center from the Sapporo Grand Hotel:
>From the Sapporo Grand Hotel to the Convention Center, the whole trip
takes about 20 to 25 minutes.
5 minutes to Odori Subway Station from the hotel,
5 minute-ride on a train (three stations away from Odori Station)
10-minute walk from Higashi Sapporo Subway Station to the venue
Teleparticipants: Sandra Carberry, Diane Litman
8:30 Breakfast
9:15 Welcome -- Mark Johnson
9:20 * Minutes of previous meeting -- Sandra Carberry
9:30 * Treasurer's report -- Kathy McCoy
ACL-02 Final Balance
Chapter shadow accounts developments (if any)
* Investments -- Martha Palmer and Kathy McCoy
9:45 * Secretary's report -- Sandra Carberry
* Electronic elections report -- Sandra Carberry
* Handbooks -- Sandra Carberry
* ACL Website -- Sandra Carberry
10:15 * Nominations -- John Nerbonne
10:30 ACL Office report -- Priscilla Rasmussen
Membership
10:45 Tea / Coffee
Conferences
11:00 NAACL
* Chapter Report -- Diane Litman (Chair)
finances, chapter events, new board members, plans, HLT update
11:15 * ACL-05 Site Selection -- Johana Moore
* ACL-05 HLT involvement -- Diane Litman, Johanna Moore
* Conference organization and difficulty of finding sites
-- Johanna Moore, Diane Litman
11:45 ACL-03, 7th - 12th July, 2003
Sapporo Convention Center, Sapporo, Japan
http://www.ec-inc.co.jp/ACL2003/
* Conference Chair -- Prof. Jun'ichi Tsujii
* Local Arrangements -- Dr. Kenji Araki
* Program Co-Chairs -- Erhard Hinrichs, Dan Roth
student workshop, workshops, tutorials, etc.
12:15 ACL-04 July, 2004
21/07/04 Tutorials and some workshops
22-24/07/04 Main conference
25-26/07/04 workshops
International Conventions Centre of Barcelona (CCIB)
Barcelona, Spain.
* Conference Chair -- Donia Scott
* Local Arrangements -- Toni Badia
* Program Chairs -- Lyn Walker, Walter Daelemans
12:45 * ACL-O6 Call for bids -- Martha Palmer
13:00 Lunch
14:00 Conference organization issues
* ACL anthology support (e.g., $500 charge to conferences)
-- John Nerbonne, Steven Bird
* Conference registration software -- Kathy McCoy
* Electronic submission and reviewing software
-- John Nerbonne (Chair), Johanna Moore, Annie Zaenen
Chapters, etc.
14:45 EACL
* Chapter Report -- Claire Gardent (Chair)
finances, chapter events, new board members, plans
15:00 * Asian Federation -- Hiroshi Nakagawa
15:15 * SIGs -- Johanna Moore
15:30 Coffee/Tea
15:45 Publications
* CL journal -- Robert Dale, Julia Hirschberg
16:00 * Distribution of proceedings / print on demand
-- David Yarowsky (chair), Julia Hirschberg, Mark Johnson
* How soon after a conference should its proceedings appear in the anthology?
16:15 * Executive Committee Tasks -- John Nerbonne, Julia Hirschberg
16:30 * Suggestions for program committee organization -- Mark Johnson
17:00 * Executive Meeting, Winter 2003 -- Johanna Moore
17:05 * Other Business
17:50 Closing
18:00 ACL Welcoming Reception (in Convention Center)
Note: Financial tables are in separate files at the end of the reports.
Mark Johnson
Our Association is forty-one years old this year, which makes it older than many of our members. I'm pleased to report that the Association is still healthy and vigorous, and is well-prepared for the next forty-odd years. We launched no major new initiatives in the past year, but instead used the time to develop and strengthen our existing activities:
The issues facing our organization are not yet major challenges, if we can act on them soon enough:
To summarize, our organization continues to be healthy in terms of membership numbers and its finances, and for this we should be grateful to the Office Manager Priscilla Rasmussen, Treasurer Kathy McCoy, and the Association's Secretary Sandra Carberry. I'd like to thank Sandra in particular for many long hours she has devoted to ACL on issues both large and small. We are extremely lucky to have someone as devoted as she is, and as anyone that has been involved in the day to day running of the ACL knows, we are a better organization thanks to her efforts.
Mark Johnson
2003 ACL President
June, 2003
Sandra Carberry
The ACL election web page is being set up. Greg Silber is now handling the election software, and the election will be overseen by Greg, Dmitriy Genzel, and me.
Drago Radev is continuing to serve as "technical webmaster" with responsibility for a variety of items, including the ACL Universe and maintaining the email aliases. Updates have been made to the ACL handbook, and the conference chairs have been asked to provide information that will be added to their portion of the handbook.
The ACL officers and Executive Board members have been given a template for recording their duties along with a timetable and suggestions for the person following them in office. These will be stored on the ACL web site for access by future officers.
Dan and Erhard used the START system. http://www.softconf.com/START/ for submission and reviewing this year. We have not made a lot of progress on identifying reviewing software for ACL, so I thought it might be useful to build on Dan and Erhard's experience and try the START system for ACL-04 and perhaps NAACL-04. Dmitry Genzel, a student of Charniak's at Brown, has agreed to provide technical support; I envision this as Dmitry helping set up the system at the new site and work on fixes or minor revisions, but not major revisions like Dan did (unless of course Dmitry wants to put more effort into it, but that is not the expectation). I have not worked with Dmitry before so I contacted Eugene who said "as far as I can remember, if Dmitry says he will do something, then he will do it. If it were me I would take him up on his offer."
Below I've summarized further information, including answers to questions that were raised by the Exec and answers to other questions that I posed to the owner of the START system. I am not pushing for the use of this system. And I cannot predict how it will work out, except that Dan and Erhard were happy with it. If the Exec and ACL-04 conference and PC chairs want to go this direction, then we needto identify the host site and then contact the START owner toget the license. If Dmitry Genzel is to help with tech support, then he will need an account at the host site so that he canaccess the system. (Dan also said that his student was now an expert at the software, so perhaps Dan's student and Dmitry might work together on this?)
DAN'S OPINION
-------------
Dan's comment about the START system is:"Yes, I think it's pretty good, and I believe Erhard likes it too. All the area chairs that came to the committee meeting were pretty happy with it."
ANSWERS TO EXEC QUESTIONS
-------------------------
The following are Dan's answers to questions posed by the Exec:
1. How much customization did you have to do? How many person hours would you estimate were involved.
Some of the customization is necessary and is actually part of
setting up the conference. This includes decisions such as:
- having two level of reviewers (area chairs and committee members)
- setting up blind reviews (and deciding who can see what)
- setting up different permissions such as - can reviewers see other reviews, for which papers, and when.
- setting up review forms
- a few more.
In addition, we made the pages looks a bit different than the original given to us, solved a few bugs the system had, etc. We also developed some tools to collect statistics on the review process and results. Between me and my assistant we easily spent 50 hours on it. (This does not include "using" the system to do the program chair work.)
2. How much effort/hours would you estimate would be involved in setting it up and making needed changes for subsequent conferences?
A new program chair can probably set it up and get acquainted with the system within a day or two.
3. Was the customer support good?
We did not use it. Once we installed the system we did all on our own.
4. Did you download it and host it at your own site, or did you host it at softconf.com (ie., the vendor). I'm assuming that you hosted it at your institution, since you customized it.
Yes, we hosted it. This, in fact, was one of the criteria I used to choose the system. I did not not a system that is hosted somewhere else. First, there are issues with having the database somewhere else. And, I preferred to be responsible on fixing problems than counting on someone doing it remotely. (The system is simply a collection of perl scripts) This turned out to be important - we found a permission related bug in the system that was discovered only after the area chairs received their assignments, and we had to solve it that night...
5. Do you recommend using it? What are the positives and negatives?
Yes, I think it's a pretty good package. We discussed it briefly in the committee meeting and all the area chairs were happy with it. That's the case also with all the reviewers I talked to.
It's going to take more time to list all arguments but briefly:
- I think the systems simplifies the work for the program chairs in terms of submission distribution of papers, simplify assignments and allows easy access to all the information.
- Likewise - it simplifies the work of area chairs - both in assignment and in following up on the status of the reviewing. [Which, along with some nagging, contributed to the fact that we got 1050/1080 of the reviews on time!]
- It allows for a better review process in that it can be set up to support reviewers seeing other reviews on the same paper, once they submitted theirs.
PRICING
-------
I contacted the owner of the START system to get pricing information. If we host it ourselves, meaning at the institution of one ofthe PC chairs for the particular conference, then the cost is $500 for the first event, $400 for the second event, $300 for the third event, etc. until one reaches the minimum per event charge of $100. If the system is hosted at softconf.com, then the starting price for the first event is $600 instead of $500 and it would take 6 events to get to the minimum per-event fee of $100. We can try START for one conference, and later decide to extend our license to cover multiple events. In such a case, he'll apply the discounted fees retroactively.
USING DAN'S MODIFIED SYSTEM
----------------------------
In discussing our needs with the owner of the START system, I said that we would probably want to license the START system as modified by Dan Roth. He did not say that we could not do that, but his argument against it follows (in addition to the fact that it could not be hosted at his site):
"Please note: Since ACL-03 used START V2, the system has been improved quite a bit. There are a lot of new features, which members of the ACL-03 committee did not have the chance to use. Also, a few bugs were found in previous releases, all of which have been fixed. There are no outstanding bugs reported. Hence, I do not think it would be a wise move to use a previous version of START V2 for an upcoming conference. You should use the latest version (the present one is V2.45.2 - about 13 revisions later than that used for ACL-03). If there is some particular feature you need in the START V2 for your confererence(s) - and it is not currently in the product - let me know. I'll probably be able to incorporate this feature myself, as an option, and make it available to all users. Anyway, I think you'll see that the benefits of using the latest version outweigh the liabilities (if there are any)."
"Perhaps you should check out a demo at softconf.com, compare it to Dan's install, and tell me what you believe is lacking in the core functionality. Then I can probably give you "Dan's Modification" as an option; meanwhile you'd be able to use the most up-to-date version of START V2. For what it's worth, I doubt that the modification was very major. It's not very easy to change any of START's core functionality, unless one REALLY knows what one is doing. Even for an expert web programmer, changing the code inside of START would be very hard to do."
TECHNICAL SUPPORT
-----------------
"All START licenses include service, wherever the program is hosted. You can count on my help while you are using the software: to resolve technical problems, or for advice on setting up the program to best fit your needs, or for any issue concerning the software (or using it)."
Kathy McCoy
FINANCES:
This report refers to 3 "charts" on the web for the financial report. See also annexes to the reports.
The charts:
net-worth-12-31-02.pdf
net-worth-12-31-02.ps
This contains a line graph of ACL's networth over time (from about 5 years ago). You can see we hit a low point in the first quarter of 1999 of $338,255.52 from which we had a jagged by fairly steady rise through the last quarter of 2001 reaching a high of $730,787. Over the last year we have had a fall back down to about $525,231.
What I cannot explain is how we possibly reached our high point, given that I think the problem was really our conferences in 1999 (which basically broke even) and in 2000 (which had a loss of about $30,000). I can only speculate that we paid most of the monies for both of these conferences IN ADVANCE of the conference. This was the opposite of what we did for NAACL2001 and ACL2002. The other thing that may be in effect is the 5-year memberships. That likely infused money into the accounts and may have given a false impression of our income from memberships at that time.
NOTE: I did not do an update on the European accounts. I will have that for the next meeting. But, these should have gone higher over the past year.
Also, I have not included NAACL's account here -- it contains about $10,000 at this point.
To see where our money is coming and going see:
income-vs-expense-by-cat.pdf
income-vs-expense-by-cat.ps
This contains our income versus expenses by category over the past 2 years. You can see that our biggest income and biggest expense remains our conferences. Membership is our next income producer, with the journal being the next largest expense. We also have a lot of office expenses (which don't have any good corresponding income categories).
As I have mentioned in the past, I have tried to categorize salaries so that they are not all included in ACL Office (e.g., it is important the Priscilla's time gets categorized to the individual conferences especially in light of the shadow accounts).
We have started to implement the shadow accounts for the chapters and for the sigs. Unfortunately, the accounting from ACL2002 is still being completed, and that affects the shadow accounting for NAACL. The European shadow will be reported on by Mike Rosner.
ON-LINE REGISTRATION
I hired a company named AgoraNet and worked with them to design and implement on-line registration software. We have used it for two conferences and it seems to be working beautifully. Currently the company has updated the pages for new conferences, and it writes all registration information to a database at Delaware. The interface seems good, and the reports that it produces for the office are great. I will also have them do the registration for ACL-04 because it will be the first time that we accept Euro credit cards (something new for the interface to handle). Once the design settles, I hope to hire a student to do the updating for new conferences.
ACL BOOKKEEPER
We hired a bookkeeper in March. Her first task was to go through the accounting for ACL-2002, and she is working out very well. She is very thorough and has uncovered some questions that we need to get answers to before we can report on the finances for that conference!
The 2003 Nominating Committee of the Association for Computational
Linguistics, consisting of Wolfgang Wahlster
(Chair, 2001-03), Ed Hovy (2002-04), and John
Nerbonne (2003-05) is delighted to present the following slate of candidates
for consideration by the membership of ACL:
VP-elect
:
Key-Sun Choi (KAIST, Korea)
-- kschoi@cs.kaist.ac.kr
Jun-Ichi Tsujii (Japan) -- tsujii@is.s.u-tokyo.ac.jp
Member at
large of the ACL Executive Committee:
Tilman Becker (Germany)
-- Tilman.Becker@dfki.de
Walter Daelemans (Belgium) -- walter.daelemans@uia.ua.ac.
