Admin Wiki - User contributions [en]

2015Q3 Reports: EMNLP 2014

2015-07-03T08:43:30Z

AlessandroMoschitti:

The EMNLP conference series is annually organized by SIGDAT, the Association for Computational Linguistics special interest group on linguistic data and corpus-based approaches to NLP.

This edition of the conference was held from October 25, 2014 (Sat.) to October 29, 2014 (Wed.) in Doha, Qatar.

In the past five years, the EMNLP conference attendance has been continuously growing, reaching just over 500 paying attendees in 2013, and it is nowadays considered as one of the leading conferences in Computational Linguistics and Natural Language Processing.

Given the growing trend, we believe it was the right time to led EMNLP 2014 into an organization structure typical of large and important conferences.
Therefore, we proposed several novelties:

first of all, a large organization committee consisting of twenty (plus twenty-six area chairs) well-known members of the ACL community, who carried out several tasks required by the new achieved scale.

Secondly, as this is the first conference edition spanning five days, in addition to six workshops, we also selected and included for the first time an excellent selection of eight tutorials. We defined a registration policy that allows the participants to attend any of the tutorials and workshops (held on October 25th and 29th) by just paying a low flat rate on top of the registration fee for the main conference. This greatly increased the participation in tutorials and workshops.

Thirdly, as a standalone conference, EMNLP required the definition of new administrative procedures and policies, regarding sponsorship booklets, double submission, scholarship assignment, and the joint EACL-ACL-EMNLP call for workshop proposals.

Fourthly, we experimented with a new approach to overcome the increasing number of papers to be presented at the conference.
Our new approach consisted in presenting posters in nine sessions each proposing a small numbers of papers: this way poster presentations received more space and consideration. Most people liked this new initiative.
However there were a couple of presenters of the last session of the last day, who felt penalized by the lack of audience. Suggesting that such typically difficult slots may not be the best fit for our new presentation format.

Next, we recorded tutorials, invited speakers and all the oral presentations of the technical papers. These videos are available at the official website of the conference http://emnlp2014.org/program.html

Finally, this was the first time that an ACL conference was largely supported by a government research foundation. The Qatar National Research Fund (QNRF) included EMNLP 2014 as one of its local funding events. This enabled EMNLP and SIGDAT to perform unprecedented student scholarship support: more than 33 students were sponsored (partially or entirely) for participating to the conference.
The obtained funds also allowed for offering a social dinner free of charge to all the attendees and closing the conference budget in active, i.e., $46,100. This has provided SIGDAT with important additional resources for sponsoring students for the next editions of EMNLP.

2015Q3 Reports: EMNLP 2014

2015-07-01T18:26:53Z

AlessandroMoschitti:

The EMNLP conference series is annually organized by SIGDAT, the Association for Computational Linguistics special interest group on linguistic data and corpus-based approaches to NLP.

This edition of the conference was held from October 25, 2014 (Sat.) to October 29, 2014 (Wed.) in Doha, Qatar.

In the past five years, the EMNLP conference attendance has been continuously growing, reaching just over 500 paying attendees in 2013, and it is nowadays considered as one of the leading conferences in Computational Linguistics and Natural Language Processing.

Given the growing trend, we believe it was the right time to led EMNLP 2014 into an organization structure typical of large and important conferences.
Therefore, we proposed several novelties:

first of all, a large organization committee consisting of twenty (plus twenty-six area chairs) well-known members of the ACL community, who carried out several tasks required by the new achieved scale.

Secondly, as this is the first conference edition spanning five days, in addition to six workshops, we also selected and included for the first time an excellent selection of eight tutorials. We defined a registration policy that allows the participants to attend any of the tutorials and workshops (held on October 25th and 29th) by just paying a low flat rate on top of the registration fee for the main conference. This greatly increased the participation in tutorials and workshops.

Thirdly, as a standalone conference, EMNLP required the definition of new administrative procedures and policies, regarding sponsorship booklets, double submission, scholarship assignment, and the joint EACL-ACL-EMNLP call for workshop proposals.

Fourthly, we experimented with a new approach to overcome the increasing number of papers to be presented at the conference.
Our new approach consisted in presenting posters in nine sessions each proposing a small numbers of papers: this way poster presentations received more space and consideration. Most people liked this new initiative.
However there were a couple of presenters of the last session of the last day, who felt penalized by the lack of audience. Suggesting that such typically difficult slots may not be the best fit for our new presentation format.

