2012Q3 Reports: NAACL 2012

From Admin Wiki
Jump to: navigation, search

NAACL 2012 General Chair Report

by Jennifer Chu-Carroll

NAACL HLT 2012 took place June 3-8, 2012 in Montreal. The conference completed successfully, attracting a healthy number of attendees. The organizing committee was excellent and put together interesting programs and I am very grateful for the effort they put into this conference.

While NAACL HLT 2012 followed the format of previous conferences for the most part, there are a number of differences worth pointing out. The program co-chairs pioneered the first ever "NLP Idol: Plucked from Obscurity", which was fun and informative, and very well-received. Also the first time we had Priscilla as the Local Arrangements Chair, assisted by a local advisory team. This worked out very well as long as future organizers are aware of the need to fill a couple of roles that are typically covered by the Local Arrangements Committee. I will discuss these in the rest of this report, which consists of my experience as well as feedback from the organizing committee.

In general, questions were raised about the 'HLT" component of the conference. Perhaps more discussions can be paid to its role in the conference in the future? I have much longer versions of reports from many organizers, which I will not include in full here, but will forward on to the NAACL board and next year's general chair for their reference.

Main Conference

The main conference received 196 full paper submissions of which 61 papers were accepted (31% acceptance rate), as well as 105 short paper submissions of which 36 were accepted (34% acceptance). The number of submissions received was lower than in past years, and, based on postmorten analysis, is likely due to the fact that the submission deadline was 9 days before the ACL submission deadline and that no double submission was permitted.

This year we also inargurated "NLP Idol", a "game show" in which four contestants each tried to convince the audience that a past paper deserves a revisit in today's NLP world. The contestants and judges all did a great job, and it was both informative and entertaining for all involved. Overall, this new session was very well received.

The PC co-chairs felt that it was very hepful to have one PC chair who had experience as publications chair in the past. Eric Fosler-Lussier as PC co-chair and past pub chair, played a very active role in the publication process which made everyone's life easier.

Local Arrangements

Local arrangements worked out very well. One lesson learned is to leave sufficient time to ship equipment to Canada and stay on top of the shipping company like a hawk. We ended up receiving our shipment on the last day of the conference, although things ran so smoothly without them that most people didn't even notice it.

Two roles typically filled by the local arrangements committee that caught me a bit by surprise is webmaster and conference handbook production. The former I realized fairly early on, and was able to enlist a summer student to take on the role (this role may be easier for someone working at a university to fill than it was for me). The latter we didn't realize until just over one month before the conference and we ended up soliciting student volunteers to produce the handbooks. This wokred out very well, though future general chairs should take note of that and make sure they recruit someone sooner than we did (it was stressful for a while there).


While the NAACL publications process is pretty well integrated with the START system, a few steps in the end were not fully supported and not well documented, which resulted in some hickups in the process. The publications co-chairs suggest that the code they developed to address the problems be integrated into START to streamline the process for future years. They also suggest better documentation for the process as well as have an experienced publications chair play an advisory role in a future year. Nizar Habash is willing to volunteer to carry out some of these tasks.


One of the tutorials had low registration and it was discussed at one point whether it should be canceled or not. We ultimated decided that it should go on and the conference should absorb any financial loss (I believe in the end there wasn't a loss), out of fairness to the tutorial presenters. However, during that discussion, it was mentioned that the 6 tutorials is generally preferred over 8. It would be good to make this recommendation available somewhere for future tutorials chairs.

Furthermore, there was some difficulty in getting tutorial slides from the presenters weeks in advance (in time for printing and for USB). The question was raised about whether it is worth the trouble to do so.


It may help to have a unified menu and submenu structure across conferences, instead of having each webmaster go through the trial-and-error process. This will also ensure that important information doesn't get left off of the website. Dirk Hovy is happy to provide code and experience for future webmasters.