How long does it take to prepare an academic presentation?
As an academic, one of the joys, and from what I hear from many other colleagues too, one of the struggles is preparing academic presentations, scientific talks that are generally offered to a group of peers who likely stand on the same sources of knowledge and expertise as yourself, or at least that’s what we think when we prepare our first few presentations. But over talks and years, you often realize that unless you have continued to publish and operate in a narrow sub-area and select bunch of peers who see you on a frequent basis (and the topic is highly specialized, and therefore you only expect the audience to consist of peers that write similar things), it is rarely the case that you can be preparing the presentation with such assumptions. In many cases, at least at scientific conferences, of course you can at least expect people to have the same set of foundational assumptions about terms – in AI, for example people would know about terms such as agency and learning and reasoning. But even here, definitions and notions can vary widely between groups and widely between sub-communities.
When a paper gets written for a conference, there is also an audience in mind, but it is often that the paper is self-explanatory, and good papers usually are to an extent. Now, it is unlikely that you will submit a paper to a journal that introduces the most fundamental concepts that one might come across in an undergraduate class. So even here, there is some expectation of the reader’s knowledge, but at least because of the length of the document, it becomes somewhat easier to provide hints about how the ideas have been developed.
All of this goes out of the window when you have to give a 20 or 30-minute talk on a topic to an audience consisting of people from different communities. I don’t think it’s ever easy to get this balance right. I’ve seen some very polished presentations, at least in terms of the visual style, as well as the points delivered, that still get lost among the audience members, just because there are such massive differences in the assumptions behind the topics. Precisely because my interest can lie at the intersection of logic and learning, I often see talks by machine learning people that are completely lost to the logicians, and vice versa. The logicians might struggle to see how exactly the data is collected, and what kind of assumptions are built into the machine learning system, whereas machine learning people might often see presentations from logicians and consider it too abstract. They might as well be talking about set theory or number theory, for all they care.
In that regard, the time it takes to prepare the presentation is non-trivial. Often these days, when I give talks, I try to include contributions and developments from some of my students as part of the overall picture, and this makes it all the more challenging to prepare a good talk. What I find somewhat frustrating here is that the assumption that all this should be quick and relatively painless (compared to course material), at least in the academic community. Because we give so many talks, the time taken to prepare this presentation must be very short. And I don’t know if this assumption is something I tell myself, or I’ve convinced myself of, or if it is also held by others. But for as long as I can remember, I’ve held on to this assumption, and there is this feeling of guilt when the presentation takes too long.
Today, for instance, I’m wrapping together content from a few talks I’ve given, and all I’ve managed to do all day is perhaps a portion of the final product. And again, this is still not accounting for the time it will take to polish, reorder, and move things around. Of course, the time I’ve spent on it is completely justified, and I don’t think I could have done any of these things differently because I needed this time to work out how to best deliver the ideas and explain them in a precise enough manner for a technically-minded yet diverse audience. And so, you really start to need to acknowledge, I think, for the sake of the enterprise, that presentations take time. And the assumption that it can be done quickly, or should be done quickly, is fundamentally wrong. If the end goal of science is to advance the society in terms of knowledge, a part of which also means communicating effectively, then presentations should be given a good amount of weight. And this is something that you want to be telling yourself as often as you need to.
