eigenspace

Still don't have the ankled person model quite right. Better, but not quite right. I'm getting to the point that I may just write my own kinematic tree code so that I can finally stop wasting so much time getting placements set right. It certainly would make things easier.

I've gotten it sampling initial states now but it's still giving fairly bad performance. I recall that I never completely tested the random generation of Gaussian samples, which might be the problem but I suspect other issues. It seems that one of my initial fears regarding the relationship between legs was necessary. There is a small (but notable) difference in the phase timings of one leg relative to the other. I'm now working on fitting a mixture model to the joint space but it is fairly tricky. The badness of fits that I found with the independent data is much worse in the joint space. Part of the problem is numerical instability since the nature of the data is many clusters with high variance along one axis and low variance along another. In fact, much of the data is almost linear in time which causes all kinds of problems. Compound that with the problem of outliers and suddenly things are really messy.

Anyway, I made spring rolls on the weekend but ran out of rice papers. They were quite good and I am craving them right now so I'm going to pack up, head home and buy some rice papers to finish off the filling and have a very tasty dinner!

Well, I'm taking the approach of fitting a Gaussian mixture model to the data and sampling from that. The fits look good but it's a bit complicated sampling from the conditional distribution using my current code. The other thing I've been working on is getting the ankled model working. Right now there is something wrong with my transformations and the feet end up in a very incorrect place but that should only take a little bit to fix up.

Anyway, Alissa is coming to join Liv and I for some margaritas and guacamole and, if people feel like it, some mayan hot chocolate! Time for me to get things cleaned up and set out.

Well, I checked how it behaved in expected value and the results weren't as promising as I could've hoped. It generally does do better than the original style of initialization but not as significantly as I hoped. Moreover, it only works well from certain starting points. The curve fits are pretty bad at the discontinuities, which could be part of the problem. Another possibility is that I may be making a false assumption about the state timings between legs. Essentially, I'm assuming that if one leg is 75% of the way through a swing phase then the other leg is 75% of the way through a support phase. I know there is a short (and insignificant, I had hoped) overlap where both legs are in the support phase but it's possible that this is more important than I had thought.

I have some ideas to investigate this which I'll explore later. At the moment I think I'm just going to get some food and maybe head to bed.

It's been a fairly productive day. I've gotten data extracted from a successful walking simulation and have been trying to fit some sensible curves to it. Unfortunately, the nature of the simulation is that much of the data is discontinuous and I'd rather not have to fit piecewise curves but it might come to that. As it is, the fit curves don't generally provide good starting points. I'm wondering if it may not be better to only fit one known good cycle to a curve and use that data instead of trying to fit all of the cycles which vary in quality and aren't quite properly aligned.

I'll meet with David tomorrow morning to see what he thinks. I've not had a chance to compare the fit initializations with the old ones in expectation though and that may be the deciding comparison. Whether it works well for one random seed doesn't matter. What matters is if it works better on average. And, of course, whether it will work on a smaller slope.

Didn't go swimming today. I got deep into hack mode early and didn't emerge until 5pm when I realized that one bowl of Cheerios doesn't make a complete days meal. I was faint with hunger and spent the next two hours snacking, making dinner and eating it. Of course, I overstuffed myself and was too full to go swimming. By the time I was comfortable I was back into deep hack mode and didn't emerge again until after it was too late to go swimming. Ah well, I'll go tomorrow.

Well, Liv's email has been acting a bit funky for a few days and finally refused to be cooperative today. I investigated and it looks like she had over 24,000 spams in her junk folder which was causing the underlying filesystem to have problems since with maildir every messages is a file. Rebooted, deleted the offending folder and everything seems to be happy now. I'm going to cross my fingers.

Been a non-weekend aside from Friday evening and swimming on Saturday. Today I pretty much did nothing but play a little Diablo II and I've been entering recipes into Gourmet which seems like a nice and promising tool, even if it does have some stability issues. If anyone has any interesting recipes, let me know and I'll give them a whirl and add them to my database. The nice thing about this program is that it can export to HTML so I'm hoping to add some of the recipes to a webpage sometime.

