Thanks, Nad. But I don't _literally_ have an ulcer! That was just a metaphor.
I really want to go to this Neural Information Processing Seminar, since there are probably lots of smart people interested in "networks" there. I was invited to give a minicourse there last year (on information geometry), and I cancelled at the last minute because I was too busy. I don't want to cancel again. If I wanted to give a low-stress talk, I could talk about my mathematical ideas on network theory. But I think for this audience it will be better if I walk about something more "practical". I feel pretty sure I can give a mildly interesting talk on climate networks and El Niño prediction. I'll do this if I can't do something more exciting. But I've been wanting to push myself into new territory... and get some of the machine learning / computation people here involved. We'll see.