I gave a talk today on my research on segmenting ontologies. It went well. There were about 20 people in the audience (in total 60 people attended the conference, but they were distributed between the three different workshops). Some people had specifically come from one of the other workshops just for my presentation.
I started off with a very good opening statement. This caught everyone's attention and got them interested. I said (view my powerpoint presentation for the graphics):
"An ontology is like a labyrinth, very complex and almost impossible to comprehend at first glance. Beneath the ancient city of Knossos on the island of Crete there was a very large and complicated labyrinth, which had a fearsome Minotaur within it. One day the hero Theseus descended into the labyrinth and killed the monster. He then found his way back out by following a trail of red yarn he had laid behind him as he was exploring the maze, thereby reducing the complex labyrinth into a simple corridor. Ontology segmentation aims to do the same thing."
After saying this I could see people nodding and agreeing. This is what Jeremy Wiessman in his book calls the "ah-ha" experience that one should aim for in any talk or presentation.
I also used the powerpoint presenter's display view. A feature which strangely no one else seems to take advantage of. It may be because it is somewhat tricky to set up. It displays a different view on my laptop's screen to what shows up on the projector. In addition to the current slide, I also get a running timer, my notes (though I didn't make any), a preview of the next few slides and the ability to jump to a specific slide immediately without having to scroll through the entire presentation (this was very useful for the Q&A session).
I wasn't particularly nervous, though I um-ed and ah-ed a bit too much for my liking. I did however manage to hold the attention of the audience for the 20 minutes of my talk. No one went to sleep, no one started checking their email, everyone was looking at me throughout. Pretty good, considering my talk was in the graveyard slot: just after lunch.
There were initially some general questions and clarifications from the audience. However, then Chris Welty bombarded me with questions. He misunderstood the filtering segmentation algorithm walkthrough I had presented. I corrected his misunderstanding of the final directed acyclic graph connectivity of my segmentation example. He also didn't understand the need for the reciprocal filtering. I answered that one by explaining that it worked well for GALEN, which is fair enough; however, on reflection, I realize that some reciprocal link filtering is necessary in every segmentation algorithm, since otherwise everything will be connected to everything else and the "segment" will end up becoming the entire ontology.
Harith Alani, one of the workshop organizers from Southampton University, told me afterwards that he liked my presentation and talked me for submitting my paper at such short notice. Another conference chair, Derek Sleeman, also seemed pleased and interested.
All in all, I'm pleased with my first conference presentation.