*shrug* Anyway, now I'm bored so I think I might look through some recipe books and look for ideas.

Well, last night was quite interesting. I made dinner (which turned out quite nicely, if I may say so) and we had margaritas with dinner and a nice bottle of port with dessert. Port is....well, it's just short of ambrosia I swear. We had a night of great conversation and indulgence and realized a long standing dream.

I'm quite proud of dinner, I spent about two or three weeks planning it, selecting dishes, tweaking them and thinking up garnishes. It had about 4 courses, each of which was planned to lead well into the next. The portion sizes were deliberately kept moderate in order that no one got uncomfortably full and the courses were somewhat spaced to allow breathing room. Here's how it went:

First Course (Appetizer) - Sweet and Sour Ceviche
This was tilapia marinated in lime and orange juice. I served it in short ice cream dishes that were layered with leaf lettuce and garnished with tortilla chips.

Second Course (Cheese) - Queso Fresco Blanco with Fruit and Water Crackers
Exactly as described, the fruit was grapes and clementine pieces. It was meant as a light prelude to the main course and helped cleanse the palette.

Third Course (Entree) - Creamy Mushroom Soup with Croutons and Manchega
The soup was not a regular cream of mushroom, it was a light wild mushroom soup which was enriched with some heavy cream. The croutons were seasoned with garlic and olive oil and topped with parmesan reggiano before being toasted. The soup was sided with a Spanish Manchega which is a tangy cheese that I thought offset the soup well.

Fourth Course (Dessert) - Molten Chocolate Cakes with Vanilla Whipped Cream and Port
Molten chocolate cakes are a chocolate cake except that the middle is literally, molten chocolate. I made my own whipped cream, flavouring the whipping cream with vanilla extract that I brought back from Mexico. (Note that I don't have a mixer, so my arm got quite a work out whipping the cream.) This was served with a bottle of late bottle vintage port which matched the chocolate well and tasted fantastic, like port always does.

Everything went off well and tasted good. The only real issue was that I slightly undercooked the molten chocolate cakes and so they fell apart when I tried to plate them. They still tasted good, regardless of the presentation.

Anyway, I've spent today recovering from all the drinking. I managed to go swimming although who knows how I managed it while half hung over. I'm starting to feel a bit better now that the evening is wearing down but it's been a very fuzzy headed day.

Life has been quite good these past few weeks and yesterday evening was definitely a highlight. My thesis topic, though not originally what I had in mind, has been progressing fairly well. I've been swimming almost every day this past week (and things look promising for this trend to continue) and have been feeling pretty good, if not exhausted, for it. I've felt more energized and had more "zest for life" than I've really had in a long while.

Life has just been good I guess. Liv is happy and enjoying her free time as well as her time at practicum. She has started planning a private clinic with a friend of hers and has been bouncing ideas around about some unique offerings. We're happy and relaxed and life seems to be generally headed in the right direction. After several months of agonizing over our future we've stopped worrying about it and started just focusing on what's happening now. Things aren't perfect (e.g., we could be in a better position financially) but what we know about and can control is under control and we've stopped worrying about the rest.

Life is good.

Well, I just wanted to drop a note here that I'm still alive and kicking. A lot has happened since October. For one, I've finished all my course work for my masters and now it's just a matter of doing my thesis. Also, my thesis topic has done a complete 180 for a variety of reasons that I might get into later.

Hopefully though, I'll be writing here more frequently.

It's been a while. Classes got started and have been going alright. I dropped down to only 2 courses. The graph theory course ended up with a final exam so I had to drop it. It's all good though, that way I'll be able to take the vision topics course next term. I've started TAing and that's been a load of work. I just finished marking the first assignment for the one course. Let me tell you, that was painful. I probably spent about 13 hours on it over this weekend. Marking over 100 assignments is about enough to make your brain melt.

