Alan Gatherer, Angel Lozano, Jeff Andrews, Ana Garcia Armada, Muhammad Zeeshan Shakir, Yingzhen Qu, RR Venkatesha Prasad, Anthony Soong
Published: 9 Jun 2020
CTN Issue: June 2020
A note from the editor:
It is the best of times and the worst of times for conference organizers. For a short while it looked like conferences would be cancelled this year. Instead they have come roaring back, online. Attendance at large conferences has more than doubled, except that no one is physically coming to them and in many cases most are not paying to attend. For this month’s CTN news feed our brave authors spent quite a bit of time picking through what is going on and tried to draw some conclusions on what all of this means to the future of technical conferences and in particular our beloved IEEE ComSoC conferences. It turned out to be a highly charged topic for many that we spoke to. We are sure you have an opinion too. Comments, as always, welcome at the end of the article.
Alan Gatherer, CTN Editor-in-Chief
A Brief History of Conferences
Back in the dim and distant past of the 1990s, there were a few massive IEEE communications conferences (chiefly ICC and GLOBECOM, and partially ICASSP), a few highly respected but smaller conferences (ISIT, ICUPC, INFOCOM), and maybe an upstart or two (PIMRC). But the opportunity to publish revolved around these relatively few venues. Being assigned to organize such a conference was a major career recognition, with which came the ability to set the tone and direction of complete fields of study. They also set the seasonal calendar for academics who, like druids of old, prepared for and worshiped their rising and setting.
Then came the Internet. Suddenly, one didn’t need the might of a large organization to organize and publicize a conference, and the number of conferences started to proliferate. IEEE conferences also grew bigger and became more exotic in their locations as the IEEE became more international. In addition, one could publish online. ArXiv started in 1991 and, by the mid-2000s, it was already a popular source for what might be considered preprints of papers that were in the process of being published . The rise of TeX and LaTex also contributed to this ability of self-publishing . The number of publications to be read on a single topic exploded in most fields, and many complain that it has become hard to sort the wheat from the chaff and find the truly relevant works. Democracy comes at the price of uncertainty; whose papers do I read in the limited time that I have?
It was with the advent of video streaming that we really started to see the lecture and the conference, those bastions of presentation and Q&A, being challenged by an online format. Video streaming really had to wait for the MPEG4 standard in the late 1990s to properly take off  with Microsoft and Apple adding it to their platforms, but it was not until the mid-2000s that use of video streaming became commonplace. For the last ten years the conference, as the place where researchers unveil their new ideas, has been under siege.
Conferences have remained relatively popular with academics as a venue where ideas are swapped and collaborations born. Also, as a way to get ideas officially recognized in print, and as part of the process of peer review towards a journal paper. Chart 1 shows attendance number from 2010 onwards and we can see that they have remained fairly stable with a slight trend upwards, large conferences expecting about 2000 attendees and more focused conferences below 1000.
Chart 1: Conference Attendance by Year 1
The bar for acceptance at a conference is lower than for a journal publication and conferences have retained that seasonal regularity that encourages students to “get on with the paper,” especially with the promise of a trip to a large conference—still a highlight for many—at the other end. But even this role has been eroding due to the emergence of short letters that might be published faster than conference papers. For industrial researchers and implementers, traditional conferences have also become less relevant; now they have competition from many new, non-IEEE industrial conferences that are more focused on keynotes, panel sessions, and demo floors. Search engines and mass emailing allow small private companies to organize conferences on very specific topics. These conferences compete for the precious time of engineers. With industrial researchers becoming less interested in classic conferences, the IEEE have tried harder, adding more keynotes, discussion panels, and demo floors, as well as “industrial tracks” organized by senior industry researchers.
It is important to note that these efforts to make IEEE conferences more relevant to industrial participants are really playing to the basic desire of using the precious time at conferences to interact, to discuss, to criticize, and to challenge; a need that smaller commercial conferences had already identified and the IEEE has been catching up to. Industrial researchers in particular do not get as much of a chance to meet one on one with their peers, who might be in competing companies.
Where Are We Now?