Next, we recoded tutorials, invited speakers and all the oral presentations of the technical papers. These videos are available at the official website of the conference http://emnlp2014.org/program.html

Finally, this was the first time that an ACL conference was largely supported by a government research foundation. The Qatar National Research Fund (QNRF) included EMNLP 2014 as one of its local funding events. This enabled EMNLP and SIGDAT to perform unprecedented student scholarship support: more than 33 students were sponsored (partially or entirely) for participating to the conference.
The obtained funds also allowed for offering a social dinner free of charge to all the attendees and closing the conference budget in active, i.e., $46,100. This has provided SIGDAT with important additional resources for sponsoring students for the next editions of EMNLP.

2015Q3 Reports: EMNLP 2014

2015-07-01T17:54:11Z

AlessandroMoschitti:

The EMNLP conference series is annually organized by SIGDAT, the Association for Computational Linguistics special interest group on linguistic data and corpus-based approaches to NLP.

This edition of the conference was held from October 25, 2014 (Sat.) to October 29, 2014 (Wed.) in Doha, Qatar.

In the past five years, the EMNLP conference attendance has been continuously growing, reaching just over 500 paying attendees in 2013, and it is nowadays considered as one of the leading conferences in Computational Linguistics and Natural Language Processing.

Given the growing trend, we believe it was the right time to led EMNLP 2014 into an organization structure typical of large and important conferences.
Therefore, we proposed several novelties:

first of all, a large organization committee consisting of twenty (plus twenty-six area chairs) well-known members of the ACL community, who carried out several tasks required by the new achieved scale.

Secondly, as this is the first conference edition spanning five days, in addition to six workshops, we also selected and included for the first time an excellent selection of eight tutorials. We defined a registration policy that allows the participants to attend any of the tutorials and workshops (held on October 25th and 29th) by just paying a low flat rate on top of the registration fee for the main conference. This greatly increased the participation in tutorials and workshops.

Thirdly, as a standalone conference, EMNLP required the definition of new administrative procedures and policies, regarding sponsorship booklets, double submission, scholarship assignment, and the joint EACL-ACL-EMNLP call for workshop proposals.

Fourthly, we experimented with a new approach to overcome the increasing number of papers to be presented at the conference.
Our new approach consisted in presenting posters in nine sessions each proposing a small numbers of papers: this way poster presentations received more space and consideration. Most people liked this new initiative.
However there were a couple of presenters of the last session of the last day, who felt penalized by the lack of audience. Suggesting that such typically difficult slots may not be the best fit for our new presentation format.

Next, we recoded tutorials, invited speakers and all the oral presentations of the technical papers. These videos are available at the official website of the conference http://emnlp2014.org/program.html

Finally, this was the first time that an ACL conference was largely supported by a government research foundation. The Qatar National Research Found (QNRF) included EMNLP 2014 as one of its local funding events. This enabled EMNLP and SIGDAT to perform unprecedented student scholarship support: more than 33 students were sponsored (partially or entirely) for participating to the conference.
The obtained funds also allowed for offering a social dinner free of charge to all the attendees and closing the conference budget in active, i.e., $46,100. This has provided SIGDAT with important additional resources for sponsoring students for the next editions of EMNLP.

2015Q3 Reports: EMNLP 2014

2015-06-30T22:21:28Z

AlessandroMoschitti: Created page with "The EMNLP conference series is annually organized by SIGDAT, the Association for Computational Linguistics special interest group on linguistic data and corpus-based approache..."

The EMNLP conference series is annually organized by SIGDAT, the Association for Computational Linguistics special interest group on linguistic data and corpus-based approaches to NLP.

This edition of the conference was held from October 25, 2014 (Sat.) to October 29, 2014 (Wed.) in Doha, Qatar.

In the past five years, the EMNLP conference attendance has been continuously growing, reaching just over 500 paying attendees in 2013, and it is nowadays considered as one of the leading conferences in Computational Linguistics and Natural Language Processing.

Given the growing trend, we believe it was the right time to led EMNLP 2014 into an organization structure typical of large and important conferences.
Therefore, we proposed several novelties:

first of all, a large organization committee consisting of twenty (plus twenty-six area chairs) well-known members of the ACL community, who carried out several tasks required by the new achieved scale.

Secondly, as this is the first conference edition spanning five days, in addition to six workshops, we also selected and included for the first time an excellent selection of eight tutorials. We defined a registration policy that allows the participants to attend any of the tutorials and workshops (held on October 25th and 29th) by just paying a low flat rate on top of the registration fee for the main conference. This greatly increased the participation in tutorials and workshops.

Thirdly, as a standalone conference, EMNLP required the definition of new administrative procedures and policies, regarding sponsorship booklets, double submission, scholarship assignment, and the joint EACL-ACL-EMNLP call for workshop proposals.