The run was yesterday (Sunday). It went well. They raised something like $3.7 million. Thanks to everyone who donated.

Lets see. In terms of research things have been coming along fairly well. I wrote up a start at a problem statement cum thesis proposal. Here is an excerpt:

The basic question which I want to address is what 
level of model (in terms of activation and learning) is 
appropriate to develop and simulate the functionality 
of the visual cortex? That is, if a neural network is 
given realistic visual input (e.g. stereo video, 
modeled as ganglion cell receptive field responses) 
what rules are needed to govern the activity and 
plasticity of this network in order that it develops 
structure similar to that of the visual cortex. 
Further, what is an appropriate criterion to judge this 
and make comparisons?

The goal here is not to create the most realistic 
neuroanatomical model of the visual cortex, but rather, 
to see how simple of a model we can use while still 
getting realistic results. There are many reasons for 
this. For one, the most accurate models of neuroanatomy 
are largely intractable computationally if one wants to 
model a significantly sized network. This makes any 
practical application of this kind of network unlikely.

Secondly, the complexity of these models makes it hard 
to analyze mathematically precisely what is happening 
in the model. As a result, such a model doesn't not 
plausibly allow for further, rigorous analysis.

A further, more philosophical point, is that a model 
such as this is simply modeled at the wrong level if 
what we are interested in is network behavior instead 
of individual neuron behavior. We want a model which 
clearly defines how the activity of neurons are 
interrelated and how those interrelations change 
relative to input and other activity.

I wrote about a page worth like this. When I showed it to David his response was: "great...now, where's page two?" Which, I guess, shouldn't surprise me. It wasn't much beyond what we'd been discussing for a while so he's really seen all of this before. It was good for me to get it down on paper and to talk about it in more concrete terms.

At any rate, while I was writing that I was also working on the neural net simulator. For kicks I wanted to see how hard it would be to implement the Hodgkin-Huxley neuron model. Turns out it wasn't too hard, although finding reasonable values for the various constants was a challenge. That was kind of cool anyway.

Damn, I knew I should've sent the email later. I just heard back from the head TA and now he's fixed it so that I can post the grades. There goes my plans for the evening.

Alright, so if you feel like helping find a cure for breast cancer go here right now, search for me (Marcus Brubaker) and give me some money. Well, not me actually, but the Canadian Breast Cancer Foundation. That way I don't have to pay the $35 registration fee. :) Thanks, now back to your (ir)regularly scheduled blogging.

Well, I obviously missed a few days. The orientation ended well, although I'm going to be quite busy. This weekend I've tasked myself with learning Maya now that I have access to ti and I'm pretty far along that route. It is a BIG and complex program but, as with all pieces of software, has an underlying logic. It's just a matter of discovering that logic. I've skipped most of the parts on modelling as I'm don't really feel a need to create new models. Rather, I'll probably use freely available models and textures. The part that I'm focusing on is the animation and scripting. I need to do some simple things which will undoubtedly require scripting and expressions. Fortunately, at this moment in time, it looks like it should be even easier than I thought it might.

Anyway, it's been an exhausting week and things don't look likely to slow down any time soon. I even skipped the island picnic today because I just needed some time to catch my breath. I need to do things like cleanup the back patio, beg for money for CBCF run which is coming up soon and schedule things like swimming and indoor soccer as well as planning out my research goals. And thats just off the top of my head.

But that will all have to wait, as now it's sleep time.

