WMT2020 will be run as a virtual conference. This is a new experience for most of us, but we hope it will still be a productive and enjoyable two days. Below we briefly describe how each of the sessions will work. For more detail, refer to the EMNLP blog on the virtual conference.
The keynote this year will be delivered by Masakhane. It will be pre-recorded, and live-streamed, followed by Q&A via Zoom. It will be immediately followed by a panel discussion. The keynote will begin at 1300 UTC.
All papers have pre-recorded talks. The overview and research papers have 12 minute talks, and the system papers have 2 minute talks. These are all available from the virtual EMNLP site, and eventually will be in the ACL anthology.
Our rocket chat channel is #workshop-wmt
These will be presented in 4 1-hour-long Zoom sessions, organised similarly to the EMNLP Q&A sessions, i.e. as panel discussions. There will be a chair, who will invite authors to summarise their work, then the audience can ask questions. Questions can be asked in Zoom or in RocketChat.
These will be similar to the research paper sessions, with a chair overseeing a Q&A session. There will be two Q&A sessions on each day, with each task presented twice (subject to organiser availability). The reason for having the two sessions is to cover as many time zones as possible -- the first will be morning (UTC) and the second will be afternoon (UTC).
On each day there will be two Gather.Town poster sessions, with each task assigned to a specific day. If you uploaded a poster, then it will be displayed in Gather.Town. If you did not upload a poster, you can still screen-share a poster to everyone on your "virtual rug".
Note that there may be a change to the poster sessions enforced by EMNLP because of space restraints. Waitch for announcements.
A video presenting Gather.Town can be found here