Remarks
from the Nominating Committee on this Slate of Candidates:
In light of ACL's continuing
outreach to Asia, the nominating committee has nominated two outstanding
Asian researchers for the position of Vice President Elect.
Two Europeans have been nominated for the open slot on the committee to balance representation from the various continents.
All of the nominees have a long history of distinguished technical
contribution and reliable professional service to ACL.
Remarks
on the Search Process:
The Nominating Committee started its search on 8 February 2003
and finished its work on 19 February by forwarding the above slate of candidates,
who have all agreed to run, to Sandra Carberry.
===============================================================
Claire Gardent, Chair
John Carroll, Secretary
1. EACL'03 Conference (Budapest, 12-17 April 2003)
EACL'O3 gathered 350 participants thus being the biggest EACL since the inception of the conference. The conference had a strong European flavour with 78% participants from Europe, 12% from America, 7% from Australasia and 3% from Africa; and a truly international dimension with 40 countries being represented (29 being European).
As is now usual in ACL related conferences, the event consisted of a main conference, a student research workshop, demos, (12) workshops and (4) tutorials. For the first time however, the EACL conference included a research notes session with the aim of opening up the conference to a wider audience. Thus whilst for the main conference, the acceptance rate was 26.5%, research notes had an acceptance rate of 60%. Demos were accepted to the rate of 80% whilst the student research workshop adopted a selective 33% acceptance rate.
2. New EACL Student Board
The current EACL Student Board -- Malte Gabsdil, Jon Herring, and Julia Hockenmaier -- have done a great job setting up and managing the EACL web site, and organising the EACL'03 student workshop. Their term of office finishes in July.
The new Student Board that will serve from July, for 2 years, is:
Leonoor van der Beek, University of Groningen
Irina Chugur, UNED, Madrid
Sebastian Pado, University of Saarbruecken
The old board is in the process of handing over to the new one.
3. Newsletter
Gertjan van Noord (EACL Chair-elect) has taken over from Claire Gardent as editor of the EACL newsletter. The fourth issue will be sent out to members later this year.
4. EACL Foundation
John Nerbonne is pushing forward with plans for a foundation associated with the ACL European Chapter. The motivation is to create a "European legal person" which would allow EACL to apply for EU funding. The foundation will be registered at the Chamber of Commerce in Utrecht (The Netherlands), and have an official postal address at the University of Groningen. The board of the foundation will consist of the three primary officers of the EACL, ex officio, the chair, the secretary and the treasurer.
The ACL executive committee has approved this, on condition that the foundation submit an annual report (which should simply become part of the EACL chair's annual report), and a clean financial relationship Ð a separate bank account.
John Nerbonne has an estimate from Lubbers and Dijk, of 800 euros for their work and the cost of registering the foundation. There is an annual registration fee of around 30 euros.
5. Sponsorships
As usual, EACL will sponsor an ESSLLI'03 foundational course, with a contribution of 750 euro towards travel costs for the presenter. In return for sponsorship, members of EACL will be eligible for a reduced registration fee.
EACL will sponsor the computational lexicography training workshop LEXICOM, to be held at the University of Brighton, 13-19 July 2003. The contribution will be 750 euro, to be administered by LEXICOM to support the participation of a student from Eastern Europe.
EACL will also provide 750 euro sponsorship to the EUROLAN 2003 Summer School 'The Semantic Web and Language Technology: its Potential and Practicalities' in Bucharest, 28 July - 8 August 2003. The support is intended for student grants for disadvantaged participants from Eastern Europe.
June 2003
Diane Litman, Chair
http://www.naacl.org
1. Elections
The NAACL election in the fall of 2002 (held electronically) had 4 candidates for 2 open Chapter Board positions (2 year terms). Graeme Hirst was re-elected, and Dekang Lin replaced Dan Jurafsky (who did not run for re-election).
The 2003 election (for Chair, Secretary, Treasurer, and 2 Board positions) will be happening this fall; the nominating committee is already setting the process in motion. Both the Chair and Secretary are ineligible for re-election to those positions due to term
limits. The Chair will remain on the board for 2 more years as Past Chair.
2. Shadow Account Status
2001 balance 33,104.27
2002 JHU workshop - 9,519.96
NAACL Board meetings, Web fees, etc. - 1,723.05
----------
BALANCE 21,861.26
TO BE RECONCILED:
+ 2002 ACL Conference (surplus expected, but apparently much smaller than ACL initially estimated)
- 2003 JHU workshop (approximately 10,000)
+/- 2003 HLT-NAACL Conference (+ expected)
3. HLT-NAACL 2003
Recall that for 2003, the NAACL Chapter Board decided to merge the NAACL conference series with the HLT (Human Language Technology) conference series. The conference was a fusion of the (NA)ACL and HLT styles (e.g., plenary demos, highly referred long papers, late-breaking short papers/posters, student workshop and tutorials, etc.), and also spanned several previously relatively independent subareas of human language technology. As a result, the conference was designed to especially encourage reports of work on synergistic combinations of language technologies, as was the selection of conference organizers:
General chair:
Eduard Hovy, USC Information Sciences Institute
Program co-chairs:
Marti Hearst, U of California Berkeley
Mari Ostendorf, U of Washington, Seattle
Local arrangements chair:
Dekang Lin, U of Alberta
Workshop and tutorial co-chairs:
IR: James Allan, U of Massachusetts at Amherst
NLP: Jason Eisner, Johns Hopkins University
Speech: Wayne Ward, U of Colorado
Demo co-chairs:
Bob Frederking, CMU
Bob Younger, SPAWAR
Publications co-chairs:
Dragomir Radev, U of Michigan
Steven Abney, U of Michigan
Student Workshop co-chairs:
Faculty Advisor: Marc Light, U of Iowa
IR: Victor Lavrenko, UMass Amherst
Speech: Costas Boulis, U of Washington
NLP: Eric Breck, Cornell U
In addition, the conference gave special consideration to papers that addressed the special topic of new approaches to and uses of unsupervised and lightly supervised learning techniques.
The conference was just held, and by all accounts was very successful. Over 400 people attended the meeting (despite the outbreak of SARS, and both EACL and ACL within a relatively short time span). We were happy with the mix of workshops and tutorials, and moderately happy with the mix of long paper area coverage. 162 full and 80 short papers were submitted, of which 37 and 41 were accepted. There were 9 workshops, 9 tutorials (2 of which unfortunately had to be canceled due to lack of registration), 17 demos (3 of which were presented in a plenary session), and a student workshop.
The conference website
contains full details regarding the conference.
4. HLT-NAACL 2004
Based on the success of this year's HLT-NAACL meeting, we are continuing the merged meeting in 2004. The meeting will be held in Boston (probably in early May). We extremely fortunate to have assembled the following conference team:
General chair:
Julia Hirschberg, Columbia University
Program co-chairs:
Susan Dumais, Microsoft
Daniel Marcu, USC Information Sciences Institute
Salim Roukos, IBM
Local arrangements chair:
Christy Doran, Mitre
5. ACL 2005
Looking ahead to 2005, ACL will again be in North America. Since NAACL is a co-host with ACL, we cannot have a separate merged meeting with HLT that year. In order to keep the momentum of the merged conference series going, both NAACL and HLT would be very interested in having a co-located HLT event during ACL.
6. Support for Summer Schools
We are in the second year of a three year cooperation with Johns Hopkins's CSLP Summer School on Human Language Technology. This year we have funded 10 undergraduate and graduate students to attend the 10 day course that takes place before the workshop.
Since ACL asked us to take responsibility for any future arrangements with the LSA Linguistics Institute, we also handled the arrangements for the 2003 Institute (although ACL still paid the bill this one last time). A member of NAACL and a member of the LSA worked together on the courses and instructors to be funded. Money was also set aside to provide financial assistance to worthy NAACL student members, but apparently none applied. LSA has promised to be more diligent about promoting the Institute to our membership in the future.
After ACL referred NASSLLI to NAACL, NASSLLI has as yet not pursued our offers of cooperation for their 2003 (or future) summer institutes.
7. Outreach to Latin America
Previously, the NAACL executive committee asked Ted Pedersen to act as a liaison to members of the CL/NLP research community in Latin America. Ted has set up a Registry for Latin American Researchers website: http://www.d.umn.edu/~rass0028/classtest/registry.cgi
8. Executive Committee Meetings
The NAACL Board met for its first ever winter meeting on February 14, via telephone (following the ACL model). The NAACL Board met again in person during our 2003 Conference. The minutes of both meetings will be made available on the web.
Junichi Tsujii
(1) AFNLP planned to hold its first conference at the Hainan Island, China in December, 2003. However, it has taken more time than we expected for Chinese colleagues to get approval from their authority of joining AFNLP. There has also been uncertainty due to SARS outbreaks in Asia, and quite a few international conferences initially planned in China have been postponed or cancelled. Due to these two factors, preparation of the first conference has not made much progress, and we now think that we have to either cancel or postpone it.
(2) However, Prof.B.Tsuo in Hong Kong and colleagues in Taiwan have made significant progress in resolving the difficulty that prevented our Chinese colleagues from joining AFNLP, and we will have a meeting during ACL03 at Sapporo to approve the revised constitution of AFNLP. Chinese colleagues tell us that the constitution we porpose is fine with them, and with approval by them, I hope we will be able to announce their membership at the time of ACL.
(3) We will have a meeting of conference coordination committee (CCC) of AFNLP during ACL03 to decide the venue and the date of our first conference. Currently, two options are being discussed. One option is to postpone the first conference at the Hainan Island to late February or March, 2004. Another option is to cancel it and re-arrange the conference in November or December, 2004. In the second option, we have to choose the venue.
(4) In either option, we would like to have close coordination with other ACL-related conferences and we appreciate all supports we will have from ACL.
===============================================================
Junichi Tsujii (Department of Computer Science, University of Tokyo, Japan)
I have constantly received help and advice from the ACL Exec, without which I could not have organized ACL03. In particular, John Nerbonne has given advice from the very beginning, i.e. preparation of the conference proposal, and Mark Johnson has given us great help such as preparing the policy for SARS. The SARS policy provided by the Exec was very effective for placating the worries of potential participants and the local organizers.
Kathleen McCoy and Sandra Carberry have helped us and given us very useful advice on every aspect of budget preparation and conference preparation. Priscilla Rammusen has given us her insights on every aspect of conference preparation such as hotel bookings, banquet, etc. The local organizing committee chaired by Kenji Araki has been energetic throughout difficult times caused by the uncertainty of SARS, while the financial committee chaired by Hiromu Hayashi was very successful in attracting sponsorships in Japan.
I am very grateful to all the people involved in preparation of ACL03. I believe that conference preparation till now has been extremely well organized by people with enthusiasm and professional skills.
The conference seems to have attracted more attendees than we predicted (522 registrants when the early regitration was closed). It has also attracted much more sponsorship than initially planned (the precise figure will be given by the local organizer), while the main conference attracted many more paper submissions that expected (360 papers). 12 WSs including two associated WSs (IRAL and EMNLP) were planned, and while one WS was cancelled due to SARS, they also attracted many paper submissions and accepted about 140 papers. Together with the main conference (71 papers), student workshops (12 papers) and demo/poster sessions (27 papers), we will have around 250 papers.
While we still have uncertainty due to SARS (SIG-HAN, one of the largest WSs, may have far less participants than they initially predicted), ACL03 will be the largest conference of our field so far held in Asia.
Since I wanted to make ACL03 as much international as possible, I tried to have balanced organization of committees, balanced in terms of geographical regions, genders and generations. While it is impossible to build a team of perfect balance, I believe that the team for ACL03 was well-balanced. One co-chair was chosen from each of the three geographical regions (Asia, Europe and North America) for the committee of workshops and that of tutorials, while Dan Roth from North America and Erhard Hinrichs from Europe agreed to be program co-chairs with Yuii Matsumoto from Japan as the chair of Poster/Demo sessions.
This policy was successful and contributed to raising awareness of ACL03 in the three regions. I have not applied the same policy to the committees of Publicity, Industrial Demos and Finance. Actually, these three committees were not only chaired by Japanese but also consist of only Japanese members (one exception is one of the co-chairs of Industrial Demos, Key-Sun Choi, from Korea). This all-Japan organization of committees may be one of the reasons why we were not successful in attracting sponsorships and industrial demos from other countries.
One of my concerns was whether ACL03 could attract reasonable numbers of paper submissions and participants, since Japan is far away from the US and Europe and since ACL has not yet been well rooted in Asia and in Japan. Fortunately, ACL03 took place at a year when Coling and LREC do not take place. The dates of the three conferences (NAACL, E-ACL and ACL03) were coordinated from the very beginning. However, the conflict with NAACL and E-ACL remained one of my major concerns.
Ideally, the conferences organized by ACL should be complimentary, but not competitive, with each other. They should provide members of ACL with as many opportunities as possible for them to present their work, without compromising the quality of conferences. In particular, one whose paper is rejected by a preceding conference should be able to resubmit a paper revised by considering reviewers' comments and/or by adding new research results.
The submission deadlines of NAACL and our conference were two months apart, which was long enough for people to submit papers with different content. The notification date of NAACL, 24/Feb, was too close to the submission deadline of ACL03, 26/Feb for people to revise their papers by considering reviewers' comments. The PC of NAACL delivered their notification earlier than their official notification date, which we appreciated very much. However, three weeks gap between the notification date of a preceding conference and the submission deadline of a following one would be the minimum.
In future, the four conferences (NAACL, E-ACL, Conference organized by AFNLP, and International ACL) should discuss not only their dates of conferences, but also the submission deadlines, notification dates, etc. in order to maximize the opportunities that these conferences provide ACL members with.
