Virtual reflections on virtual conferences

I'm sure many scientists have already blogged about their experiences at virtual conferences. Certainly, I know a few of my labmates have! Nonetheless, as I've just had my first experience with a virtual conference, I thought I'd add to that dataset. (We always need more data, right?)

I am currently attending the Fall Meeting of the American Geophysical Union. This is a massive conference bringing researchers from across the Earth and planetary sciences, including geology, weather, climate, hydrology, the environmental sciences, and more. This year, they have 27,942 attendees, which I think is about on par with previous years that required in-person attendance, and nearly all of them are presenting (25,629 presentations). As you can imagine, this conference was always a bit overwhelming in person! 

Mostly Pro: Hybrid Poster Session

The poster session at which I presented started with a 3-minute live talk by each presenter. There was then a live Q&A session for about 15 minutes, followed by a 1-hour virtual poster session using Gather.town. Personally, I think this was an excellent format, and at least in some ways, an improvement on traditional poster sessions. The opportunity to talk about my science to everyone in the session for 3 minutes is certainly more exposure than I've usually received at poster sessions. Although that "lightning talk" format has been used at in-person conferences before, I think it could be much more utilized in a virtual format, where there are unlimited presentation "rooms" and the schedule can span much more time. Indeed, this year the Fall Meeting was spread out across 17 days rather than its usual 5 days, and it utilized a much larger fraction of each day than usual. Today, for example, I see presentations spanning 4 AM to 12:30 AM in my time zone.

It also somehow more comfortable to me to present when everyone was virtual, as though everyone be contained to a little box on a screen made us somehow more equal, though it's hard for me to separate that from the fact that I knew both chairs and a few of the other presenters, which has not commonly been the case for me in past oral presentations.

One possible negative is that no one visited my poster during the Gather.town portion, which simulates a physical space-in this case, a poster session-in an 8-bit, 2D world. Each attendee has an avatar. When your avatar walks near another avatar, their video and audio fade into your screen (and vice versa for them), so that it's much like walking up to a conversation circle in physical space. Each poster had a dedicated area, but no one visited mine. However, I think this is probably, on balance, only a minor negative, because I still reached more people with my science than usual via the lightning talk, and I saw very few posters being visited. For the most part, attendees clumped into one or two talking circles, which one could argue may be more productive than one-on-one talks at individual posters anyway. Unfortunately, the theme of the session wasn't an ideal fit for my presentation: The Use of Analog Environments in Preparing for the Exploration and Characterization of Planetary Surfaces II eLightning. Whereas my science was focused on refining remote techniques by using terrestrial analog sites as test examples, most of the other presentations seemed more focused on simulating or preparing for human exploration of other planets. My focus on remote-sensing application may have made me a bit of an odd duck.

Mostly Pro: Jumping between Sessions

As a giant conference, the in-person Fall Meetings were spread out between massive buildings. Getting from one session to another took time, and precisely because there are so many presentations on so many topics, it's always the case, at least in my experience, that I have to forego some talk or even session of interest to attend another one. In a virtual format, with posted schedules, I can easily jump from a talk in one session to a talk in another, with just a few clicks.

Pro: Tutorials

As a technically-minded scientist, I've been curious about machine learning for some time. When I saw a 6-hour tutorial offered at the Fall Meeting, I jumped at the opportunity. I must say, it was an excellent experience. Certainly, there was a lot that was above my head, but there was also a fair bit that I already knew, too, which is probably unavoidable when trying to address the needs and interests of a group with diverse degrees of experience in a subject area. Each of the presenters and panelists were excellent. They clearly were very knowledgeable, generally had a knack for explaining concepts well (or perhaps just a lot of practice), and clearly were passionate for what they do. The tutorial was also well planned, with a mix of hands-on activities (with some check-your-understanding polls) and presentations by different people from different perspectives.

I think the virtual format was massively advantageous. I have to imagine that interest in machine learning was also very strong one year ago, but they had to cut off registration at 100 people due to room size then. This year, they had over 1200 registrants! That would have been impossible to accommodate in person. However, their peak simultaneous online attendance was only around 700, indicating that many registrants prioritized attending science presentations or other responsibilities during the time window, or may have been sleeping. The flexibility to jump in and out of the tutorial to suit your interests or level of prior acquaintance, or to accommodate other demands on your time, is certainly much facilitated by the virtual format. Nonetheless, I would guess that many of the registrants simply registered for the tutorial to gain access to it on their own time, as everything will be available online later. Even so, the virtual format helped in this case by offering that opportunity; last year, even someone opting not to attend would still have taken one of the limited 100 allowed registrations. And lest you think that the 1200 registrants vs. 700 online discrepancy could be largely explained by people who simply lost interest or forgot, there was a $50 cost to the tutorial, so I doubt that was a major factor.

The $50 registration fee also means that, with 1200 registrants, AGU brought in $60,000 from that tutorial! Hopefully, that revenue will get some attention, and perhaps AGU will find ways to take the best of virtual tutorials into future in-person Fall Meetings.

Pro and Con: Time Management

The flexibility of a virtual conference spread over 2.5 weeks definitely avoids the usual necessity to block out an entire week of one's schedule for a conference. However, I also found that it makes it easier to simply skip talks that you might have attended otherwise. It's also more tempting to try to multi-talk, and work while attending a talk, which can be advantageous to productivity but is also disadvantageous to engagement, both for you and the presenter.

Con: Networking

Undoubtedly, one of the most important things that happens at conferences is the ability to chat one-on-one with a potential collaborator, grad school/postdoc advisor, or employer. Although Gather.town somewhat presents that possibility, there really isn't any substitute for an in-person conversation, I don't think, and that is definitely a major negative for virtual conferences.