This year of course, along came COVID-19, forcing all of us to look towards the logical conclusion of this 30-year journey. 2019 saw the conference circuit already under strain from global superpower tensions leading to issues with visas for many as well as the fear that leaving your country of study for a conference might make it hard to return. COVID-19 has democratized this problem by making it impossible for us all. Have conferences become redundant and unnecessary? If not, what will they become after COVID-19? The world will find a vaccine for COVID-19 and we will get back to some normality over the next few years, but will conferences ever go back to where they were? We think this time of sheltering in place is an opportunity to take stock of what we like and what we miss about conferences.
Conference organizers are rising to the challenge, even for the large conferences that support thousands of attendees. The traditionally massive ICASSP and ICC conferences just came to a close, and we saw that many of the difficulties that conferences faced early in the COVID-19 crisis have been solved. Live streaming of presentations followed by on-demand viewing allows participants to watch more presentations than ever before, no matter what time zone they are in. The IEEE staff have certainly shown they can run an online show. Commercial conferences are also responding well, some reporting their best, if virtual, attendance ever . Much of what we used to travel to a conference for can be done remotely. In some respects, virtual presentations are easier to attend. There is no running from ballroom 5 to meeting room 7 on the other side of the conference center. One can even attend two sessions that have been scheduled at the same time. You can simply click through from the conference agenda to the paper and, if it looks interesting, to the talk itself. We suspect that many papers that would have been presented to a live audience of just a few people in the past will receive a lot more visitors online over a longer period. This idea of a conference that lasts well beyond its initial opening has already been successfully reported in . Note that the attendance online was over 10 thousand, roughly five times the size of a large IEEE physical conference. Keynotes are also more easily watched online, as anyone who has ever ended up in the back left-hand corner of the main auditorium can attest to. Let us also not forget that many who could not afford to travel to Taiwan or Hawaii can now afford to attend a big international conference remotely.
The Future of Conferences
So why have physical conferences at all? We do not believe the physical conference is dead. If anything, it seems we might be entering a new era for conferences, dragged there by this global disaster. There is still a strong need on both the academic and industrial side to meet, discuss current trends, and network. Given all that we have laid out above, we would like to suggest a template for future IEEE ComSoc “hybrid” conferences that would combine both in-person and remote attendance.
- The virtual part of the conference would stretch over a certain period (say a few months). Attendance at this virtual part would only require a nominal fee. Longer and cheaper conferences are bound to attract a much wider worldwide virtual audience.
- All technical paper presentations would be online and take place during the virtual period. Any virtual attendant would be able to browse and comment on presentations. Authors would be expected to go to the effort of responding to posted questions, just as in the past they were expected to physically attend and present. And, with presentations being virtual, it could be enforced that they be given by the lead author. This would add quality to the exposition and richness to the Q&A, especially for presenters who might have struggled to respond in English. No more would we be plagued by “we have two minutes for questions” from a session chair valiantly trying to keep things on time. Likewise, the twin plague of the presenter who runs over time throwing the session into disarray would be eliminated.
- Keynotes tend to be associated with sponsorships and may continue to serve this purpose. They would be delivered online too. The keynote speaker might be required to attend the subsequent physical conference for a longer and more informal discussion and Q&A session. Conference attendees would be able to view the keynote beforehand and then come prepared with questions.
- The physical part of the conference would be held right after the virtual period, and the larger trends would be discussed there and then. The organizing committee might invite “best papers” to present longer versions with more discussion. There should be large spaces for networking and interaction, with demos, networking, and training, perhaps focused on young professionals and underrepresented groups.
- Proceeding papers, which are often redundant versions of more complete journal articles, could be eliminated altogether and the presentations be linked directly to those journals once available. In lieu of proceeding papers, the presentations themselves could be peer reviewed and later added to Xplore or some video-specific resource center. Most importantly, they could be cited and referenced. There is an opportunity here for IEEE to seize the market for peer-reviewed video content, and indeed some other publishers are starting to make moves in this direction. The IEEE would need to make a plan for archive and contribution recognition, which remains an important part of the conference process.
We hope that this speculative article can start a longer conversation on the future of IEEE conferences. We have probably only scratched the surface here, but we are sure that major changes are coming and now is the time to anticipate them so as to ensure that the IEEE emerges much stronger.
1 Thanks to Jimmy Le of IEEE ComSoc for this data.
Statements and opinions given in a work published by the IEEE or the IEEE Communications Society are the expressions of the author(s). Responsibility for the content of published articles rests upon the authors(s), not IEEE nor the IEEE Communications Society.