On the efficacy of virtual seminars
Hugo Camargo, Michal P. Heller, Ro Jefferson, Johannes Knaute, Ignacio Reyes, Sukhbinder Singh, Viktor Svensson
OOn the efficacy of virtual seminars
Gravity, Quantum Fields and Information research group ∗ Max Planck Institute for Gravitational Physics (Albert Einstein Institute)Am M¨uhlenberg 1, D-14476 Potsdam-Golm, Germany
Abstract
During the SARS-CoV-2 pandemic, theoretical high-energy physics, and likely also the major-ity of other disciplines, are seeing a surge of virtual seminars as a primary means for scientificexchange. In this brief article, we highlight some compelling benefits of virtualizing research talks,and argue for why virtual seminars should continue even after the pandemic. Based on our ex-tensive experience on running online talks, we also summarize some basic guidelines on organizingvirtual seminars, and suggest some directions in which they could evolve. ∗ Web: aei.mpg.de/GQFI; Contact: [email protected]; Members: Hugo Camargo, Michal P. Heller (PI), Ro Jefferson,Johannes Knaute, Ignacio Reyes, Sukhbinder Singh, and Viktor Svensson. a r X i v : . [ phy s i c s . e d - ph ] A p r Introduction and overview
One of the main driving forces behind scientific progress is the free and timely exchange of ideas. Areal breakthrough in this direction happened already about 30 years ago with the launch of the arXiv preprint server, which allowed for a free, timely, and scalable dissemination of the newest knowledgepresented in research papers. Despite the enormous progress in technology since that time however,other equally essential channels of research dissemination namely, research seminars and conferencesdo not seem to have evolved proportionally to technological advancement.The advent of the SARS-CoV-2 pandemic has, however, dramatically laid bare the severe limitationsof traditional seminars and conferences .As a result of SARS-CoV-2, several scientific meetings and visits have since been canceled and replacedby a tangible sense of vacuum and lack of established alternatives. But innovative solutions haveemerged. To keep science progressing in these difficult times, an increasing number of groups areswiftly migrating to online communication platforms. Several groups have, in fact, already upgradedtheir regular seminars to virtual talks, and the first virtual conferences have been organized, withmore brewing in the pipeline. But one wonders if groups and communities will resort to their formerhabits after the pandemic has passed, or whether virtual seminars and conferences will permanentlyjoin our traditional array of science communication channels.In this brief article, we outline reasons for why we believe that virtual seminars and conferences might,or rather should, be here to stay.Ours was one of the first groups in our discipline to have launched an entirely virtual seminar serieslong before the current pandemic appeared on the horizon. In the past 30 months, we have hostedmore than 80 virtual talks. Since the first broadcast of our group seminar, we have also launchedthree new specialized seminar series in collaboration with partner groups:(i) Virtual Seminars on Quantum Gravity and Information (QGIseminar.org) in partnership with the Amsterdam String Theory group ,(ii)
Virtual Seminar on Tensor Networks in High Energy Physics (hepTNseminar.org) in collabora-tion with
DESY Zeuthen , MPQ , and
Perimeter Institute ,(iii)
Virtual Seminar on t-, T- and µ -dependence in Quantum Field Theory (tTmuQFTseminar.org).The recording of the vast majority of our virtual talks are made available afterward on our YouTube channel (youtube.com/c/GravityQuantumFieldsandInformationAEI) and are free to watch by anyone.Some of us are also presently organizing a virtual conference on
Complexity from Quantum Informationto Black Holes under the aegis of Leidens Lorentz Centre in June 2020.The entire machinery and impact of virtual communications and how they relate to traditional methodsof communication is, of course, an active area of academic research in social sciences. Our viewspresented here are based entirely on the first hand experience that we have accumulated from ourvirtual seminars. The purpose of this article is to make use of this timely opportunity to share a We are writing this article from the perspective of our field, i.e. theoretical physics at the intersection of gravity,quantum field theory, and quantum information science. The relevant arXiv category is primarily hep-th and to a lesserextent quant-ph , nucl-th and gr-qc . Also, the scientific environment we refer to is Germany, where our group is located. Reduced cost:
Perhaps the most apparent benefit is saving taxpayers money. If we assume thatorganizing an average speakers visit within Europe costs 500 EUR, a rather modest estimate,then eighty virtual talks save costs amounting to more than one year of a Ph.D. student salaryin Germany, or equally a significant fraction of a post-doc salary. One can therefore imaginethat three geographically- and scientifically-close group virtualize their talks and invest savedresources into hiring a joint Ph.D. student or a post-doc. The savings from curbing long-haultravel are, of course, much more pronounced. (Our virtual seminars have had fifty speakersfrom the Americas and ten from East Asia.) Streaming a seminar also obviates the need topresent the same talk multiple times to geographically disconnected audiences, and thus thereare multiplicative cost savings for every talk. Such savings can be crucial for small-scale groupswith fixed, restricted budgets such as ours. They can also provide additional employment op-portunities in a field where positions are scarce and the number of applicants large. We alsonote that the costs of organizing traditional conferences can be extremely high.2.
