The first TCS+ seminar of the semester just took place today, and I’m happy to report that the delivery went well! The baby has all its fingers and can be seen here. The parents are still recovering from their emotions (which included the fire alarm of MIT’s Stata center setting off right in the middle of the seminar: check out the flashing lights appearing in the video feed from MIT starting around 27 min in! Luckily, it was a false alarm…The hangout was hosted on my laptop and I had to leave it open in our seminar room for the live stream to continue during our forced break) and swear they’ll never be caught at it again (we all know how these kinds of promises go…next talk, by Anup Rao, in two weeks!).
I want to thank the speaker Ronald de Wolf, first for a great talk (no surprise here), but also for his unbounded patience and determination in preparing for this event and helping us make it work. The talk was joined live by eight quite diverse groups, from ETH in Zurich, Columbia and NYU in New York, Max Planck Institute in Munich, University of Edinburgh, University of Utah, IQC in Waterloo, MIT (before the fire alarm) and University of Michigan (after the fire alarm). Thanks to all the organizers for setting this up in their respective universities; there were relatively few technical glitches and we’re glad almost everyone who wanted to join eventually made it. We had roughly 35 additional external viewers at any point during the talk, which for a first is pretty good! Let me also thank the moderator (and organizer) Oded Regev who did a fantastic job. I’m especially grateful to Oded for keeping things under control when I was getting too confused to keep up with my share of the work (what with the fire alarm going off, the comments to monitor, the extra groups trying to join…). Parts of my job was to decide on which group got the focus (in the main window) and I admit to somewhat losing track of this during the questions session at the end. Finally, I thank Daniel Burgarth and Matt Leifer, organizers of Q+, for their invaluable advice.
The main negative comment we got so far has to do with the video quality. It is quite good when you are inside the hangout, but the YouTube feed is not great. It is good enough for the slides to be readable (and the sound quality is quite good), so overall the experience remains pleasant. I wish we could somehow improve the resolution, but this seems to be locked in by default. We’ll also experiment with different ways to make the slides bigger (as a percentage of the speaker’s screen) in the future. For now, a pdf of the slides can be found on the TCS+ website.
What with all this, did I get anything out of the talk? Well, I really like the paper, and Ronald is such a clear presenter that you should really go and watch the talk — no spoilers. Looking at the list of authors for the paper, it’s natural to wonder how the result came about. Here’s an urban legend about it. One of the authors (let me not give names) was walking his dog in one of Brussels’s public parks. A researcher at ULB, unbeknownst to him he happens to bump into a colleague. The two men had never met, but quickly realize they come from the same institution, and immediately start exchanging research topics. The first exposes a deep problem he has been thinking about for some time; he is stuck: how can one prove lower bounds on (a variation on) the rank of a certain class of matrices? The second sees a connection: this notion of rank precisely appears in quantum communication complexity. Could techniques from that area help? He happens to be in close contact with an expert in the field. As it turns out, this third protagonist has worked on a very closely problem…there is hope ;-)
All in all, it was quite fun! I hope all who watched enjoyed the experience and will come again. Suggestions for improvements in the format, the choice of future speakers, and general comments are very welcome! (Please post these directly on the TCS+ page on Google+, that everyone is encouraged to follow :-))