I realized that policies on several issues had not been explicitly stated before. Those are:
[A] Confidentiality
It is a common assumption that PC members and reviewers would never divulge the content of papers that they read in reviewing process. However, people have become more aware of and concerned with confidentiality of their research. We should have a written policy of confidentiality and make them public in CFPs or conference Web site in future conferences.
[B] Cancellation and Liability
Policies for cancellation and liability were agreed upon and made public at the late stage of conference preparation, due to the uncertainty caused by SARS. These policies have to be made explicit from the beginning in future.
We introduced the fee for WS-only attendees this year. I think the level of the fee was appropriate and this arrangement generally worked well. However, it should have been made clear from the beginning that there are two different arrangements for WSs, that is, an official WS/Associated conference like EMNLP and a WS collocated by ACL03, like the one organized by SIG-Dial. It would be useful for WS organizers if we could give a guideline on these two arrangements from the beginning.
While I have not heard any complaints, some of the attendees of IRAL who attend only IRAL may not be so happy since they have to pay the WS-Only registration fee, which is not so cheap. Quite a few attendees of the past IRALs are IR specialists and are not so much interested in NLP or CL per se.
Since our fields are rapidly expanding to interdisciplinary fields such as those with information retrieval, multi-media communication, speech processing, Bio-Medicine, etc., we will have more cases that are similar to IRAL in future.
It is not clear, either, what WS-only attendees will get, i.e. whether they get the CD-Rom and conference bag, whether they can attend banquet, tutorial and reception if they like. There should be an agreed policy. Since some students want to attend tutorials without attending the conference part of ACL03, it would be nice if we could have special fees for Tutorial-only attendees.
The early registration deadline of ACL03, 14/June, was too close to the conference (7/July) this year. Three weeks were not long enough for the local organizer to decide the number of conference kits to be ordered. It was particularly problematic this year due to the volatility of the situation due to SARS.
To estimate the number of participants is the major aim of having early registration. However, the local organizer had to order conference kits before the early registration was closed. We should have closed the early registration at least one month before the conference.
The publicity committee issued Newsletters every month from November. They issued 5 newsletters plus two CFPs. They all contained useful, timely pieces of information. These newsletters contributed to general awareness to ACL03.
The publicity committee has carried out their work effectively. However, if we had co-chairs or members of the committee from Europe and the US, we would have had more effective publicity in these two regions.
We have not had effective publicity of these two funding. Announcements of student support only appeared in the website of student research workshop, and some eligible students may have missed the announcement.
The two co-chairs have done extremely good jobs. The effectiveness and efficiency of the publications this year were mainly due to their personal devotion and hard work. I hope their personal experience and enthusiasm will be transferred to future conferences.
The division of labors this year between the publications co-chairs and the WS organizers seems the right one, which should be adopted in future conferences
Advice given by the Exec was very useful for selecting chairs of committees. Since sometimes it was not clear who decided what, it would be desirable to have agreed protocols in the ACL manual how to decide chairs of committees.
Once such protocol is established, the general chair can go on his/her own with appointment of chairs with less communication with the Exec, while s/he should be given advice in the form of a list of possible candidates and general principles. The same should be the case for the PC co-chairs. A protocol of communication among the PC co-chairs, the general chair and the Exec should be established for appointment of area chairs of the PC committee.
The committee of sponsorships and industrial exhibitions of ACL03 successfully attracted sponsors and exhibitions. They exceeded the target we initially set. However, the sponsors are exclusively from Japan except for Microsoft. Even contact with Microsoft was made in Japan. While we planned to attract sponsors from other Asian countries like Korea, Taiwan, etc., this did not materialize.
We should haveincluded members from countries other than the host country in the financial committee or the committee industrial exhibition to broaden the support base.
The PC committee recruited its members from the three regions, which contributed to balanced distribution of paper submissions and accepted papers. However, such geographical distribution of area chairs increased the travel expenses for their meeting. We may have to think different arrangement of PC meetings in future.
It is also useful to have standard formats of e-mails of asking people to serve as chairs and members of various committees which state duties, obligations and standard arrangements for expenses incurred.
The result of early registration shows that we attracted more attendees than we predicted. This is an accomplishment considering the uncertainty caused by SARS. While the early registration figures show that we will have many more participants from Japan than our prediction, we attracted far less participants from China, Taiwan and Europe than we predicted. This may be partly due to SARS in China and Taiwan, but we need careful analysis for future strategy for conferences in Asia.
Kenji Araki
See separate Excel document for budget.
Registration figures as of June 20, 2003.
MAIN CONFERENCE:
485
WORKSHOPS:
EMNLP:
131
IRAL:
48
Multilingual Summarization
and Question Answering
53
Natural Language Processing
in Biomedicine:
48
The Lexicon and Figurative
Language:
22
Multilingual and Mixed-language
Named Entity Recognition
Combining Statistical and
Symbolic Models:
34
The Second International
Workshop on Paraphrasing, Paraphrase Acquisition and Applications (IWP2003)
56
Second SIGHAN Workshop on
Chinese Language Processing:
41
Multiword Expressions: Analysis,
Acquisition and Treatment
40
Linguistics Annotation:
Getting the Model Right:
19
Workshop on Patent Corpus
Processing:
18
TUTORIALS:
Finite State Language Processing: 43
Maximum Entropy Models, Conditional Estimation,
and Optimization without the
Magic:
95
Knowledge Discovery from Text: 83
Spoken Language Processing: Separating Science
Fact from Science Fiction: 43
BANQUET:
251
STUDENT LUNCH:
88
Student Volunteers:
6
(We could not ask
many student volunteers since ACL, GC and LOC can not have responsibility
of SARS for the student volunteers. We employ part-time job students and
asked professor volunteers and Sapporo citizen volunteers.)
Executive Board
Meeting, Sapporo Covention Center, 8:30 - 6:00 15/20
Contact: Sandee
Carberry, ( Hiroshi Nakagawa : local)
Welcome party,
Sapporo Convention Center, 18:00-20:00
Contact: Kenji Araki
EMNLP Reception,
:
Contact: Michael Collins : mcollins@ai.mit.edu
EXEC Dinner, KANI-YA,
19:00-22:30, 40
Contact: Kenji
Araki and Hiroshi Nakagawa
Banquet, Sapporo
Grand Hotel, 19:00-22:00, 251 guests
Contact: Kenji
Araki
(Generally the
time for a party is 2 hours at a hotel in Japan. I negotiated the extension
of the party time. The hotel accepted my request. But the banquet has to
finish at 22:00 completely.)
ACL Conf. Comm.Mtg.
Breakfast, Sapporo Grand Hotel, 8:30-12:00
Contact: Kenji
Araki
Student Luncheon,
Sapporo Convention Center, 12:15-14:00
Contact: Kenji
Araki
Budget to date (June 28)
See attached Excel sheets
Prepared by Erhard
Hinrichs (University of TŸbingen, Dan Roth (University of Illinois)
June 20, 2003
The first action
item for the program co-chairs was the selection of program areas and program
area chairs. The following areas and area chairs
were chosen:
Word Segmentation,
Shallow Parsing, Tagging and Chunking
Thorsten Brants,
Google Inc., U.S.A.
Syntax, Semantics,
Grammars, Morphology and Phonology
Anette Frank, DFKI, Germany
NL Applications
Koiti Hasida, Cyber
Assist Research Center, Japan
Ming Zhou, Microsoft
Beijing, China
Machine Translation
and Multilinguality
Kevin Knight, Information
Sciences Institute at USC, USA
Speech, Language
Modelling, Statistical Methods
Chris Manning,
Stanford University, USA
Computational
Psycholinguistics; Lexical Semantics and Ontologies for NLP
Paola Merlo, University
of Geneva, Switzerland
Machine Learning
for Natural Language
Raymond J. Mooney,
University of Texas at Austin, USA
Information Extraction/Information
Retrieval
Hwee Tou Ng, National
University of Singapore, Singapore
Generation and
Parsing
Gertjan van Noord,
University of Groningen, the Netherlands
Language Resources
Hae-Chang Rim,
Korea University, Korea
Discourse and
Dialogue
Bonnie Webber,
University of Edinburgh, UK
Due to the high
number of expected submissions in the program area of Natural Language Applications,
two program chairs were chosen for this area. Each
area chair recruited the help of 20-30 associated reviewers. Overall there were 237 reviewers, in addition to
the area chairs.
The timetable allowed
for reviewing, notification and camera ready papers -- determined in consultation
with the general conference chair, the area chairs, and the publication co-chairs
-- was a bit tighter than usual; dates were set to February 26, April 26
and May 20th, respectively. Our schedule also allowed for about a week of
discussions between reviewers and area chairs, before the committee meeting
took place. Over all, with significant effort of the area chairs, over 97\%
of the reviews (1050/1080) were completed on time. All notifications to authors
went out on time.
Electronic handling
of papers by the START ConferenceManager Tool}
Following last
year's practise, we allowed only electronic submission.
This year we used the START Conference Manager package to help with
the submission, distribution of papers to area chairs and reviewers, and
the review process itself.
Details on the
package's capabilities and issues involved in using it are discussed in another
document, so here we detail only decisions made when setting up the software,
which are relevant from the reviewing and decision making perspective.
o
The review process
was double blind. Reviewers did not know the identity of the authors. Area
chairs knew the authors of papers in their area.
o
Authors declared
an area for their paper (along with key words). With the aid of the software
package, these were used by the program chairs as guidelines when splitting
papers to area chairs. Deviation from authors' declarations was decided due
to mistakes, conflicts of interests and in an attempt to divide the work
among area chairs more uniformly. This was done manually by the program co-chairs.
o
Once papers were
assigned to areas, reviewers for each paper were assigned manually by the
area chairs, with the aid of a convenient user interface in START. The software
allows also automatic assignment and a bidding process but we chose to do
it manually.
o
A revised review
form was developed, in consultation with area chairs and approval of the
exec committee.
o
A reviewer for
a given paper could read other reviews written for this paper, and know the
names of the other reviewers for the paper, once his/her review was complete.
We felt that this contributed both to a high level of reviews and helped
facilitating discussions.
Out of 376 papers
that were originally submitted, 360 papers were distributed to area chairs
for review. All papers were pre-screened by the
program co-chairs to ensure that they conformed to the submission guidelines.
16 papers were
either withdrawn by the authors after the submission deadline or had to be
rejected because they did not conform to the guidelines for submission..
The remaining 360
papers were assigned to the program areas by the program co-chairs. Table 1 shows the distribution by program area. The columns Assigned and Declared reflect the manual
assignment done by the program chairs vs. the declared areas. Acceptance rate is computed relative to the assigned
papers. Tables 2 and 3 show the regional distribution
of reviewed papers (based on first author).
The PC meeting
took place on April 20-21 at the Centrum fŸr Informations- und Sprachverarbeitung
(CIS) of the Ludwig-Maximilians-Universit\"at Munich, Germany. Our special
thanks go to Franz Guenthner and Klaus Schulz for their local co-organization.
The committee meeting
was attended by the two program co-chairs and by eight area chairs. Chris Manningg, Raymond J. Mooney, Hwee To Ng and
Bonnie Webber were not able to attend the meeting. Michael
Strube (EML Heidelberg) kindly agreed to represent the program area of Discourse
and Dialogue at the meeting.
The committee unanimously
agreed on the proposed agenda and decision procedure.
It was also decided to delay discussion of any submitted papers that
was authored or co-authored by a program area chair until the end of the
meeting.
The meeting started
with a report by the area chairs on the outcome of the reviewing. Each area chair was asked to nominate definite and
possible accepts on the basis of the external reviews.
At this stage, the numerical scores of the program were not used any
more and each paper was called individually in order to allow for comments,
and to achieve uniformity across areas. The first
pass of accepted papers was followed by a detailed discussion of borderline
cases and cases of conflicting opinions. A few
papers were read by additional reviewers.
Out of 360 reviewed
papers, 71 papers were accepted. This amounts
to an acceptance rate of 19.72\%. Table 1 shows
the acceptance figures per program area. Table
2 shows the regional distribution of accepted papers.
Dual Submissions
There were quite
a few cases of dual submissions. In all cases, the authors of accepted papers
chose to present at ACL-2003.
In several cases
the program chairs were not made aware of dual submission by the authors.
A similar situation happened in EMNLP. In addition
to a waste of resources by our community, this also results sometimes in
similar papers accepted to multiple conferences. We
believe that there is a need to (1) come up with a policy that will discourage
people from sending almost identical papers to several conferences, and (2)
coordinate the reviews to save reviewers' time.
This year's conference
also continues the tradition of recognizing one of the submitted papers with
the award of Best Paper of ACL-2003. A subcommittee
of program area chairs and the two program co-chairs was formed to select
the best paper.