So today I went to the area meetings for Graphics & HCI and Theory & Discrete Math. Both were interesting but veeeery different. The Graphics/HCI one was a few people (most of their faculty weren't there) who got up, said what their research area was, showed some pretty pictures/toys (as per usual with graphics people) and said what courses they were teaching. To boot, they finished an hour early, probably due to a lack of half of their faculty members. In comparison, there was the theory group which...well, was painful. They had all of their faculty there for one. Each of them spoke. Unfortunately, most of them don't seem to know how to talk to a group of people without lecturing. So what was the theory area meeting? Roughly speaking, it was the first lecture of an introductory course which surveyed the area. All but maybe two of the faculty members actually LECTURED, giving examples and mini-intros to the courses they were teaching and their research interests. It hurt....soooo bad. Plus, they went a bit over time. I mean, wow was that bad. Lots of interesting courses but the meeting was just dull.

On other notes today, I got some information about how I'm going to generate the test data I need for my thesis. Now I just need to get access to Maya and throw together a scene with some of the freely available models out there and I will be set. Of course, I need to learn Maya in there somewhere but that shouldn't be TOO painful. I hope. I reallllllly hope.

I also decided on the courses I'm taking (more or less). They are:

Introduction to Graph Theory - Graph theory...well, it's the theory of graphs. If you have a set of nodes and a set of edges which connect those nodes what do you have? Something like that anyway. There are loads of interesting applications and it's an area that I know little about from a theoretical standpoint even though I (and every other CS student worth his/her salt) know a fair amount about it from a algorithmic/applied standpoint. Should be fun.

Topics in Computer Graphics: Machine Learning and Computer Graphics - Probably the most topical course I'll be taking with respect to my own research. It's taught by a graphics person so it will have a different perspective than I'm used to. It's studying how to do things like learning how people (or things in general) move and look automatically instead of having to program a computer to do it by hand. Some neat stuff.

Topics in Machine Learning: Bayesian Methods - A slightly less topical course but still more topical than the graph theory course. It should be interesting, Bayesian methods are extremely powerful as a class of ML techniques. I'm not sure how much I'll use it in my own research but ML is always an interesting area.

Anyway, I'm also going to be auditing two courses for a bit but I'll talk about those more later.

Well, the first day of orientation is over and wow was it busy. Things started at 9am (*bleary eyed look*) with an overview of the department, the program and the student organizations that we're a part of. After that we met the "student mentors" which was quite funny since no one knew their mentor and about a dozen students and mentors were MIA so some people were just left milling around. After that was pictures and registration. That's when things got interesting. I got my office assignment, course listing and various other bits of paper work. My desk is one of six in a medium sized hole in the wall. Basically, I have a desk, a book shelf, a chair and the little square that this creates. It's a bit...dark but it'll do. For one, I'll have a laptop so I don't really need to be in my office if I don't want to. Plus, I'm hoping to get a desk in the Pratt building with most of the rest of the vision students. Or at least, the more senior ones.

This was the first half of the day. It's really hard to say what the last half of the day was to be honest, although it was really hectic. I met with my supervisor, David, to fill out the TA contract but had to run home and back in order to get some needed information. The reason for the urgency was that if the forms weren't returned today the odds were good that I wouldn't get paid this month which, to be honest, would SUCK. I met with Karen Reid to talk about TAing for her next term. It looks like I should be able to so that was nice. She recommended that I take a course by Professor Corneil so, since I was at a loss on what to take, I decided to look into it. Consequently, I met with his to discuss the course that he's teaching this term (a topics course in graph theory). He was just a spectacular guy and a really smart individual. His son is doing neuroscience research at Western so between that and just discussing computer science in general and graph theory in particular we had a really nice conversation. I'll probably audit his course while I take an introductory graph theory course since his course is going to be a bit...heavy for me.

Anyway, there was more but I'm too tired to write it up. After all that I went and made ribbons for the CBCF for 3 hours then helped clean and count for another 30 minutes. And, really, that about brings me to now. Oh, and Eric has been having trouble finding a job in Allentown so everyone send good employment vibes his way.

Alas, I guess a trip up north was not in the cards for me. I stayed home today to rest up and desperately try not to get sick. Liv went on ahead (no point of spoiling both of our fun). I expect that she'll be back in a couple hours. Fortunately, I'm not incapacitated or anything. I'm just feeling the precursors to full illness and I really can't afford to be sick this week.

