2020Q3 Minutes
Reviewing -- feedback from the community and strategy
Revising the rolling review proposal based on the feedback (Hinrich)
Finances -- ACL's financial status given the COVID-19 crisis, and necessary adjustments
David put forward a proposal for limited-hybrid *ACL model, which is under discussion on the Exec via email.
Diversity / inclusiveness -- situation, issues, and plans
Time sensitive proposals:
- Founding for D&I initiatives permanently integrated
- ACL-X sub-events planning
- D&I training for ACL
Plan in 2020-2021:
- Status (2020 Q3-Q4)
- Measure development (2020 Q4 - 2021 Q1)
- Implementation (2021 Q2 onwards)
- Evaluation (biannual from 2021 Q3)
Emerging issues:
- Increased gender disparity in submission
- Ethics protocol in publication
Further discussion will be made via email.
Proposals for year-long Mentorship & ACL Speakers
- Year-long mentorship:
- Provide mentorship opportunities for new NLP researchers throughout the year
- ACL speakers:
- Facilitate interaction with well-known NLP researchers for a broader audience interested in NLP
The above initiatives will also be discussed via email after the Exec meeting.
Environmental sustainability & Ethics
- Environment sustainability:
- Tim will form a working group to put together a document that identifies different sources of carbon and impact on environment.
- Ethics:
- Suggestion: ACL has some basic ethics requirements that recipients of ACL awards (LTA, DSA, fellows...) should obey by. If the recipients breach the requirements, the ACL reserves the right to deprive them of the honors.
- It is difficult to define "gross violation" and "non-gross violation".
Management office & Appoint a Sponsorship Director
The Exec approved of the candidate of the sponsorship director, which is an ad hoc leadership on exec-plus. Priscilla will talk to the candidate to verify his willingness to take this position and about the future plans.
ACL2022 updates
The venue of ACL2022 will be announced at EMNLP2020.
ACL2020 -- the first virtual conference
Main points can be found in Dan's Q3 report: https://www.aclweb.org/adminwiki/index.php?title=2020Q3_Reports:_General_Chair
Some other questions:
- Did the organizing committee compare SlidesLive with other options? (Priscilla)
- Dan: There are quite a few problems with SlidesLive. But have not compared it with other similar solutions, since it was decided on before determining to be a full-virtual conference. Whether we choose SlidesLive or a competitor, I advice that future virtual conferences should use professional people and solutions.
- Is the review quality a big problem? (Hinrich)
- Natalie: The review quality is overall good. Only 2-3 complaints about this. Two practices that help improve the review quality are: (1) automatic paper matching, (2) author rebuttal.
- Most people are in Q&A sessions. Fewer people explored the exhibition and booths (Priscilla)
- Natalie: In the virtual conference, many people precisely chose one or two sessions to attend one day. Dan: probably we should indeed find a way to increase sponsor booth attendance, although a lot of people did show up; the Diamond sponsors, for example, got 550 unique page views each, and had 115 users average each on their rocket channels.
- Zoom app has to be downloaded, which cannot be visited via browser (Priscilla)
- Dan: Zoom can be visited via Chrome and Firefox browsers. So it should have been possible for people to avoid downloading zoom (which is important because some companies like Google don't allow downloading the Zoom app). I heard no complaints about this during the conference, but it's possible that people from some regions were not able to use either Zoom or the Chrome or FIrefox browsers. This may or may not be related to the issue of some countries needing VPNs.
- Social chair is important (Hinrich)
- Natalie, Dan, Priscilla: In the virtual conference model, we need a social chair to work together with the other organizers, e.g., the exhibition chair, sponsorship chair, D&I chair, to improve the social activities and experience.
- What is the outcome of the new reviewer training experiment? (Hinrich)
- Natalie: Generally positive experience, but not scalable now. New reviewers did a good job, while many bad reviews are from senior people.
PCC update
(1) Pain point in ACL2020: PCC was involved at the last minute.
In the future, PCC should be involved earlier in the decision of conference format. ACL2021 has already taken this into consideration when making the conference organizing timeline. Barbara will update the conference handbook and add the checking point for the future conferences.
Besides, it should be built into the software of the virtual conference the harassment alert function. (Priscilla)
(2) Ongoing work: to finalize the GDPR compliance to cover the CoIs data and PCC case information.
(3) Softconf doesn't serve TACL. TACL doesn't implement CoI policy in a robust way. PCC chairs are contacting the Softconf people to support TACL's CoI detection. PCC will coordinate with Amanda Stent.