Table 1 Paper
submissions by program area- assigned, declared and acceptance rate
|
Area |
As |
De |
Ac |
Rate |
|
Machine Translation
and Multilinguality |
29 |
30 |
7 |
24.14 |
|
Syntax, Semantics,
Grammars, Morphology, and Phonology |
26 |
34 |
6 |
23.08 |
|
Comp. Psycholinguistics,
Lex. Semantics and Ontology |
37 |
34 |
7 |
18.92 |
|
Natural Language
Applications |
41 |
33 |
4 |
9.76 |
|
Word Segmentation,
Shallow Parsing, Tagging and Chunking |
33 |
27 |
10 |
30.30 |
|
Speech, Language
Modeling, Statistical Models |
36 |
25 |
10 |
27.78 |
|
Generation and
Parsing |
33 |
38 |
4 |
12.12 |
|
Language Resources |
22 |
13 |
5 |
22.73 |
|
Information Extraction,
Information Retrieval |
36 |
63 |
5 |
13.89 |
|
Machine Learning
for Natural Language |
34 |
40 |
7 |
20.59 |
|
Discourse and
Dialogue |
32 |
23 |
6 |
18.75 |
Table 2 Paper
submissions and acceptance rate per region
|
Asia and Oceania |
144 |
22 |
14.01% |
|
America |
109 |
32 |
29.36% |
|
Europe |
93 |
17 |
18.28% |
|
Africa |
É1 |
0 |
0% |
Table 3 Papers
submissions per country
|
Asia |
|
|
|
|
|
|
Japan |
71 |
India |
5 |
|
|
China |
26 |
Thailand |
4 |
|
|
Korea |
17 |
Hong Kong |
3 |
|
|
Singapore |
9 |
Philippines |
1 |
|
|
Taiwan |
7 |
Indonesia |
1 |
|
America |
|
|
|
|
|
|
US |
103 |
Canada |
6 |
|
Europe |
|
|
|
|
|
|
Germany |
28 |
Bulgaria |
2 |
|
|
UK |
20 |
Greece |
2 |
|
|
France |
14 |
Spain |
2 |
|
|
Netherlands |
5 |
Belgium |
1 |
|
|
Italy |
4 |
Estonia |
1 |
|
|
Finland |
3 |
Norway |
1 |
|
|
Ireland |
3 |
Switzerland |
1 |
|
|
Sweden |
3 |
Czech Republic |
1 |
|
|
Denmark |
2 |
|
|
|
Oceania |
|
|
|
|
|
|
Australia |
12 |
New Zealand |
1 |
|
Africa |
|
|
|
|
|
|
Tunisia |
1 |
|
|
We would like to
thank everyone who helped us with this year's main program. We greatly appreciate the advice and support by the
General Conference Chair, Jun-ichi Tsujii (The University of Tokyo), and
the Local Organizing Chair, Kenji Araki (Hokkaido University). We would also like to express our gratitude to the
For the preparation
of this volume and for the production of the CD-ROM version of the proceedings
we would like to thank the Publications Co-Chairs, David Yarowsky (Johns
Hopkins University) and Sadao Kurohashi (The University of Tokyo).
We gratefully acknowledge
the advice of the ACL Executive Committee, the experience and guidance of
Kathleen McCoy, ACL Treasurer, of Sandra Carberry, ACL Secretary, and the
help from both the past and present presidents of the ACL, John Nerbonne
and Mark Johnson. We could not have done the
job without them.
We would also like
to express our gratitude to Dekang Lin (University of Alberta), the Program
Co-Chair of ACL-2002, for sharing tools and resources from last year's meeting
with us, and to the Seminar fuer Sprachwissenschaft at the University of
Tuebingen and the Department of Computer Science of the University of Illinois
at Urbana-Champaign for financial support for administrative and technical
tasks.
Special mention
for their sustained and dedicated help over many months are due to our students,
Scott Wen-tau Yih (Computer Science, University of Illinois) and Julia S.
Trushkina (Seminar fuer Sprachwissenschaft at the University of Tuebingen). Scott installed and maintained the conference management
system which handled electronic submission of papers, the entire reviewing
process and communication with authors and with reviewers.
Julia assisted us with the reviewing process, with the organization
of the program committee meeting, and with the structuring of the conference
program.
ACL'03 Tutorials Co-Chairs:
John Carroll (University of Sussex, UK)
Keh-Jiann Chen (Academia Sinica, Taiwan)
Lauri Karttunen (PARC, USA)
The call for tutorial proposals was prepared 25 September 2002 and circulated by the local organisers soon after. The deadlines were as follows:
Submission deadline for proposal: 16 December 2002
Notification of acceptance: 6 January 2003
Tutorial description due: 17 February 2003
Tutorial course material due: 5 May 2003
Tutorials date: 7 July 2003
We are using the same model for tutorial presenter reimbursement as in previous years: US$500 per session plus $25 per registrant in the range 21-50 plus $15 per registrant in excess of 50. Registration fees for participants are:
Early Late
Regular $100 $120
Student $75 $90
The discount for student tutorial registrations is a change over last year, and is intended encourage students attend the tutorials. Each tutorial participant will receive a copy of the presentation slides. Abstracts of the tutorials are published in the conference proceedings companion volume.
In response to the call for tutorials there were two proposals submitted, both of which the tutorial co-chairs considered to be of high quality. To fill out and balance the programme we approached experts in other areas, two of whom accepted to give tutorials. The final programme is:
Gertjan van Noord (University of Groningen, the Netherlands)
Dan Moldovan (University of Texas at Dallas, USA)
Roxana Girju (Baylor University, USA)
T2: Maximum Entropy Models, Conditional Estimation, and Optimization without the Magic
Dan Klein and Christopher D. Manning (Stanford University, USA)
Roger K. Moore (20/20 Speech Ltd, UK)
The local organisers put up an attractive web page on the conference site to advertise the tutorials
http://www.ec-inc.co.jp/ACL2003/tutorials.html
, containing photos and bios of the tutorial presenters in addition to the tutorial descriptions. At the close of early registration there were 263 tutorial registrations.
Lori Levin, Takenobu Tokunaga, Alessandro Lenci
As a response to the Call for Workshop Proposals (distributed in October, 2002), 15 proposals were received, for a total of 18 workshop days. Thirteen proposals were accepted for 15 workshop days. Two proposals were unified into one workshop. Two proposals were rejected. Later on, two workshops were promoted to associated conference (EMNLP and IRAL), since they are quite large and already well-established. At the last minute, one workshop was canceled. Two of the workshops/associated conferences are organized by ACL SIGs (SIGHAN, SIGDAT).
All CFPs of the workshops were announced around mid-January 2003, after a check by the workshops chair for consistency and uniformity, and were distributed on the ACL workshops website (via the webmaster) and by email via Priscilla Rasmussen. The workshop organizers also distributed the CFPs to the major mailing lists.
A few workshops received external funding. One workshop supported lunch for all the participants and partial travel expenses for student participants.
Number of Papers Number of Early Registrants
Accepted
EMNLP: AC1 28 131
IRAL: AC2 22 48
MLSUM: WS1 11 53
NLBIO: WS2 16 48
LEXFIGf: WS3 8 22
MLNER: WS4 10 34
IWP: WS5 14 56
SIGHAN: WS6 31 41
MWEXP: WS7 13 40
LINGLAN: WS8 6 19
PATENT: WS9 8 18
Our strongest suggestion to the executive committee is to make sure that the conference handbook and reports from previous years are given to all of the conference organizers at the time they are appointed, before they start their tasks. The following problems could have been avoided.
o There was some confusion caused by a delay in distributing the instruction for compiling proceedings to the workshop organizers. In addition, unfortunate miscommunication between the workshop chairs and the publication chairs made the situation worse. There should have been more information sharing between them from the early stage of the process. The tools developed by David Yarowsky are an excellent way of achieving uniformity and efficiency in this process, and should preferably be distributed to the workshop organizers at an earlier stage as well.
o Another point unclear to both of the workshop and publication chairs was the requirements for the copyright transfer agreement forms. We ended up following Priscilla's suggestion that the authors should submit the ORIGINAL designated copyright transfer agreement form with at least the first author's signature. A FAXED form is tentatively accepted for printing but the original should be submitted by the conference at latest.
o Last but not least, we did not know where to find reports from previous workshop chairs when we started this task, and some of us didn't know that such reports existed. We are very sorry we repeated the same mistakes.
Takenobu Tokunaga
Alessandro Lenci
Lori Levin
ACL-03 Student Research Workshop
Kotaro Funakoshi, Sandra Kuebler, Jahna Otterbacher
1. Program Committee
The co-chairs of the ACL-03 Student Research Workshop, Kotaro Funakoshi (Tokyo Institute of Technology), Sandra Kuebler (University of Tuebingen) and Jahna Otterbacher (University of Michigan) were nominated by the general chair, Jun-ichi Tsjuii, and approved by the ACL Executive Committee. Chris Brew (The Ohio State University) was appointed by the ACL Executive Committee as Faculty Advisor. The program committee was formed by the co-chairs and approved by the ACL Executive Committee. The final program committee consisted of 23 student members and 25 non-student members. Of the 48 reviewers, 12 were from North America, 13 from Europe and 23 from Asia.
2. Submissions and
Acceptance
We received 31 submissions from 15 different countries for the Student Research Workshop. The submission length was 6 pages, using the ACL style files. Each submitted paper received three or four reviews, with 30 receiving reviews from two or more non-student reviewers. The reviews were assigned to the reviewers by our submission web site, and reviewers also uploaded their reviews to the site.
We accepted 12 of the submitted papers. Two of the accepted papers were replaced with alternates since their authors were unable to attend the workshop. The length limit for the camera-ready papers was 8 pages.
Statistics for this year's submissions and acceptance are presented below (Tables 1-3).
3. Presentation Format
The Workshop will be organized into three parallel sessions, with each author presenting his or her paper in a 26-minute slot. Continuing the tradition of a panel workshop, we will invite at least two senior researchers to serve as panelists to provide detailed feedback to the student authors. Each presentation will begin with the author's presentation (approx. 18 minutes), followed by the panelists' feedback (4 minutes) and finally, general Q&A (4 minutes).
4. Panelists
The co-chairs are currently in the process of recruiting the panel from the list of those who have pre-registered for the conference.
5. Funding
As in previous years, our faculty advisor obtained funding for student travel from NSF. An application was submitted in early December using the current call for papers and historical information about the workshop. $21,990 was requested and awarded.
We were able to fund all of the 9 students who applied for funding. We received applications from 7 of the 12 authors with accepted papers in the SRW. We received applications from two other student authors (one was the first author of an accepted workshop paper and the second was the second author of an accepted main conference paper).
We offered grants of $2,000 to student authors traveling from North America and Europe and $1,650 to those traveling from Asia, due to the differences in airfare.
6. Problems
Overall, the co-chairs learned a great deal from organizing the workshop and found it to be a rewarding experience. We are grateful to the ACL conference organizers who helped us with many issues such as scheduling and the preparation of the workshop proceedings. We would like to thank the ACL general chair, Jun-ichi Tsjuii, and the ACL Executive Committee for allowing us the opportunity to serve as this year's co-chairs.
Below, we note a few problems we encountered as well as some suggestions for future student organizers.
6.1 Soliciting submissions
Originally, we used the same submission deadline as the main conference (February 26th). However, since we had received very few submissions by that time, we extended our deadline to March 15th. Therefore, we would like to recommend that future co-chairs use a later deadline for workshop submissions, in order to facilitate more submissions.
6.2 Communication between
previous/new co-chairs
We would like to suggest that new student co-chairs be instructed to contact the previous year's co-chairs right away (before sending out the Call for Papers). The previous co-chairs would be able to sugggest possible changes and advice to the new co-chairs. In our case, we are grateful to last year's co-chairs, Alexander Koller and Gideon Mann, for sharing their experiences and materials with us. In addition, ACL co-chairs might also communicate with co-chairs of the EACL and NAACL student workshops. For example, Malte Gabsdil, one of the EACL student co-chairs, helped us immensely by providing us with an electronic paper submission system.
6.3 Reviewing process
In the reviewing process, we made use of a modified version of last year's reviewing form that askedreviewers a series of questions about each paper on a five-point scale. In addition, reviewers were asked to provide open-response feedback to authors. Our form had a section that was explicitly labeled "Comments to the author," which was the only section seen by authors. However, in the future, co-chairs may want to consider making more of the reviewers' comments open to them, in order to facilitate the learning process. We suggest that they consider which items they would like to return back to the authors beforehand, so that they can let the reviewers know explicitly which portions of the form the authors will be able to see. Otherwise, reviewers may refer to their comments in other sections of the review form that are not seen by authors, resulting in some confusion. In addition, some reviewers may not provide any comments in the section "Comments to authors," if they are not made aware that this section is the only one seen by the authors.
TABLES
Table 1: Submissions/acceptance by country (31 subs, 15 countries)
|
Country |
Submitted | Accepted
|
| Australia |
2 |
2 |
| Brazil |
1 |
0 |
| Canada |
2 |
1 |
| France |
3 |
1 |
| Germany |
2 |
2 |
| India |
1 |
0 |
| Japan |
4 |
2 |
| Korea |
1 |
1 |
| Malaysia |
1 |
0 |
| Netherlands |
1 |
1 |
| New Zealand |
1 |
0 |
| Russia |
2 |
0 |
| Sweden |
1 |
0 |
| US |
5 |
2 |
Table 2: Papers by geographical area
| Area | Submitted | Accepted | Acceptance rate |
| Asia |
11 |
3 |
0.27 |
| Europe |
9 |
4 |
0.44 |
| N.America |
7 |
3 |
0.43 |
| S.America |
1 |
0 |
0.00 |
| Oceania |
3 |
2 |
0.33 |
|
Topic |
Submitted | Accepted |
|
Machine Translation |
2 | 0 |
|
Generation |
0 | 0 |
|
Information Extraction |
3 | 1 |
|
Lexicon |
2 | 0 |
|
Models of Language |
1 | 1 |
|
Corpus-based Language Modeling |
1 | 0 |
|
Morphology |
0 | 0 |
|
Syntax |
1 | 1 |
|
Discourse/Dialogue |
6 | 4 |
|
Information Retrieval |
4 | 1 |
|
Semantics |
6 | 3 |
|
Speech Recognition/synthesis |
0 | 0 |
|
Summarization |
0 | 0 |
|
Message & narrative understanding |
0 | 0 |
|
Pragmatics |
0 | 0 |
|
Question Answering |
1 | 0 |
|
Multi-lingual Processing |
0 | 0 |
|
Language in Multi-modal Systems |
2 | 1 |
|
Text Classification |
1 | 0 |
Hiromu Hayashi (Fujitsu Laboratories Ltd.)