Fourtly, we experimented a new approach to overcome the increasing number of papers to be presented at the conference.
Our new approach consisted in presenting posters in nine sessions each proposing a small numbers of papers: this way poster presentations received more space and consideration. Most people liked this new initiative there were some of the presenter of the last session of the last day, who felt penalized by the lack of audience. Suggesting that such penalizing lost may not be the best for our new presentation format.

Next, we recoded tutorials, invited speakers and all presentations of the technical papers. These videos are available at the official website of the conference http://emnlp2014.org/program.html

Finally, this is the first time that an ACL conference is largely supported by a government research foundation. The Qatar National Research Foundation (QNRF) included EMNLP 2014 as one of its local funding events. This enabled EMNLP and SIGDAT to perform unprecedented student scholarship support: more than 33 students were sponsored (partially or entirely) for participating to the conference.
The obtained funds also allowed for offering a social dinner free of charge to all the attendees and closing the conference budget in active: $46,100 were given back to SIGDAT thus creating additional resources for sponsoring students.

2014Q3 Reports: EMNLP 2014

2014-06-09T05:53:24Z

AlessandroMoschitti: Created page with "The upcoming conference, Empirical Methods in Natural Language Processing (EMNLP) of SIGDAT, will take place October 25–29, 2014 in Doha, Qatar. It will span three days of ..."

The upcoming conference, Empirical Methods in Natural Language Processing (EMNLP) of SIGDAT, will take place October 25–29, 2014 in Doha, Qatar.
It will span three days of paper presentations plus two days of workshops and tutorials.

This year EMNLP proposes several novelties:

1. a large organization committee, typical of very large conferences such as ACL and NAACL.

2. A considerable number of student scholarships for attending the conference will be made available.

3. For the first time, EMNLP 2014 will have tutorials.

4. A large program of seven workshops

The submission deadline has just passed and we received 840 submissions.
There were 37 withdrawn papers until now, which modifies our counts to 525 long and 278 short papers.

Everything is proceeding smoothly with the organization.

This thread is of course preliminary and it will be updated at the end of the conference.

2014Q3 Reports: Sponsorship Committee

2014-06-05T17:51:01Z

AlessandroMoschitti: Created page with "The sponsorship activity is typically divided in two subtasks to make the entire process more efficient: (1) search for local sponsors, i.e., roughly, those who have location ..."

The sponsorship activity is typically divided in two subtasks to make the entire process more efficient: (1) search for local sponsors, i.e., roughly, those who have location and specific interests in the area of the conference and (2) global sponsors, who typically are all the others. The latter can be in turn subdivided in recurrent sponsors, i.e., those who have had a continuous relation with ACL and other sponsors. Given such subgroups, we assigned our team composed by local sponsorship co-chairs, Priscilla (ACL business manager), (E)ACL co-chairs, to the three types above, respectively.

The global sponsors for EACL, ACL, NAACL and Asia conferences constitute a similar topic for ACL. Thus, there were preliminary email exchanges between the sponsorship co-chairs of the regions above to define: a common sponsorship booklet, a common document listing potential sponsors (who were successfully contacted in the previous years along with new ones) and a common invitation letter. Although, this activity was beneficial for producing common sponsorship material is not completely suitable for EACL co-chairs as their sponsorship activities must start before, given its earlier schedule. The overall suggestion for the future co-chairs is to adopt a basic booklet version and a deadline for discussing it. If such deadline has passed, the EACL co-chairs can start to utilize the current version. One interesting outcome of the overall discussion is the availability of a booklet document in Google docs.

To gain efficiency, we created a specific spreadsheet for EACL, which keeps track of candidate organizations, people inside these organizations responsible for marketing or sponsorships. We populated the table with some organizations that have been (E)ACL sponsors in the past, e.g., search engine companies, NLP OEM vendors and solution providers and some academic institutions. In the future, we recommend reaching out beyond these (e.g., car companies nowadays have internal dialog systems groups). We used the table to track the status of the various sponsorship leads as well as to avoid multiple people approaching the same company. Since the global sponsor team for (E)ACL was composed by three members, we could efficiently subdivided the different sponsors among us. We also expect this table to act as a means to preserve knowledge about successful approaches and contact details for the benefit of our successors in subsequent years.

In general, it was not easy to obtain commitments for sponsorships. Nevertheless, we were able to have an increase of 40% over the largest money amount raised with global sponsors by any previous EACL conference.
A final remark regards the role of the two-packet sponsorship option (including ACL and EACL). This has benefit more ACL: our main explanation is that it can be more attractive for sponsors than EACL. However, it would be surely worthwhile to stress the double packet option more.