As every other aspect of human life, science communication has suffered a significant setback due to the ongoing Covid-19-induced pandemic. While regular meetings of scientific teams can be effectively held online, through zoom or skype, it is the big conferences that are suffering the biggest blow. And this is not good, for several reasons.

In the past I have expressed in this blog my observation that there appeared to be too wide an offer of conferences in my discipline, fundamental physics. The new situation however suggests that the optimal number of conferences is probably much closer to the value we had before than what we have now. Many events this year have been converted into an online format, a few have mutated into a hybrid form, and a significant number have been canceled altogether. But to me, online conferences only partly meet the needs of healthy scientific communication.

While an online event still allows scientists to deliver talks and disseminate their research results, the real dissemination impact is seriously hampered. At a conference you usually sit through a session, and get to listen to many more talks than you would in principle be interested to hear, which is good. It is easy to be intrigued by a good presentation even if the topic is not exactly overlapping with your specific field of research, and this leads to useful exchanges of ideas and discussion. During coffee breaks and social activities those conversations continue, often leading to building bridges with other groups of researchers, creating the conditions for further fruitful collaboration. All of this is, unfortunately, impossible to achieve through videoconferencing tools.

I am currently sitting (in person!) in the hall of a conference that is taking place in Kolympari, on the north-west coast of Crete. The conference organizers decided to opt for a hybrid formula, but the uncertain situation of the epidemic in Greece, combined with travel restrictions from several countries, kept the in-person attendance very low. I myself could only attend by traveling to Crete with my personal funds and during my vacations, as my research institute did not yet raise its travel restrictions policy (a good decision, if you ask me).

In a hybrid conference all talks are delivered online, and only very few also in person; the conference hall becomes a rather empty space, both physically and metaphorically. Even questions must be asked through your laptop, as it makes things easier logistically (I imagine other conferences chose a different setup, though here this is the rule). And more to the point, the very scarce in-person attendance practically kills the healthy coffee-time conversations and exchanges, as the number of possible one-to-one interactions decreases dramatically (with 10 persons physically attending, you have 45 possible eye-to-eye discussions; with 100 persons, you would have 4950).

So I am led to wonder what other model could preserve the benefits of our old normal practice, if a virus like Covid-19 were to keep us in this sorry situation for a prolonged period of time. What can we bring in as a reserve weapon to foster scientific exchanges? I believe the internet must come to the rescue, and maybe I have even a simple idea on something which might work.

The idea is based on the old concept of the "academic quarter-of-an-hour" (I suspect there are other ways to call this rather than my dumb translation from the Italian "quarto d'ora accademico", but you get the point anyway). Every researcher could pick a hour-long slot per week when they make themselves available for up to four discussions with any colleague, specifying in a very brief itemized list, maybe endowed with pointers to recent papers, what is their current research focus. A simple web interface could allow participants to search through the topics they are interested in, and book time for conversations with the available researcher, with the option of allowing third parties to join.

The above would not substitute an in-person interaction, of course, and yet it would be a step in the direction of an easier interaction with people you do not know who have similar research interests and goals to your own. It would be an affordable time investment (one hour per week is 3% of a typical work schedule). I envision two different models for the conversation: a didactical one, useful for students, where the available researcher would describe in simple terms the topic of interest; and a collaborative one, where it is supposed that the interlocutor has knowledge to share on the matter, and the discussion is a two-way one.

Now, since I believe people should put their money where their mouth is, I will give this a try. My skype name is tonno923, and my email address is firstname.lastname (at) . I will offer three slots for anybody who is interested in research topics given below, and who would like to chat for 15' with me on those subjects. Interested colleagues should post a comment in the thread below specifying when they wish to talk to me, on what subject, whether the conversation could be open to others. They should pick a time slot among those offered, send me an email with the same information posted in the thread, and wait for my confirmation. Then at the agreed time we will meet on skype.

The three offered time slots are:
- Wednesday, Sep. 9, 1PM CET
- Wednesday, Sep 16, 6PM CET
- Wednesday, Sep 23, 6PM CET

The topics I am happy to discuss with you - either in "didactical" or in "collaborative" mode, are:
1) New ideas in machine learning for research in fundamental physics
2) Taming systematic uncertainties with machine learning methods
3) Optimization of detector design with differentiable programming
4) Measuring muons through their energy deposit in calorimeters

For topic (1) I do not have supporting material to suggest; for topic (2), I point you to a recent preprint I published with P. de Castro Manzano, and which will become a chapter of a book on artificial intelligence for HEP (a collaborative effort); one recent publication of mine of relevance to topic (3) is this one, although it does not involve differentiable programming; and for topic (4), you might want to check this recent preprint.

I might add that you will find more information on my research output in my personal web page. So, let's give this a try, and see if it makes any sense! Who knows if some of you out there are working on the same things and are willing to join forces or exchange ideas on them?