Sponsorships & Exhibits:
We had several
levels of sponsorships and exhibitors-- Platinum, Gold, Silver and Bronze
sponsorships. We had 2 Platinum sponsors, 5 Gold sponsors, 7 Silver sponsors
and 15 Bronze sponsorships. The lowest Sponsor category, for \100,000($850),
was newly established this year, and most of the universities' sponsors signed
up for that category. The total value of sponsorships was \8,100,000($69,000).
Several companies who had been sponsors in the past were able to sponsor
this year's meeting. This was more than twice of the last meeting($31,000).
Points for discussion:
1. Many Bronze sponsors
Sponsorships
and exhibits are given equal privileges. So, we had received many exhibitors
from non-profits(universities and national research institutes) as follows.
Private
companies
Non-profits
Platinum
2
0
Gold
4
1
Silver
5
2
Bronze
1
14
Two Platinum
sponsors are Just System and Microsoft. Five Gold sponsors are OKI, Fuji
Xerox, NTT, IBM and NII. Seven Silver sponsors are Fujitsu, NEC, Toshiba,
Matsushita, Hokuto Systems, CRL and Tokyo Univ.
2. Just one foreign entry for exhibitors
We had 28
entries from domestic(Japanese) organizations, but just one entry from a
foreign exhibitor(LanA Consulting). We have only put an advertisement to
Japanese organizations, but we would also have to make an advertisement to
overseas organizations.
David Yarowsky and Sadao Kurohashi
ACL-2003 Publications Chairs
The publications chairs were responsible for (a) the physical production of all 13 ACL-2003 proceedings, (b) the design, data collection and creation of the CDROM, (c) jointly coordinating all proceedings preparation with the relevant program and workshop chairs, and (d) designing and compiling the companion volume, exhibition brochures and other general publications
In total, 19 hard copy publications were processed in Tokyo:
Main conference proceedings
Companion volume (student research workshop, poster/demo, and tutorial abstracts)
CD-Rom
Exhibition brochure
9 workshop proceedings
2 associated conference proceedings
4 tutorial handouts
For the main conference and the companion volume, their front material (toc, index, program, preface) were organized by the publication chairs. Toc, index, program were automatically generated by 'db' file, which is based on authors' electronic bibliography submission. Hard-copy papers were sent to the local publication chair from the authors directly, and then their order was arranged and the consistency was checked with the toc and index by the local publication chair (page numbering was done by the printer).
Daisuke Kawahara and Nobuko Machii greatly helped process these.
9 workshop proceedings and 2 associated conferences proceedings were fully prepared by workshop/conference organizers, then sent to the local publication chair and handed to the printer.
The initial publication time-schedule was designed to have a margin of two weeks: one week before handing manuscripts (including CD-ROM) to the printer, and one week after sending everything to Hokkaido. However, all of the two weeks were exhausted, because of several late submissions: some papers from the authors, some hard-copy proceedings from workshop/conference organizers, and some requests of correction and replacement (so, everything will arrive at Hokkaido just on time).
ACL-2003 CDROM:
The ACL-2003 CDROM constitutes a complete archival record of all 275 papers (in both PDF and postscript) for the main conference, its companion volume, its 2 associated conferences and 9 workshops all in a unified format.
The CDROM also contains Bibtex citation entries for all papers, PDF versions of all papers stamped with citation information, complete copies of schedules and other cover material in multiple formats and a unified author index to all events. The CDROM also contains a complete archive of all ACL-2003 conference and workshop web pages, current as of June 8, 2003.
All tables of contents, index pages and bibtex entries on the CDROM were created automatically using a set of tools developed by David Yarowsky for the first ACL CDROM in 2000, and used for most ACL, NAACL and HLT conferences since then. Improvements this year include the generation of ACL anthology structures, and a web interface for paper and bibliographic data submissions by authors. This interface generated a master paper information database ('db' file), over which additional tools more-or-less automatically generated the table-of-contents, author indexes and meeting schedules for both the hardcopy proceedings and CDROM, as well as converted automatically to ACL Anthology format.
The design of the CDROM was based heavily on an original 2000 design by Silviu Cucerzan. Silviu also created and implemented the final design for the CDROM banner image. The original CDROM banner design, and the design and implementation of most of the archived website pages on this CDROM are due to Eiko Yamamoto, ACL-2003 webmaster. The automatic stamping of the PDF files with citation information was accomplished using programs designed by Jason Eisner. Special thanks are due to Drago Radev for his numerous contributions to the ACL CDROM creation processes over the past 2 1/2 years.
Masashi Okamoto did a very nice job designing the CDROM jewel case and cover. Yoji Kiyota physically burned this CDROM from an archive finalized in Baltimore on June 11, 2003, and he also helped with several other aspects of its preparation, including format and link verification. Daisuke Kawahara helped prepare electronic front materials for several volumes of the proceedings.
Additional special thanks are due to the ACL-2003 program chairs, all workshop organizers and the interactive poster session chair for providing all of the proceedings preface materials, verifying and augmenting the standardized 'db' files, and for their careful proofreading of their sections of the CDROM. Numerous others have contributed to this CDROM in important ways. They include Jun'Ichi Tsujii (for his very welcome support and advice), Takenobu Tokunaga (who helped coordinate all workshop material submissions), Hitoshi Isahara (for the multi-faced work of his publicity committee), Steven Bird, Priscilla Rasmussen, Kenji Araki and other members of the ACL-2003 committees.
David Yarowsky and Sadao Kurohashi
ACL-2003 Publications Chairs
Yuji Matsumoto
The aim of the sessions:
The aim of these sessions is to provide a forum of academic and technical exchanges of researchers and developers of Natural Langauge systems. We called for original and significant work in progress, on-going research projects with novel ideas/applications, and late-breaking results that are best communicated in an interactive environment.
The Program Committee:
We organized an international Program Committee of 28 distinguished researchers.
Paper Submission Format:
The submission format is four pages in the same format for ACL main conference. We strongly recommended electronic submission either in PDF or PS format.
Reviewing Process:
There were 62 paper submissions, of which one was withdrawn during the review process because it was accepted by a different conference. Each paper were read by at least two reviewers, and when there were conflicts in the review results I asked the reviewers to talk over emails. The PC finally selected 28 papers. Later, one paper was withdrawn because the author of the paper cannot attend the meeting.
Session Format:
We allocated to each paper four pages in the Companion Volume, and in the meeting each paper receives a five minute preview presentation and two and half hour presentation of poster with/without demonstration. The previews and posters/demos take place in parallel but in separated rooms.
Problems:
As we specified the submission format either in a PDF or PS file, some of the authors send the papers as a Word file. We decided to accept such papers as is.
We didn't clearly specify the rule of anonymity. This caused confusion to the authors and some papers are send with the names of the authors on the first page. We also accepted such a paper as is.
===============================================================
Hitoshi Ishara
(1) November, 2002, we printed 300 posters and 1000 flyers. Most of them were sent to American, European and Asian institutes. The rest of the flyers are distributed at the workshop and meetings on NLP.
(2) We issued "ACL2003 News Letter" and distributed them via several related mailing lists
No.1 (27th of December, 2002)
No.2 ( 7th of February, 2003)
No.3 ( 4th of March, 2003)
No.4 (15th of April, 2003)
No.5 ( 3rd of June, 2003)
Reminder: Early Registration (12th of June, 2003)
No.6 (will be issued on the 24th of June, 2003)
(3) Maintenance of the ACL2003 Homepage
We constantly maintain most of the contents on the Web.
Note that there is also
information about previous meetings in the NAACL and EACL reports.
Donia Scott
and Toni Badia
Donia
Scott
ACL'04 will be held in Barcelona from July 21st to 26th, 2004.
It will be hosted by a committee composed of computational linguists from
several institutions in Barcelona, chaired by Prof Toni Badia of University
Pompeu Fabra. The Conference Chair will be Prof Donia Scott of the University
of Brighton. All technical chairs are in place:
Programme Chairs: Walter
Daelemans and Marilyn Walker
Tutorial Chair:
Inderjeet Mani
Workshop Chair:
Srinivas Bangalore
Demos Chairs:
Philippe Blache and Horacio Rodriques
Publications Chairs:
Owen Rambow (main), Sergei Balari (local), Sadao Kurohashi (ACL'03)
Publicity Chair:
Felisa Verdejo
Exhibition Chair:
Toni Mart’
The search for sponsorship chair(s) is still on; volunteers or nominations
are welcome.
There will be a number of other significant CL-focussed conferences in 2004:
LREC (in Lisbon in the Spring), NAACL (in the US in the Spring) and COLING
(in Geneva in August), and we are in discussion with the organisers of these
conferences to coordinate our respective deadlines. The deadlines for LREC
and NAACL will be in the Autumn, and so will not be problematic. Our
tentative schedule for deadlines is:
Preliminary Call for Papers: September
Paper submission deadline: 25 February 2004
Notification of Acceptance: 26 April 2004
Camera ready papers due: 31 May 2004
Main session: 22-24 July 2004
<note: the proposed
submission deadline for COLING is April 28th>
Submission Deadline for
Proposals: 1 February 2004
Notification of acceptance:
19 February 2004
Tutorial Announcements due:
19 March 2004
Tutorial Notes due: 19 April
2004
Tutorials: 21 July 2004
Submission Deadline for
Proposals: 16th February 2004
Notification
of acceptance: 8th March 2004
Workshops: 25-26th
July
Submission Deadline: 1st
April
The PC co-chairs have expressed the wish to increase
the acceptance rate in the main session, with some papers being presented
as posters (and also included in the main proceedings).
Toni Badia and other local
committee members
We are planning for ACL 2004 to be held at
the new Barcelona Forum Convention Center, which is scheduled to be completed
in December 2003, officially opening in April, 2004.
The ACL meeting is included in the programme
of the Forum of Cultures that will take place in Barcelona from April to
September 2004. As part of this event, ACL obtains a number of benefits, including
the use of the rooms and other spaces needed free of charge. ACLÕ04
is nevertheless going to be charged the amount of 10Û per day and person
registered; this fee will allow participants to enter the Forum precinct
and visit and use the facilities, exhibitions, etc., within the precinct.
ACL, Barcelona 2004 and FUPF are going to sign a cooperation contract in order to fix their respective obligations concerning the ACL conference. From April to June we have been discussing the terms of this contract. Basically,
Rooms reserved at the Barcelona Forum Convention
Center are as follows:
|
DATE |
21.07.04 |
22.07 |
23.07 |
24.07 |
25.07 |
26.07 |
||||||
|
|
MORN ING (M) |
AFTER NOON (A) |
M |
A |
M |
A |
M |
A |
M |
A |
M |
A |
|
Rooms of congress |
|
|
|
|
|
|
|
|
|
|
|
|
|
ROOM A1
(650 pax) |
|
|
P |
P |
P |
P |
P |
P |
|
|
|
|
|
ROOM A2
(650 pax) |
|
|
|
|
|
|
|
|
|
|
|
|
|
ROOM B1
(330 pax) |
|
|
P |
P |
P |
P |
P |
P |
P |
P |
P |
P |
|
ROOM B2
( Ò ) |
P |
P |
P |
P |
P |
P |
P |
P |
|
|
|
|
|
ROOM B3
( Ò ) |
|
|
P |
P |
P |
P |
P |
P |
|
|
|
|
|
ROOM B4
( Ò ) |
|
|
|
|
|
|
|
|
|
|
|
|
|
ROOM B5
( Ò ) |
|
|
|
|
|
|
|
|
|
|
|
|
|
ROOM C1
(75 pax) |
P |
P |
|
|
|
|
P |
|
P |
P |
P |
P |
|
ROOM C2
( Ò ) |
P |
P |
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C3
( Ò ) |
P |
P |
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C4
( Ò ) |
P |
P |
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C5
( Ò ) |
P |
P |
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C6
( Ò ) |
P |
P |
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C7
( Ò ) |
P |
P |
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C8
( Ò ) |
|
P |
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C9
( Ò ) |
|
|
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C10
( Ò ) |
|
|
|
|
|
|
|
|
P |
P |
P |
P |
|
ROOM C11
( Ò ) |
|
|
|
|
|
|
|
|
P |
P |
P |
P |
In addition, ACL will have:
Two rooms for secretarial activities
Space for exhibitions and posters
Two rooms with PCs and Internet connections for demonstrations
Space for the reception (700 attendants foreseen)
Space for the banquet (700 attendants foreseen)
Audiovisual equipment (Internet, projectors,
OHP) and technical assistance in rooms will also be available.
Planned schedule of ACL 2004:
July 21, 2004: tutorial
sessions
July 22 - 24, 2004: main
conference (3 concurrent sessions)
July 25 - 26, 2004: 11
workshops
The details of social events such as banquets
and receptions have not been fixed yet; we are exploring different possibilities.
There will be plenty of social activities going
on those days in Barcelona, many of them within the Forum precinct. We will
advertise them in due course.
The Local, Regional and Spanish Governments
contribute to ACLÕ04 through their participation in the funding of
the Forum 2004 event.
We expect to obtain some subsidies from the
Catalan and Spanish research agencies. The figures cannot be foreseen in
advance and we have not taken them into account in the provisional budget
we are presenting below.
A conference Website is currently under development
(URL
to be announced soon.). It is going to include information on accommodations
and Barcelona.
This is a difficult part because of the early
reservation that is going to be needed in order to guarantee the rooms.