Ease of organization and timely research dissemination:
Virtual seminars and confer-ences can also save the sometimes massive time and effort devoted to organizing academic visits.The reduced organization time not only eases administrative workload but also demonstrablyleads to timely dissemination of new research. Many speakers of our virtual seminars were ableto present a paper within a couple weeks of releasing it on arXiv . In contrast, there is usually asubstantial time gap between the release of new research and presenting it in-person at a seminaror conference. Thus, virtual seminars can help disseminate new research faster.2.
Scale and quality of presentation:
The next obvious benefit is that virtual seminars canreach out to a much larger global audience across different time zones. For example, one of ourrecent
QGI seminars had 180 participants, of the order of a medium-sized conference in ourdiscipline. An easily accessible large-sized audience can accelerate the visibility of new researchand enhance scientific networking. Moreover, virtual mediums could also enable broadcastingcolloquium style talks to the general public, which can help scale-up scientific outreach moregenerally. (In fact,
Perimeter Institute has already been broadcasting its on-site outreach talksfor quite some time.) Access to scientific seminars can also be of immeasurable benefit to lessprivileged research audiences from developing or geographically isolated countries.We have already noted that, somewhat unexpectedly, the quality of a virtual seminar does notseem to be compromised in any meaningful sense when compared to in-person delivery. In fact,there seem to be some small but significant advantages. Firstly, the visual and audio deliveryof a virtual seminar can be more effective since it offers everyone in the audience the first-row experience. (In a popular seminar or a large conference, latecomers often find themselvesstraining to catch the presentation at the end of a large room.) It is also simpler to use propsin a virtual seminar: one can simply deliver the talk from where the accessories are locatedinstead of carrying them to the audience. Imagine an experimentalist giving a talk directlyfrom their lab or from the LHC, or a theoretical physicist showcasing a real-time simulation ona supercomputer or on their local computing cluster. This may seem like belaboring a point,but it illustrates very clearly that virtual seminars are not merely replacements for traditionalseminars, but can potentially serve as a platform for a more vibrant and impactful delivery.4.
