2010Q3 Reports: Workshop Chairs

The six workshop co-chairs responsible for NAACL HLT 2010, ACL 2010 and Coling 2010 jointly drafted a call for workshop proposals which was sent out on July 31st. By the deadline of October 30th we had received 56 submissions of which 29 specified ACL as their first choice, 13 gave HAACL HLT as their first choice, 12 gave Coling as first choice, and 2 did not specify any preference. Submissions were made through START. The six workshop co-chairs gave feedback on proposals, and then, for each of the three conferences, the co-chairs for that conference took responsibility for the selection of workshops for their conference.

The following workshops were accepted for ACL-2010 (the first two were "pre-approved"): SemEval-2010 (2 days); WMT’10/MetricsMATR (2 days); The LAW IV (1.5 days); BioNLP2010 (1 day); CMCL (1 day); NLPLing 2010 (1 day); SIGMORPHON-11 (1 day); TextGraphs-5 (1 day); NEWS 2010 (1 day); ATANLP 2010 (1 day); DANLP 2010 (1 day); CDS (1 day); GEMS-2010 (1 day).

Given the way that the selection process was carried out (in particular the fact that the co-chairs for a given conference made their initial selection from those workshops that specified their conference first), and the disproportionate number of proposals that specified ACL as the preferred location, it turned out that the workshop location preference information had a significant impact on the chances of selection. This issue could be addressed in two ways: by giving a clear indication in the call for proposals that specifying a preferred location was optional, and would be taken to indicate a strong preference; or by having all co-chairs jointly produce a (partial) ordering of ALL proposals (without consideration of preferred location), and then allocate workshops to conferences in a way that maximised the rankings of the accepted workshops.

Given the substantial number of strong proposals that specified ACL as first choice, a second issue arose having to do with the number of workshops that could be accepted. Two contrasting points of view were expressed: accept as many workshops as we could find space for; or limit the number of workshops in order to improve attendance rates and avoid the need to spatially distribute the workshops. In the end, we opted to use all available space, and it appears that (based on pre-registration figures) attendance at all of the accepted workshops is strong.

David Weir and Pushpak Bhattacharyya