At the moment we are negotiating with the Forum
organisers to have rooms blocked for ACLÕ04 attendants until a reasonable
date (early spring). Alternatively we are looking at the cost having a private
company blocking the rooms for us; the cost and the date of relieving the
blocking have not been determined yet.
This applies to both standard hotel rooms and
rooms in students residences and hostels.
Costs for student accommodation (2003 prices)
are:
single room: 40Û
double room: 75Û
700 attendees
Registration Fee
Regular 270Euro x 500 =
135,000Euro
Student 120Euro x 200 =
24,000Euro
total 159,000Û
We are planning the following registration
fees:
regular
early: 270Euro
late:
300Euro
on
site: 350Euro
student early: 120Euro
late:
140Euro
on
site: 160Euro
Entrance to Forum2004
25 000Euro
Financial handling and legal advice (FUPF)
12 460Euro
Handling of credit card registration
3 975Euro
Costs of PCO contracts or costs required for
pre-/on-site registration
15 659,88Euro
Cost of blocking rooms
??
Printing
14 000Euro
(Including Proceedings, Circulars, etc.)
Telephone and Internet connections
??
Hire of poster panels
??
(Including signs, booths, lighting equipment,
etc.)
Costs for invited speakers (4500 x 3)
13 500Euro
Estimated Local Committee expenses
21 000Euro
(Including costs for expendable supplies, conventions
bags, bookbinding, communications, transportation tickets, web elaboration
and maintenance, etc.)
Meeting of Program Committee
4 300Euro
Signs, banners, etc.
2 150Euro
Refreshments (coffee, tea, juice, cakes, etc.)
13 482Euro
Welcome Reception costs
12 425Euro
Total (with some expenses pending)
137 952Euro
We may be able to offer the ACL-04 participants:
- reduction on flight prices (if they fly with IBERIA)
- reduction on town transportation tickets
Johanna Moore
I broadcast a request for bidders to host ACL/NAACL 05 shortly after ACL 02. Avoiding the saga of how each promising bid shriveled to nothing, I can now report that I have the following weak leads:
1. Michigan -- still a possiblity, but there are personal problems for one of the key organisers
2. San Antonio -- this is actually Jill Burstein of ETS in Princeton.
They apparently have conference facilities in San Antonio. She's considering it, but I really think this is a LONG shot.
3. Brandeis -- James P indicated an interest. However, I have not actively pursued this as NAACL 2004 is going to be in Boston and so I've been discouraged from holding 2 ACL confs in a row in Boston.
I've contacted 25+ sites and have got a resounding "no" in most cases. (There were a few who did not respond to repeated email.)
I really think we have to re-think whether we should just run the conference with PCOs in the city that puts up the best bid, independent of academia. I think academics are burned out, and that ACL has gotten too big for anyone to consider hosting without a lot of admin help, which most Universities can't provide.
This is not something we can ponder indefinitely. We may need to go this route for ACL 05. I was contacted about 6 months ago by an independent conf org service. Unfortunately I can't seem to find this email now, but of course I have no reason to assume this service was any good. In any case, I think the exec need to discuss this at Sapporo. I have no idea how we'll proceed if no one comes forth to do ACL 05. I can't possibly be expected to organize or manage a conference in NA from Edinburgh, so who's going to deal with it?
Johanna Moore
Conference Report, June 2003
Eduard Hovy, USC/ISI <hovy@isi.edu>
HLT-NAACL03 was an attempt to prevent the appearance of two Language processing conferences in North America every spring. The conference fused the HLT and ACL conference styles and traditions (e.g., plenary demos, short and long papers from the former, and student workshop and tutorials from the latter).
Tuesday May 27, day: Tutorials and Reception
Wednesday-Friday May 28-30: Main conference
plenary demos on Wednesday evening
banquet on Thursday evening
Saturday-Sunday, May 31 and June 1: One- and two-day workshops
Despite losing some attendance to SARS, the conference attracted approx. 435 attendees, including people from Asia and Europe. Its very high standard of papers, coupled with unusually large number of tutorials (7) and workshops (9), helped establish the conference at the highest technical rank. The invited talk by Prof. Elissa Newport on language learning, and the panel on the Surprise Language Exercise, were well received.
Unfortunately the conference did not succeed in attracting significant numbers of papers or attendees from either the SIGIR or Speech communities. Increasing outreach will be a goal of HLT-NAACL04.
A full report of the conference is available from the conference chair.
Summary:
All of the SIGS have been active over the past year, sponsoring SIG-related conferences and workshops, participating in resource gathering and dissemination, and providing valuable information to their respective audiences through websites.
Several of the SIGs bring up issues about building relationships to other organizations, or expanding the SIGÕs charter. The Exec may wish to provide some guidance regarding these issuess. In particular:
SIGLEX: Adam Kilgariff is exploring the possibility of making EURALEX a sister organisation to ACL, so that, e.g., there is discount membership for people who are members of both (precedents and pointers most welcome).
SIGMedia: SIGMedia members have been involved in the preparation of a Human-Machine Interaction Network on Emotions (HUMAINE). If this proposal is funded by the European Commission, SIGMEDIA will organize workshops and tutorials in cooperation with this network.
SIGPHON: There is no ACL SIG devoted to computational morphology. SIGPHON's most recent workshop covered morphological and phonological learning, and SIGPHON may consider expanding our charter to cover morphology as well as phonology. This is arguably a natural interest group, as the two problems are somewhat intertwined, especially when dealing with corpus data. Consolidating the two communities would increase submissions and attendance at our workshops. SIGPHON is interested in the thoughts of the ACL Exec on this question.
SIGDAT (Ken Church, David Yarowsky)
Kenneth W. Church , kwc@research.att.com; David Yarowsky, yarowsky@cs.jhu.edu
SIGDAT - 2003 Annual Report
SIGDAT is ACL's special interest group for linguistic data and corpus-based approaches to NLP.
In 2003, SIGDAT organized the 2-day Conference on Empirical Methods in Natural Language Processing (EMNLP-2003). The meeting was scheduled immediately after ACL-03 in Sapporo on July 11-12 at the Sapporo Convention Center.
Michael Collins served as conference chair and Mark Steedman served as co-chair. 28 papers were accepted for presentation out of 121 submissions (a 23% acceptance rate), equivalent to the main ACL sessions. 280 copies of the proceedings were printed.
In 2002, SIGDAT organized the 2-day Conference on Empirical Methods in Natural Language Processing (EMNLP-2002). The meeting was held immediately prior to ACL-02 at the University of Pennsylvania in Philadelphia on July 6 and 7.
Jan Hajic served as conference chair and Yuji Matusumoto served as co-chair. 35 papers were accepted for presentation out of 142 submissions (a 24.6% acceptance rate), yielding a program and proceedings equivalent in size to full ACL conferences before 1997. Over 230 people attended.
SIGDAT also co-sponsored the Sixth Conference on Natural Language Learning, with SIGNLL, held in Taipei on August 31-September 1, 2002. Dan Roth, Antal van den Bosch and Erik Tjong Kim Sang were co-chairs.
David Yarowsky, Secretary-Treasurer
SIGDIAL (Laila Dybkjaer)
Laila Dybkjaer , laila@nis.sdu.dk
SIGDIAL 2003 Annual Report
SIGdial is the ACL and ISCA Special Interest Group on Discourse and Dialogue which was formed in November 1997. This report covers the period from July 2002 to June 2003.
PEOPLE:
SIGdial is being run by the following team: President: Laila Dybkj¾r; Vice President: David Traum; Secretary: Jennifer Chu-Carroll; Science Advisory Committee: Jan Alexandersson, Kristiina Jokinen, Akira Kurematsu, Alexander Rudnicky, Ronnie Smith, and Jan van Kuppevelt; Information Officer: Claudia Soria assisted by Kristiina Jokinen, Eric Chang, and Pamela Jordan; ACL Student Liaison: Weiqun Xu and Holmer Hemsen; CHI Liaison: Nils DahlbŠck; ICSLP/ISCA Liaison: Julia Hirschberg; Mailing list maintenance: Laurent Romary.
Elections are held every second year and the next elections will take place this summer/autumn.
SIGdial has about 300 members registered on its membership web page coming from at least 20 different countries across the world.
WORKSHOP ENDORSEMENT
During the past year several workshops with strong relations to the SIGdial field of interest have been endorsed by SIGdial. The workshops endorsed by SIGdial which have taken place within the reported period include:
September 4-6 2002: EDILOG 2002: Sixth Workshop on the Semantics and Pragmatics of Dialogue (Edinburgh, Scotland).
March 24-26 2003. AAAI Spring Symposium on Natural Language Generation in Spoken and Written Dialogue (Palo Alto, CA, USA).
Workshops which have been endorsed by SIGdial but havenÕt been held yet include:
July 20-21 2003: AI in Education 2003 Workshop on "Tutorial Dialogue Systems: With a View Towards the Classroom" (Sidney, Australia).
August 28-31 2003: ISCA Tutorial and Research Workshop on Error Handling in Spoken Dialogue Systems (Chateau-d'Oex-Vaud, Switzerland).
September 4-6 2003: DiaBruck 2003: Seventh Workshop on the Semantics and Pragmatics of Dialogue (SaarbrŸcken, Germany).
SIGDIAL WORKSHOP SERIES
Following the first and successful SIGdial workshop on Discourse and Dialogue held on 7-8 October 2000 in Hong Kong in conjunction with ACLÕ2000, and the second SIGdial workshop held on 1-2 September 2001 in Aalborg, Denmark, in conjunction with EurospeechÕ2001, the third SIGdial workshop took place on 11-12 July in Philadelphia, USA, in conjunction with ACLÕ2002. This workshop was a successful event. The workshop received 38 paper submissions. Of these 19 were accepted for presentation as long papers and 9 as short papers. The online workshop proceedings are available at http://www.mlab.uiah.fi/~kjokinen/sigdial/publications.htm
Hardcopies can be bought via ACL.
The fourth SIGdial workshop will take place on 5-6 July in Sapporo, Japan, right before ACLÕ2003. Information on SIGdial workshops can be found at
http://www.sigdial.org/workshops.php
WEB
SIGdial has a web site at
The web site is hosted by NISLab. Some pages are maintained by a particular person, such as the information officer, while other pages are intended to be maintained by the membership collectively. For instance, the web site enables members to easily update their personal information in the membership database. Also a resource page has been established which members are encouraged to collectively maintain by adding references to resources they are aware of and which may be of interest to the SIGdial community. This page includes references to e.g. coding schemes, language resources, spoken dialogue systems, and tools.
OTHER ACTIVITIES
Since autumn 2000 SIGdial has a collaboration with Elsnet, the European Language and Speech Network. Elsnet has a quarterly newsletter, Elsnews, and Elsnet has generously offered one page in each issue of Elsnews for SIGdial purposes. SIGdial contributions in Elsnews can be found in issue 9.3 and onwards. Elsnews can be freely downloaded in pdf format from
http://www.elsnet.org/elsnews.html.
Contributions from the reported period include
A flyer on SIGdial was produced for distribution at ICSLP'2002 and the COCOSDA workshop. The flyer is available for download in pdf format from www.sigdial.org.
SIGdial has a business meeting each year in connection with the SIGdial workshop. Agenda, slides and minutes are available at http://www.sigdial.org/businessmeetings/
Laila Dybkjaer, president
===========================================================
SIGGEN (Owen Rambow)
Owen Rambow, rambow@cs.columbia.edu
SIGGEN report 2003
CONFERENCES AND
WORKSHOPS: The major NLG event of the year was INLG-02, which SIGGEN sponsored
and supported. It was held in Columbia University's
Arden House, near New York City, just before ACL (which took place in Philadelphia). While submissions were as in previous years (with,
as in previous years, an acceptance rate of about 50%), attendance was below
expectations (about 75 paying attendants, down from the record 100 at the
workshop in 1998, with the first conference in Israel in 2000 having a lower
attendance, presumably due to the location). An
attempt to extend the audience by having special paper tracks on both dialog
and summarization were not successful, perhaps due to lack of sufficient
publicity outside of the standard NLG communities, and/or lack of NLG-related
work in these areas outside of the NLG community (which would submit anyway). A further factor for low attendance may have been
the co-location with ACL, from which it was separated by the July 4 holiday.
It was decided
that INLG04 will be organized by a group at the University of Brighton and
take place near Southampton, England.
There has been
a request by Susan Haller to co-sponsor a special track on user modeling
and HCI approaches in natural language generation at FLAIRS.
MAILING LIST AND
WEB SITE: The web site needs to be updated.Helmut Horacek and Tilman Becker
have jointly taken over this task.
ELECTION: A contested
election was held for new board members. Tilman Becker (SaarbrŸcken) and
Marilyn Walker (Sheffield) replaced Maria Milosavljevic and Ehud Reiter as
permanent board members, and Stephen Wan (Sydney) replaced NoŽmie Elhadad
as student member.
SIGHAN (Richard
Sproat)
Richard Sproat, rws@research.att.com
SIGHAN Report 2003
In 2003 SIGHAN is sponsoring two events:
The first is the 2nd SIGHAN Workshop on Chinese Language Processing to be held in conjunction with ACL 2003, in Sapporo, Japan, July 11-12. The CFP and description of the workshop can be viewed at
The following is the breakdown of authors of accepted workshop papers by region:
Mainland China: 8
Taiwan: 2
Singapore: 1
Japan: 1
USA: 5
Europe: 2
------------------
Total: 19
The second event was the First International Chinese Word Segmentation Bakeoff. While there have been several contests comparing word segmentation methods on standard corpora held within Mainland China, this is the first event of its kind to be held internationally. Details of the bakeoff can be found at
http://www.sighan.org/bakeoff2003/.
