Well, an informative day in college. Aside from my normal lecture,… - John C. Kirk
Feb. 20th, 2003
Well, an informative day in college. Aside from my normal lecture, there was an extra seminar this afternoon. I didn't have to go, but it's relevant to the area I'm doing my project in, and my supervisor emailed me specifically about it (as opposed to all MSc students), so I figured I'd earn some brownie points by going along. There was some interesting stuff there, although a lot of it went over my head. Anyway, I didn't ask any questions at the end, as I was worried that everyone else would already know the answer. However, I did ask a question afterwards, when there were a few people standing around. Basically, a standard approach in anaphora resolution is that if you get "John did x. Bob did y. He did z." then it will match "he" to "Bob", as the nearest name. My question was whether there were exceptions for dialogue, where it would alternate between two people. Turns out that this area does cause major problems, which suggests some good scope for research :)
Anyway, the prof had said at the end of the lecture that a few people would be going off for coffee afterwards. However, there were only five of us left at the end, so he decided it was better to use his office, rather than go to Starbucks or something. He said I was welcome to come along, so I did, although I don't drink coffee (I just had water). I felt like a bit of an intellectual pygmy, since the others were two PhD students, one postdoc (who'd given the talk), and the prof, so I kept my mouth shut most of the time, to listen to them. I did also notice how many textbooks the professor had, and how many of those had his name on the spine... Random comment at one point was when Shalom (the professor) asked Judita (the speaker) "Where's Charniak nowadays - are you in touch with him?" Now, when I did my undergrad degree, the main textbook for the AI course was by Charniak and McDermott (now out of print, unfortunately), and he defined some key techniques that were referenced in the day's talk. So, hearing this is the equivalent of answering the phone, and being told that Steven Hawking wants to speak to one of my Physicist flatmates :)
Anyway, on to the key point. We moved onto the subject of PhD funding. It turns out that the NLP group at Kings is entirely reliant on the EPSRC and the ESRC - they don't get any private sponsorship. So, the others were having a shared moan about the way that those organisations are really cutting back on the money they give out. So, basically, the group won't accept any PhD candidates unless they're self-funded. Which counts me out (at least for now). Hmm. So, time to start looking around elsewhere. I'll investigate the other London universities initially (particularly Imperial), but if that doesn't pan out then I may need to look further afield. I've never seriously considered leaving the country before, but this may be the time to start...
By the way, quote of the day from yesterday's NLP lecture: "A language is a dialect with an army".
Slightly random dance class - Dorothy got us doing a party game (more or less) to teach us a foxtrot step. Basically, all the guys lined up in a row. Then the bloke at the front had a rucksack, and took a step back, and a step to the side, then passed the rucksack to the person behind/next to him, who repeated the process. It took us several attempts (about ten, I think) to get the rucksack all the way down the line, much to the amusement of the girls who were watching on the other side of the room :)