It occurs to me that this might be easier if I just wrote short entries every day instead of these long ass updates every couple days. Of course, that would be the easy way, now wouldn't it? Anyway, lets see what's been up. Well, for one I'm getting sick again. This is not only annoying, it's also confusing. I almost never get sick and I figured the strep that I had last month should be enough for the next couple years. Well, maybe it travels in batches because I seem to be getting strep or something similar again. It started last night and has stuck around for today. I'm hoping it goes away. I REALLY can't be sick now, I just don't have the time.

Yesterday I went golfing with Rolf, Chris and Peter which was good. We played Centennial again which, while not perfect, is cheap. I had what was probably my best game. I got two pars (in a row no less), beat my best score on that course by 5 strokes and finished second. I didn't feel like I played that much better than before but the numbers say otherwise. Beyond that I didn't hit the stroke limit on any holes. (Got really close on one or two, but never hit it.) Chris and Peter are both great guys. I don't know them really well but it was a pleasant round with them, certainly moreso than with Uri. Nothing against Uri but he can grate on my nerves at times.

In the afternoon I went to get my new student ID. It's definitely a better picture than my old one. After that Liv and I did a small amount of shopping for dinner (Orzo with Roasted Vegetables) which was quite good and hung out for the night.

Thursday evening we went to see As You Like It at High Park. The show was really nice and watching Shakespeare outdoors is really the only way it should be done. They did a 1950's musical spin on the play which was quite neat, if not somewhat surprising at first. It was also our first time back in High Park in about a year. The last time was when we went last summer to see Twelfth Night or, What You Will. Before we went to see the play this year we walked down to a smaller grocer that we used to go to when we lived there.

Anyway, Thursday was also notable in that the initial TA assignments were posted. Somehow (I'm not really sure how or why) I've ended up TAing two courses for which I feel utterly unprepared. One is a 2nd year course on algorithmic complexity and data structures. I don't think it will be too bad. It is, after all, only a 2nd year course. But still, complexity is one area that I've always been a bit shaky on. I'm also TAing a 4th year graphics course which I only just completed myself four months ago. Graphics is not my strongest area so I'm somewhat concerned. If I had been assigned that I probably would've requested a different course but I was asked personally by David to TA his UTSC section so I figure it can't be too bad. And the UTSC part of it certainly adds a different perspective on things as I've never been to that campus. At any rate, it should be interesting.

Now, back to today. We got books for the class Liv is starting next weekend and then did a little bit of grocery shopping. I nearly passed out half way through the trip from heat and feeling a little sick but I managed to get through it. The rest of today I relaxed, watched some World Cup Hockey and fiddled with stuff on the computer. Of particular note is that I figured out how to burn VCDs. I also, soon after, realized that our DVD player doesn't play VCDs. Sigh. Ah well. I need to ask if Eric has one which does since, if he does, I would then have the perfect Christmas present for he and Amber.

Anyway, it's getting late. I need to go to sleep because tomorrow Liv and I are going with Nicole and some other people to a beach up in Barrie and will be leaving home at around 9am. That and I'm trying not to get sick, right?

It's going to be a loooong time until life "settles down." Liv and I have been talking about our plans for the next years and beyond and it may be a while until we're in a situation where things will be stable. I guess in a way that's a good thing. We really have the rest of our lives to settle in, we should enjoy the next 10 years of being unsettled. The current plan is something like this.

For the next two years we will be here, in Toronto. We'll finish up our masters. I'll do my thesis and Liv will finish her coursework, do her practicum and take her qualifying exams. Before her practicum (most likely at the Clarke) Liv will have to quit her job at the CBCF, likely sometime in April or May next year. During that time who knows how we'll afford rent but...it'll happen somehow. After that she'll finish her coursework and do her exams. Meanwhile, I'll probably be out of funding by around January 2006 at which point I will have to be done with my masters. That means I'll need to find some work until I start my Ph.D. in order for us to have things like food and a roof.

