The lost utility of academic conferences

First drafted: 2024 May 14
Last updated: 2024 May 20

I don't get big academic conferences.

Major events like the International Conference on Robotics and Automation (ICRA) and the International Conference on Intelligent Robots and Systems (IROS) are meant to be a great opportunity for the academic and research community to get together on a yearly basis to discuss what's new and what's needed in a fast-paced field like robotics.

I don't know if John F. Jarvis and R. P. C. Paul, chairmans of the first ICRA conference in 1984 (back then simply the International Conference on Robotics), ever envisioned their 2-track, 75-papers¹, and less-than-400-attendees conference becoming the gargantuan affair of academic indulgence and inebriation that it is today.

This year's ICRA conference is taking place in Yokohama, Japan, as we speak. With what appears to be more than 6000 attendees, 30+ parallel tracks, 1765 papers presented, 10 competitions, 4 technical tours, and a showroom hosting about 135 companies and institutions, it's hard to tell what the goal of this type of event is anymore; or what to attend and where to be at any given time.

I know why I come to these events. I want to learn.

I attend these events so that I can spot a few new interesting trends and approaches that I may not have thought about. Thoughts and ideas that we could immediately apply to our research back home. Networking and meeting old friends are a plus, but not my priority. I can do that through other (much cheaper and efficient) means. Showrooms are cool the first time you walk through one. Afterwards, they all look pretty much the same (although, to be fair, a few cool things were showcased at this year's ICRA, like the LEV-2 microrover at the SONY booth. Very cool).

But with 30+ papers being presented at any given time with 5-to-7-min presentations and little to no time for in-depth discussions, I'm rarely successful at learning anything new from these events.

On top of this, and to make matters worse, I recently noticed a pattern toward increasingly mismanaged and poorly organized conferences: food run out on the first day of ICRA24 leaving ⅓ of the attendees without lunch or force to find it elsewhere, drinks or water was no where to find next to presentation rooms, sessions that should be close together are 10 min apart by foot, increasing number of speakers not showing up or with poorly prepared presentations (although this isn't news in academia), live streams broadcasted through low-resolution cameras pointing at screens, etc.

Remember, this wouldn't be an issue if these events were close to zero-cost for attendees. But those pay anywhere from $500 to $2000…for the registration alone.

So I think it's about time as academics that we start to demand to organizations like IEEE much higher quality standards and take upon ourselves to bring back rigorous scientific discussions to these events.

For me the ROI of attending these conferences has been barely positive, and it's rapidly decreasing every year. Large academic conferences like ICRA and IROS aren't optimized for learning. Not anymore.

Maybe I need to change my expectations or stop attending these and focus on smaller, niche events, like the Int. Conf. on Space Robotics (iSpaRo) coming up in a month, which I helped organized (shameless plug, I know).

The issue is always the same. How do you play enough of the game to remain relevant while actively avoiding most of its downsides.

To end this random thought on a constructive note, and if I had to choose, I believe these would be the key features of my ideal academic conference:

• One to at most three parallel tracks with sessions oriented around specific problems (e.g., place recognition under uncertainty) rather than general knowledge domains (e.g., perception).
• Acceptance rates of 20% if not lower for oral presentations, up to 35% for posters. Posters exclusively reserved to early career researchers (i.e., students up to PhD level; no postdocs or profs presenting posters).
• Longer paper presentations and Q&As. Top (awarded) papers discussed in roundtables with the authors and other experts in the field.
• Debates on key technical and non-technical topics (e.g., are conferences still a useful vehicle for scientific discussion?).
• Online first with worldwide local in-person gatherings. No need for long-distance traveling or cumbersome visa application processes.
• Some form of group physical activity that helps break the pattern of intellectual conversations and pretentious appearances (a good example of this was the 5K charity run organized in the middle of the ERF2024)
• Real-time translation for non-native speakers who choose to present in their mother tongues (forcing them to present in English can be a huge barrier for deeper scientific discussions)

Ok, maybe with the last one I went a little too far…but hey it's 2024.

D.R-M

¹as per https://staff.aist.go.jp/k.koide/acceptance-rate.html

Note: These opinions are my own. They are also highly dynamic and their second time derivative, albeit decreasing with age, is large. Experience provides me with new insights that drift former convictions. By the time you read this, assume my opinion on the subject has most likely changed.