Towards a global online video arXiv : Virtual seminars are much more quickly recordedthan in-person seminars, and do not require expensive camera and recording equipment. Someavailable software solutions allow recording seminars locally on any participating computer withthe press of a button. If virtual seminars take-off as a regular scientific activity, it is not hardto envision the emergence of a shared catalog, similar to arXiv for research papers, of linksto past and upcoming virtual seminars across the globe. Links could be posted by researchgroups anywhere in the world. (Some informal catalogs, e.g., social media pages, are beginningto appear, but they are not conveniently organized for scalable access.) At some stage, it wouldalso be natural to incorporate individual submissions, that is, a single researcher can record apresentation of their work and submit the link. (This may require moderation of some sort.)One new possibility on this front is providing either a modular format for a talk of, e.g., 15minutes of introduction and summary and, e.g., 30 minutes of detailed explanations, or evenrecording one very short talk with the key punchline (an analogue of a letter) and a longer talkexplaining details (an analogue of a technical paper).Several physics institutes, such as
Perimeter Institute in Canada and
Kavli Institute for Theo-retical Physics in USA, already maintain substantial video repositories of their local seminars ,which have proved to be a tremendously valuable resource. But these recordings are of seminarsthat were presented by researchers either working at or visiting these institutes. Therefore, suchrepositories are less encompassing and can only grow at a rate that is once again limited by thefrequency of academic visits; thus, they do not currently serve to disseminate new research insuch a timely and unrestricted way. Contribution to climate change activism
Perhaps the most significant environmental impact of theoretical physics comes from academic travel.Large-scale virtual talks can dramatically reduce the associated carbon footprint. Our own virtualseminar series provides a small-scale but encouraging illustration. If we had instead opted to organizeall the seminars as traditional on-site visits, we would have had to bring twenty-four speakers fromEurope, fifty from the Americas, and ten from Asia. A crude estimate of the carbon footprint fromflights would correspond to seventy tons of carbon dioxide in under two and a half years. This emission,according to publicly available data , equates to around thirty tons of carbon dioxide per year, whichis roughly speaking twice the contribution of an average North American and three times more thana European. The above estimates ignore the emission costs of the global computing and transmissioninfrastructure, which are undoubtedly significant (back-of-the envelope calculations estimate them tobe of the order of 10% of the saved emissions), but more easily transitioned to green energy than airtravel. The experience of giving and listening to a virtual seminar is, of course, different from that of atraditional seminar. It may be a particularly odd experience for some speakers, given the absence ofvisual cues about how the seminar is progressing. Below we share some guidelines that we learned whiledeveloping our virtual seminars, which might be useful for groups who are contemplating transitioningto virtual seminars. a) Our seminars were broadcast only to active researchers who subscribed to our mailing list.b) The virtual seminar room link was only shared with subscribed participants by email, and wasnot listed on our seminar advertisements on public domains.c) Seminars were advertised at least a week in advance, and a reminder was sent on the day of theseminar.d) We invited the speakers to set-up their talk at least 15-30 minutes before the talk. a) Seminar lengths of up to 1 hour (including questions) seemed to work well. Most people wereusually happy to stay up to an hour in our seminars, but longer talks did not work as well.We also experimented with 45-minute seminars and found that they were significantly morecomfortable to watch and follow.b) Digital board talks generally worked better than blackboard talks, due to better lighting andreadability. See, e.g., data.worldbank.org/indicator/en.atm.co2e.pc.
4) Blackboard talks work better if the speaker divides the blackboard into sections and manuallyadjusts the camera position to zoom to one section at a time. (The entire board is difficult toread on a computer screen.) a) We asked speakers to pause periodically in their presentations to break for any questions ex-plicitly. Sometimes including ”blank” slides to break for questions at regular intervals helpedregulate the pace of the talk.b) To improve the lack of visual cues, both for the speaker and the audience, we felt that it wasbetter to include a live visual (in a small window at the top corner) of the speaker while theywere speaking, and also of some of the organizers (for the speaker to collect cues). a) We found it better if people virtually raise hands (e.g., using such a built-in facility in therelevant software) before a moderator calls out their name to ask their question.b) It is better to make any announcements at the beginning of the talk, as the audience tends todissipate very quickly at the end of a virtual seminar.c) We announced at the start of the seminar that the seminar will be recorded and uploaded onour
YouTube channel.d) Excessive text exchanges over publicly available during the seminar chat are distracting. a) If the software allows, setting up multiple online hosts for the seminar helps in case there is atechnological failure at any one end.
In the end, we would like to stress that we do not, of course, believe that all scientific meetings mustbe virtual. In particular, the benefits of the human interaction component of traditional seminars andconferences can hardly be neglected. But we do suggest to keep using the virtual medium of scientificexchange whenever deep interactions are not necessary, and to maximize the impact (and cost-benefit)of in-person seminars by sharing them virtually.
Acknowledgements:
Our group is supported by the Alexander von Humboldt Foundation and theFederal Ministry for Education and Research through the Sofja Kovalevskaja Awardthe Sofja Kovalevskaja Award