The bakeoff included four sets of training and test corpora from four different institutions. Twelve sites participated, broken down geographically as follows:
Mainland China: 2
Taiwan: 1
Singapore: 1
Japan: 1
Hong Kong: 3
USA: 4
The corpora used in the bakeoff, as well as the scoring software, are being made available for research purposes via the SIGHAN website. The results of the bakeoff will be presented at the SIGHAN Workshop.
Richard Sproat
SIGLEX (Adam Kilgariff)
Adam Kilgarriff, Adam.Kilgarriff@itri.bton.ac.uk
SIGLEX Report 2003
Since ACL 2002 SIGLEX has endorsed and contributed to five workshops: IWCS5 (Computational Semantics) in Tilburg in January (with SIGSEM), The Generative Lexicon workshop in Geneva in May, the NAACL workshop on learning word meaning from non-linguistic data, and two workshops (Figurative language and Multi-Word Expressions) at Sapporo. We are also endorsing and contributing to the Second Global WordNet conference in Brno next January.
The planning of SENSEVAL-3 has continued, under the most able leadership of Phil Edmonds and Rada Mihalcea. The evaluation will take place in early 2004, with the workshop probably at ACL-04 in Barcelona. (SENSEVAL is, constitutionally, a sub-committee of SIGLEX.)
Last summer I was elected onto the EURALEX Board and I am planning to explore making EURALEX a sister organisation to ACL, so that, e.g., there is discount membership for people who are members of both (precedents and pointers most welcome).
Ken Litkowski, as SIGLEX webmaster, has improved the website and provided indexing into the CORPORA mailing list archive.
Elections and a new Chair are due: this will not happen this year now, as there are no further relevant events, and I have been unable to attend any of the possible ones in 2003. The new Chair and committee will be put in place at the SENSEVAL workshop next year.
Adam Kilgarriff
SIGMEDIA (Elisabeth Andre)
Elisabeth Andre«,andre@informatik.uni-augsburg.de
SIGMEDIA ANNUAL REPORT 2003
SPECIAL INTEREST GROUP ON MULTIMEDIA LANGUAGE PROCESSING
June 15th 2003
CHAIR: Elisabeth Andre (University of Augsburg, Germany, andre@informatik.uni-augsburg.de)
MAILING ADDRESS: sigmedia@dfki.de
CONFERENCES AND WORKSHOPS:
In the last reporting period, SIGMedia has been serving as a sponsor for two new events at the crossroads of computational linguistics and virtual agents.
SIGMEDIA has been serving as a sponsor of the 4th International Working Conference on Intelligent Virtual Agents which will take place at Kloster Irsee, Germany from Sept. 15th-17th 2003. The organizers of the Conference are: Ruth Aylett (University of Salford, UK), Daniel Ballin (Radical Multimedia Lab, BTexact Technologies, UK), Jeff Rickel (USC Information Sciences Institute, CA) and Thomas Rist (DFKI, Germany). The URL of the working conference is:
The 2nd new event, SIGMEDIA is currently preparing is the Dagstuhl Seminar on Evaluating Embodied Conversational Agents, Dagstuhl Germany, March 15th-19th, 2004. The organizers of the Seminar are: Dr. Zs—fia Ruttkay (CWI, Amsterdam, The Netherlands), contact person, Prof. Dr. Elisabeth AndrŽ (University of Augsburg, Germany), Prof. Dr. Kristina Hššk (SICS and IT-University, Sweden), Prof. Dr. W. Lewis Johnson (USC/ISI, USA), Prof. Dr. Catherine Pelachaud (University of Paris 8, France). The URL of the Seminar is:
In addition, SIGMEDIA successfully conducted two events whose organization already started in the preceding reporting phase. The 2nd Conference on Computational Semiotics for Games and New Media, Augsburg (Germany), Sept. 2nd-4th 2002,
http://www.kinonet.com/conferences/cosign2002
was attended by computer scientists, artists and media specialists from Europe, North and South America and Asia to discuss phenomena of coding meaning in a digital form.
The ISCA Tutorial and Research Workshop on Spoken Dialogue in Mobile Environments, Kloster Irsee, Germany, June 17th-21st, 2002,
http://www.sigmedia.org/ids02/
attracted an unexpectedly high number of participants both from academy and industry. Due to the great success, a follow-up event with a slightly different focus is planned for 2005.
COOPERATIONS WITH OTHER NETWORKS AND INSTITUTIONS:
SIGMedia members have been involved in the preparation of a Human-Machine Interaction Network on Emotions (HUMAINE). A proposal for funding has successfully passed the first round of the reviewing process. In case it will be accepted by the European Commission, SIGMEDIA will organize workshops and tutorials in cooperation with this network.
Elisabeth Andre, chair
SIGMOL (Dick
Oehrle)
Dick Oerhle,
oehrle@linc.cis.upenn.edu
SIGMOL Report 2003
During the past year, the principal activity of the Mathematics of Language (MOL) SIG was organizing the eighth Mathematics of Language Conference, which was held June 20-22, 2003, in Bloomington, Indiana, in conjunction with the North American Summer School in Logic, Language, and Information. The conference included:
* thirteen submitted papers on a broad range of topics;
* an invited lecture by Aravind Joshi on connections between formal language theory and the analysis of
the structure of complex biochemical structures;
* an invited lecture by Ed Keenan & Ed Stabler on A Mathematical Theory of Grammatical Categories
(coinciding with the final lecture of their NASSLLI course);
* a symposium on Language and Game Theory, with presentations by Robert van Rooy, Rohit Parikh, and
Martin Nowak, and commentary by Larry Moss;
* a symposium on Statistical and Symbolic Approaches to Natural Language Learnability, with presentations
by Makoto Kanazawa, Ed Stabler, Jason Eisner, and Dan Osherson.
Further information about the conference, including the details of the program, is available at:
http://www.cis.upenn.edu/~ircs/mol/mol8program.{pdf,html}
The preprinted proceedings of the conference are available at:
http://www.cis.upenn.edu/~ircs/mol/mol8prcd.{pdf,ps.gz}
A special issue of the Journal of Logic, Language, and Information containing a selection of the papers from MoL8 will appear sometime in 2004.
The incoming President of MOL is James Rogers (jrogers@cs.earlham.edu). An election will be held in the near future to select the incoming Vice-President/President-Elect.
(submitted by Dick Oehrle)
SIGNLL (Walter Daelemans)
Walter Daelemans, walter.daelemans@uia.ua.ac.be
SIGNLL Report 2003
ACL SIGNLL - President's Report 2002-2003
In 2002 current president and secretary Walter Daelemans and Dan Roth extended the SIGNLL advisory board in order to increase geographical and thematic diversity. The advisory board now consists of Thorsten Brants, Claire Cardie, James Cussens, Ido Dagan, Diane Litman, Rada Mihalcea, Yuji Matsumoto, Raymond Mooney, John Nerbonne, Hwee-Tou Ng, Grace Ngai, Miles Osborne, David Powers, Adwait Ratnaparkhi, Ellen Riloff, and Erik Tjong Kim Sang. Our SIGDAT Liaison Representative remains David Yarowsky, and our Information Officer Antal van den Bosch. In the autumn of 2003, new elections for president and secretary will be held.
In 2002, SIGNLL has an estimated 350 members (judging from the on-line information, and controlled for inactive email addresses). The SIGNLL home page is located at
URL http://www.aclweb.org/signll/
The goals of the SIG (promoting of and informing about research on computational language learning) are served by the following two activities: (i) the maintenance of an informative and up-to-date website and associated mailing list, and (ii) the organization of at least one yearly event (CoNLL), and support of other related activities.
The web-pages, maintained by Antal van den Bosch, remain an important source of information, complemented by an email list for conference announcements. On the web-site, links can be found to relevant associations, networks, research cooperations, research departments, groups, institutes, individuals, 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.
The main events in 2002-2003 were the sixth and seventh CoNLL (SIGNLL Conference on Natural Language Learning). More information about these events can be obtained from their web-pages, linked from the SIGNLL website.
CoNLL-02 was organized August-September 2002 with COLING in Taipei by Dan Roth and Antal van den Bosch. It was a successful event, with 45 submissions of which 21 could be accepted, and an invited talk by John Lafferty (sponsored by STIL, http://ilk.kub.nl/~stil). A shared task on language-independent named entity recognition was organized by Erik Tjong Kim Sang as part of the CoNLL, to which 12 systems participated. 65 persons registered officially for the workshop.
The seventh CoNLL was organized with HLT-NAACL May-June 2003 in Edmonton by Walter Daelemans and Miles Osborne, with a shared task organized by Tjong Kim Sang and Fien De Meulder, again on language-independent named entity recognition. 35 papers (for the first time full papers instead of long abstracts) were submitted of which 18 were accepted for presentation and publication in the proceedings. 16 systems were submitted for the shared task, and their descriptions included in the proceedings. An invited talk was given by Steven Abney (sponsored by CLIF, http://clif.uia.ac.be) and attendance was an all time high for CoNLL with 85 official registrations.
We think SIGNLL is still unique in its multiple focus (computational models of language learning both for language engineering and for testing psycholinguistic and linguistic theories; formal and empirical aspects of learning of both artificial and natural languages). We keep striving for complementarity with related SIGDAT events such as EMNLP, and have contributed to interdisciplinary communication by our conference colocation policy.
Walter Daelemans
Antwerp, June 4, 2003
SIGPARSE (Harry Bunt)
SIGPARSE 2003 report
In the period of June 2002 - June 2003, the activities of SIGPARSE have, as usual, focussed on matters concerning the biennial `International Workshop on Parsing Technologies' (IWPT) series. Ensuring the continuity of this series is the main aim of SIGPARSE.
First, the 8th International Workshop on Parsing Technologies (IWPT'03) was held from 23 to 25 April at INRIA-LORIA in Nancy, France. Gertjan van Noord was program chair; Guy Perrier local arrangements chair, and Harry Bunt general chair. Invited speakers were Eugene Charniak and Gregory Grefenstette. Both the quantity and the quality of the submitted papers was satisfactory, and gave rise to a high-level workshop program. The attendance was in the range that is usual for IWPT: between 60 and 70. The workshop proceedings were available at the start of the workshop and were sold out at the end, but the local organisers planned to print an extra set of copies. The first preparations have started for IWPT'03, which is planned to be held in North America.
Second, effort has been spent on putting a book together on the basis of selected papers from IWPT 2000, which took place in Trento, Italy, in February 2000, and IWPT'01, which took place in Beijing in October 2001. This book is edited by Harry Bunt, John Carroll and Giorgio Satta, and is scheduled to be published by Kluwer Academic Publishers in 2003.
Harry Bunt, June 2003.
SIGPHON (Jason Eisner)
Jason Eisner, jason@cs.jhu.edu
SIGPHON Report 2003
MEMBERSHIP
SIGPHON is ACL's special interest group for computational phonology. Membership currently stands at 237 (previous years: 214, 190, 176) with 53 declaring computational phonology as their "primary" interest (previous years: 47, 39, 36).
WORKSHOP
SIGPHON generally runs biannual workshops. 2003 is an off-year; thus we expect to hold a workshop in 2004. The workshop at ACL-02, chaired by Mike Maxwell, had the theme "Morphological and Phonological Learning" and was cosponsored by SIGNLL.
ONLINE ACTIVITIES
SIGPHON continues to serve the community by maintaining a mailing list and online bibliographies at its web site,
The website is likely to migrate soon to sigphon.org.
FUTURE PLANS
SIGPHON is interested in continuing to increase connections with related communities, such as speech technology, "pure" phonology, and morphology. A computational morphologist and a speech technologist have been recruited to the executive committee, which already includes the linguist John Goldsmith.
There is no ACL SIG devoted to computational morphology. SIGPHON's most recent workshop covered morphological and phonological learning, and we may consider expanding our charter to cover morphology as well as phonology. This is arguably a natural interest group, as the two problems are somewhat intertwined, especially when dealing with corpus data. Consolidating the two communities would increase submissions and attendance at our workshops. We would be interested in the thoughts of the ACL Exec on this question.
We hope to arrange for a special issue of a journal such as _Phonology_, focusing on the contribution of computational phonology to phonology proper. We are also considering holding our next workshop at a linguistics conference such as the LSA.
Jason Eisner
SIGSEM (Patrick Blackburn; Harry Bunt)
Patrick Blackburn, patrick@aplog.org; Harry Bunt, Harry.Bunt@kub.nl
SIGSEM, Report 2003
The biggest SIGSEM event held since the report to the ACL in June 2002 was IWCS-5, the fifth International Workshop in Computational Semantics. As usual, this was hosted by Tilburg University, and took place from January 15-17, 2003 in Tilburg, the Netherlands. The workshop, organised by Harry Bunt (both PC and local arrangements chair), was well attended (there were about 70 participants) and it is clear that the event has become a regular part of the computational linguistics calendar.
In addition to the regular IWCS workshop, on Tuesday January 14, the day before the start of IWCS-5, there was the first meeting of the SIGSEM Working Group on the Representation of Multimodal Semantic Information. IWCS-5 participants were invited to attend this meeting. About 20 WG members participated and 15 IWCS-5 participants attended. It was agreed to formulate a proposal to ACL for the next WG meeting as a SIGSEM event, open to SIGSEM members, in conjunction with the ACL 2003 conference. More information about the Working Group is available on the SIGSEM website (www.sigsem.org).
SIGSEM elections were held at the start of 2003. When nominations for SIGSEM officers closed on 15 February 2003 the following nominations had been received by Kristina Striegnitz, the returning officer for SIGSEM election.
