Technical Communication and Research Methods
This week I arrived on time. We talked a lot about how to effectively present. A lot of which we've been exposed to before, but good practises always bear repeating :) Last Thursday involved introductions (which I missed) and a seminar on LaTeX, which I'm thankfully somewhat comfortable with already.
Being Helpful
One of my favourite things to do is to be of help to others. I got to help a classmate diagnose part of a problem, though I'm not sure if it was the fundamental problem for their project. Good luck to them. It unfortunately took a lot of time I needed for something else, though.
Graduate Student Representation
I feel like a very ineffective Graduate Student Representative so far, mostly because Graduate Students are far flung. There was a department meeting which was full of funness, where I discovered a Grad Info day was happening yesterday night (for undergraduates who are considering graduate studies at Guelph). I went about and harassed some undergrads, enticing them to come on short notice with the promise of free pizza.
Neural Networks
Neural Networks was enjoyable. We got dates for our projects now, which I should really enter into my calendar before I forget. A lot of it is familiar from Computational Intelligence and from the course I sat in on in New Zealand, but I try not to take knowledge for granted here. I've made that mistake too many times in the past :D
Grad Info Night
After class, I waited for Liv to be done her blood donation, but timing was confused. I ended up being of use at Grad Info night and was there talking to students for two hours, and I enjoyed two slices of pizza. I even secured a slice for dear Livi. Nom nom nom. Attending department meetings and this event helps me understand the scale of CS at Guelph. I mean, the numbers for a year are in the dozens, and at any given time for the whole department, a few hundred perhaps. This in a country of over 33 million. (Yes, I should reduce that to the population for the relevant demographic, but oh well.) Things operate on a more individual level than I thought.
OCUS and dodging hammers
Afterwards, I checked what time the dodge ball game was going to be. (I'm on an intramural team for OCUS, the union for off-campus students.) It turned out it was going to be at 11:30PM! I thought I might stick around and work, but a movie was playing in the lounge, "Dinner for Schmucks.' I only saw the last ½ hour, but it was less painful than I would have expected. After that came "The Hangover," so I decided to go home for a snack and see Livi.
I made it back to campus in time to practise a bit for dodge ball, and that's when I remembered I sucked. We ended up walking to the Athletics Centre in shorts and t-shirts, which was amusingly cold. We had to meet a female quota, and initially we weren't going to, but then magic happened. Also, some creative aliasing went on. We ended up playing about 5 matches and I only went in on two, since we have too many guys, and we had to switch out, and I lack confidence.
However, those two matches I played were glorious, I say! No hammer could hit me. Except for those that did. I even managed to hit a couple people out and catch a ball aimed right at me. It was great. Afterwards, I walked and talked my way back to the UC and waited to finally be ferried home.
CIS*2750 Software Systems Development and Integration
This is the course I TA. I was hoping to have a test plan done for the automarker I have to write Real Soon Now, and have it done for Monday. However, I ended up spending all of Sunday celebrating my birth instead. So, I spent Monday working on it, but due to car troubles, I didn't have as much time as I needed, and Tuesday (as described above) was quite busy (as evinced above). So, I'm trying to finish it today. Yay!
Natural Language Processing
I really wanted to smooth out the issues with my implementation before this semester really got under way. Sadly, that may not get to be. I am going to write a post about its development at some point, though, and then perhaps updates on it. One of the latest things I added was a display of the correct derivation of a sentence as though it had been parsed from a model. I now can find sources of error where certain actions are scored with a probability of 0! I next need to find out how valid, observed actions can fail to be learned by the model, resulting in p(a|b)=0, grr. Hopefully there's another small bug I get to fix.
Now
I've just had a pleasant morning working from home with Liv about. Now to go be a TA and hold office hours, though.
No comments:
Post a Comment