Patrick Blackburn (President)
Harry Bunt (Secretary)
Johan Bos (Information Officer)
Ivana Kruijff-Korbayova (Information Officer)
Paul Piwek (Information Officer)
Matthew Stone (Information Officer)
As the SIGSEM constitution calls for the election of one president, one secretary, and four information officers, the above nominees were declared elected without contest.
Currently SIGSEM is organising ICoS-4, the fourth International Workshop on Inference in Computational Semantics. This will be held in Nancy, France, on 25-26 September 2003. ICoS-4 will be combined with a meeting of SIGSEM Working Group on Multimodal Meaning Representation. All IWCS-5 participants were invited to attend this meeting. More information about these meetings will soon be available on the SIGSEM website (www.sigsem.org).
SIGSEM is also endorsing a one-day school on Description Logics which will be held in Nancy on 24 September 2003. The school will introduce Description Logics from scratch and examine applications in NLP and text analysis.
Patrick Blackburn, Harry Bunt, June 2003.
REPORT by Dragomir Radev
INTRODUCTION
The ACL Web site is getting significantly more popular. From June 26, 2002 to June 20, 2003, the top-level page has been accessed 286,296 times. That is an average of 795 hits (accesses) per day (or an increase of 6.3% over the previous year).
NEW ACTIVITIES
Online secure registration for HLT-NAACL was organized by Sandee Carberry - this experiment proved very successful.
CONTINUING FEATURES
The ACL NLP/CL Universe, the catalog of Web-based NLP/CL resources, has grown at a steady rate. Since last year there has been a 15.3% increase in the number of pointers, bringing their total from 2509 to 2893. There are now 352 personal pages, 772 pointers to various resources (corpora, software, etc.), 391 academic or industrial labs, departments, and institutes, 76 professional organizations, etc.
This year, Erin Doumpoulaki of the University of Michigan has helped annotate all new entries.
The mailing list acl-news is used to announce new additions to the NLP/CL Universe. Currently, the list includes 651 subscribers, an increase of 10.9% since last year.
The unofficial Natural Language Processing FAQ (list of Frequently Asked Questions and Answers) is still available through the ACL page. Volunteers are sought to contribute to the list.
A facility allowing for the creation of mirrors of the ACL anthology was put in place this year.
ANNOUNCEMENTS
ACL members should make an attempt to announce all events, resources, or personal and lab pages to the aclweb.org site using the on-line URL submission mechanism.
Please feel free to send me mail with suggestions or comments
(radev@umich.edu)
===============================================================
John Nerbonne, Johanna Moore, Annie Zaenen
The ad-hoc committee discussed various experiences informally. We got input from Jan Hajic, Mari Hearst, Dekang Lin, Drago Radev, and Dan Roth. We recommend that the ACL acquire a license for the START package and (i) recommend it to conference program chairs (with Dan Roth's additions); and (ii) keep asking for feedback from conference chairs in the near future.
For a justification of this recommendation, see SandeeÕs ACL secretary report.
We include below some remarks by Jan Hajic leading to the same conclusion
Jan suggested that the following would be very useful.
- automatic summary to chairs of # submissions, status of referee reports (+/- rec'd)
- checking for all information necessary in submissions, incl. ID, length
- "rough first ranking of papers" incl those with only very positive reviews, only very negative reviews, and those with some very high reviews (unclear cases) Jan was clear about NOT advocating automation of decisions, however, just a guide for the chairs.
- reminders to referees about impending and past due dates
- ability to show reports to other referees (at least in some cases)
- web forms for reviews (useful for regimentation, even if sometimes inconvenient).
===============================================================
Robert Dale and Mary D Taffet
12 June 2003
Following consultation and discussion with a steering committee consisting of Chris Brew, Lillian Lee, and Dragomir Radev, the online ACL Course Survey ( http://www.clt.mq.edu.au/survey/ ) was created in mid-2002 as a spin-off from the 2002 ACL Workshop on Effective Tools and Methodologies for Teaching Natural Language Processing and Computational Linguistics. Site programming was carried out by Mary Taffet.
At the time, the survey was created with three aims in mind:
- to serve as a source of information, ideas and possibly material for
teachers of NLP;
- to serve as a source of information for students who might want to use this information in choosing an institution at which to study; and
- to serve as a set of data from which statistics can be derived, thus providing a snapshot of the character of the field that can be used to map trends.
At the end of its first year, the site lists 176 courses from 114 institutions in 49 countries. We know there are a lot more courses out there, and that the information currently on the site will rapidly become out of date, so periodic reminders of the site's existence sent to the ACL membership would be appropriate.
The site holds what we think is a wealth of information on how computational linguistics and natural language processing are being taught in universities and colleges. We are aware of a number of things that could be done to enhance the site and the usefulness of the information it contains, and we hope to make some minor changes in the near future.
We are interested in exploring how the information on the site can be put to use, and we welcome any proposals to use the data or other suggestions along these lines.
===============================================================
OLAC - Open Language Archives Community Ð http: www.language-archives.org
Steven Bird and Gary Simons
OLAC, the Open Language Archives Community, is an international partnership of institutions and individuals who are creating a worldwide virtual library of language resources by: (i) developing consensus on best current practice for the digital archiving of language resources, and (ii) developing a network of interoperating repositories and services for housing and accessing such resources.
Following the pilot phase in 2001 with rapid development of archiving infrastructure, the standards were frozen for a 12 month period of experimentation in 2002. At the end of this period the standards and infrastructure was revised once more, and the OLAC operational phase was launched in 2003. OLAC now invites language resource providers of all kinds to participate by contributing resource descriptions in a standard XML format. These descriptions are harvested, and cross-archive search is available on the OLAC and LINGUIST sites.
OLAC repositories of particular interest to ACL members include:
- ACL Anthology
- European Language Resources Association (ELRA)
- Natural Language Software Registry
- Linguistic Data Consortium (LDC)
For more information, please visit www.language-archives.org and join the OLAC-General mailing list.
No report on ACL Archives: nothing has changed.
Julia Hirschberg, Editor.
65 papers were submitted to Computational Linguistics in 2002. The mean time to first decision was 127 days, as compared to 110 days for 2001. Of the total submissions, there were 11 in the short paper category with a mean time to first decision of 96 days. The overall acceptance rate for 2002 was 35%. There are 17 papers submitted so far in 2003.
Disposition of manuscripts by type of first decision:
2002 2001
2000 1999
1998 1997
1996
Submitted 65 57
64 47
48 69
57
Accepted 23
18 15
13 9 21
9
Rejected 20
12 11
9 7 15
15
Resubmit 18 22
27 12
4 27
26
No decision yet
0 0
7 8 28
5 6
Withdrawn 2
3 3 2
0 1 1
Resub as squib 2 2 1
3 0 0
2
Of all submissions, 16 were for a special issue on the Web as Corpus, edited by Adam Kilgarriff and Greg Grefenstette.
For 2001 papers, the completed disposition of mss was, with the acceptance rate 42.1%:
Submitted 57
Accepted 24
Rejected 14
Resubmission 15
Withdrawn 3
Resub as Squib 1
Distribution of submissions by location of first author for 2002:
ASIA/PACIFIC
| P.R.China |
3 |
| Japan |
4 |
| Hong
Kong |
3 |
| Total |
10 |
EUROPE
|
Germany |
4 |
|
Belgium |
1 |
|
Netherlands |
2 |
|
Spain |
7 |
|
Finland |
2 |
|
Turkey |
1 |
|
UK |
12 |
|
India |
1 |
|
France |
2 |
|
Israel |
1 |
|
Ireland |
1 |
|
Slovenia |
1 |
|
Italy |
2 |
|
Greece |
1 |
|
Sweden |
1 |
| Total | 39 |
NORTH AMERICA
| USA
|
13 |
| Canada |
3 |
| Total |
16 |
USA 13
Canada 3
____________________________
Total 16
GRAND TOTAL 65
The new 2003-5 editorial board is:
Thorsten Brants, Google Inc
Rebecca Hwa, University of Maryland
Emiel Krahmer, Tilburg University
Gina-Anne Levow, University of Chicago
Ted Pedersen, University of Minnesota, Duluth
Gerald Penn, University of Toronto
Brian Roark, AT&T Labs -- Research
Amanda Stent, Stony Brook University
Antal van den Bosch, Tilburg University
The new Editor is
Robert Dale, Macquarie University, Australia
===============================================================
Pierre Isabelle, Squibs editor
At the beginning of year 2002, there was only one submission in the pipeline. In the course of year 2002, 4 additional papers were (re-)submitted. At the end of the year, there was again only one submission left in the pipeline. Thus, 4 decisions were made during year 2002. The results were as follows:
* 1 paper accepted
* 3 invitations to revise and resubmit
The mean time taken for these decisions was 135 days.
Thus far, only 2 papers have been (re-) submitted in 2003.
Graeme Hirst
June 2003
BOOK REVIEWS PUBLISHED
In 2002, we continued to publish an average of about five book reviews plus a couple of brief notices in each issue of Computational Linguistics.
This number will drop in 2003, as the number of books published in computational linguistics and closely related areas has dropped noticeably over the last year or two.
Most reviews are published in a timely manner -- that is, within 12 months of receipt of the book. This allows six months for the reviewer (some take less) and five months for journal production.
MATERIAL REVIEWED
I am continuing to be fairly strict in deciding if a book is to be reviewed, but try to include all books that are in "core" computational linguistics, as well as a variety of books from adjacent and overlapping disciplines that are likely to be useful in CL. We do not review technical reports, doctoral theses, conference proceedings, or workshop proceedings, except if revised for publication as a book by a recognized publisher.
PRODUCTION MATTERS
I am indebted to Nadia Talent for long hours of reading out loud withme to check the galleys.
Steven Bird
The ACL Anthology is a digital archive of research papers in computational linguistics, sponsored by the CL community, and freely available to all. 44,000 pages have been scanned and OCRed, corresponding to 200 volumes, 6,400 papers, or a 13ft/4m shelf. It includes the CL journal, proceedings of the ACL, EACL, NAACL, ANLP, TINLAP, COLING, and ACL-sponsored workshops. All materials have been OCRed, permitting full-text search. Each publication series in the anthology has an introduction explaining its history and purpose.
PERSISTENT URLs: The ACL website supports persistent URLs for all papers that are resolved to a copy at the selected mirror site. These URLs have the form http://www.aclweb.org/anthology/P99-1012, and they may be used for citation purposes. Individuals and special interest groups can easily construct annotated bibliographies of ACL publications with pointers to the full text. (Thanks to Dragomir Radev and Haejoong Lee for setting this up.)
CITESEER: The anthology index pages have been submitted to CiteSeer (http://citeseer.nj.nec.com/cs), and about 1000 of the more recent papers have now been indexed.
REMAINING MATERIALS. Recent additions to the anthology include the ACL-01 workshops and the EACL-03 conference. HLT-NAACL-03 and ACL-03 will be ready very soon. The following materials have not yet been incorporated into the anthology: NAACL-01 workshops, ACL-02 workshops, EACL-03 workshops, COLING-02 proceedings and workshops, CL journal prior to 1980 (microfiche), Mechanical Translation and Computational Linguistics journal (some issues not yet obtained). The following materials are also missing from the anthology and I am trying to obtain the physical or digital materials.
* W9801 EMNLP
* W9804 Computational Terminology
* W0105 EMNLP
* W0106 SENSEVAL Workshop
* W0111 Arabic Language Processing: Status and Prospects
* W01?? Collocation: Computation Extraction, Analysis and Exploitation
SPONSORSHIP. In the last six months some outstanding pledges have been paid up. Priscilla Rasmussen will report on total sponsorship income.
BIBLIOGRAPHY. Doug Arnold is compiling BibTeX data for the anthology.
FUTURE MATERIALS. David Yarowsky has developed a process by which new anthology content will be generated as part of the process of compiling the conference CD-ROM. Julia Hirschberg has arranged for electronic content from the Computational Linguistics journal to be supplied directly to the ACL by MIT Press, one year from the original publication date.
OUTSTANDING TASKS:
* Finalize content (obtaining materials, scanning, uploading)
* Validation, pursuing technical problems with the scanning company
* Finalize bibliography
* Improve the design of the anthology webpage
* Finalize mirror sites
* Construct author index
* Make corrections as they are reported by users
* Publish pagescans and OCR text
ONGOING MAINTENANCE: Maintaining the anthology is becoming more streamlined, but will continue to involve a myriad of small tasks. To support this ongoing effort, I propose that each major ACL conference make a donation to the ACL Anthology Fund. Suggested level: $500 per conference (i.e. about $1 per participant).
Bran Boguraev and Steven Bird
The goal of Studies in Natural Language Processing is to identify and publish state-of-the-art work on topics of interest to the CL/NLP community. Feedback from sales suggests that interest in the community is maintained at level similar to prior years; levels of sales are in line with the types of scholarly books that have been published recently.
The series has a new acquisitions editor at Cambridge University Press (Andrew Winnard), based in the Cambridge office. The series is undergoing a renewal and reprofiling process with the Press, and we solicit proposals for books addressing current research topics in CL/NLP. Breadth of scope and audience is of particular importance.
STATUS OF CURRENT PROJECTS
1. Books published during the last 12 months:
0521650585 HB ASHER/LASC Logics of Conversation SNLP 7/1/03
2. Book projects approved, with manuscripts under review:
* Wilks,Y., "Machine Translation; Its Scope and Limits"
3. Forthcoming titles, without projected publication dates:
* Basili,R. et al., "Lexical Acquisition for Practical NLP Systems"
* Carroll,J., "Practical Natural Language Parsing"
* Daelemans/Bosch "Memory-Based Language Processing"
* Pollard,P. "Generalized Phrase Structure Grammars"
* Wilks,Y. "Language, Cohesion and Form
Selected Papers of Margaret Masterman"
===============================================================