After that, most likely, Montreal. Liv is now looking at doing a Ph.D. in counselling psychology at McGill although a research masters/Ph.D. combo at Concordia hasn't been completely ruled out yet. I'll probably do my Ph.D. at McGill. By the time all of that is finished it will probably be around 2009 or 2010. And that's when the fun starts because, unless I get really lucky and land a position straight out, I'll start doing post docs and who knows where that will take me.

So, in a nut shell, that's our plan for the next few years. Of course, things will probably change but who knows. We have no idea how money will work out. I'm not sure if this is a sign, but I just got an email about a private sector opportunity for a vision researcher/engineer from Sven that I could do right now if I wanted....hmmmm. Liv thinks it's a sign, and I think she might be right.

While these things are good to think about, I need to start focusing on my thesis. I started looking at available software today. It may end up being necessary that I write my own as what is out there seems to be more complex than I need. Of course, I still haven't settled on what I'm doing. For one thing I need to decide what my goals are. More over, and more relevant to the software issue, is if I'm going to run a spiking neural network or to use a sigmoid activation function. The sigmoid function allows for easier analysis at the expense of less biological accuracy. Spiking neural networks are more realistic but harder to analyze. And if I go with a SNN I need to decide just how much anatomical detail I will model. Will it just be a thresholded spike? Will there be a refactory period? What will the threshold be? If it's not set how will it be learned or determined?

One thing I've been thinking about with respect to the threshold issue is to set the threshold based on the weights. That is determine an expected firing rate of an individual neuron (e.g. a neuron should fire 50% of the time over a period on average) and then, either by making assumptions about the input or by calculating the firing statistics of the inputs, set the threshold to have that expected firing rate. If we assume the inputs are iid (not always true but...) and we normalize the weights to sum-square to 1 and we want the probability of a neuron firing to be p then the threshold T should be p. (Assuming each input neuron has a similar probability of firing.) Now, things aren't quite so simple if we stop making such strong assumptions. I'm going to go think about it and write more when I get back...

Well, a lot has happened since I last wrote. To be honest, I'm not sure what has really happened but it's certainly felt like a blur. Wednesday or Thursday we started looking for tickets to Brazil. We weren't expecting trouble at all. The last two times we've gone we've both booked at least a month or two later than we're booking this year and have had little to no trouble finding flights with air miles or cheap flights otherwise. This year was different. I called Air Canada to check about a flight with air miles. Nothing. Absolutely nothing. No way to get from North America to Brazil in December or January with air miles. We figure that Air Canada, being bankrupt, has cut back on air miles seats and resign ourselves to having to pay for the ticket. Unfortunately, after calling half a dozen travel agents the cheapest thing we find is $1500 which is outrageous. Then, by freak chance, Roxana sees a place which advertises cheap flights to Brazil. We talk to them and find a ticket for $1085 which is fantastic.

By the time all of this has been sorted out it is Saturday so we walk to the agent to get the tickets. Throughout the walk and while we're getting the tickets Liv keeps saying "I love Nelson" and the like. Nelson, of course, being the travel agent. The poor guy, I don't think he knew how to respond. He was nice though. Anyway, on our way back we're hungry and decide to get some food. By chance we see a place that does all day breakfast and say what the hell. It turns out to be the most fantastic breakfast I've had in a while. Both plates came with garlic fries which were just amazing and it was all really reasonable. We're definitely going to be going there again.

Thursday we cleaned for no apparent reason. We organized our books and papers in the back room and just tried to make things look better.

Friday was Roxana's last full day in town before she heads to Montreal so she and Tatiana came over that night to hang out and spend the night (since Rox's apartment was in boxes and chaos). I wasn't certain they were coming over so I opened a bottle of Wolf Blass South Australia Merlot 2002 and did a bit of coding. Let me tell you, that is a spectacular merlot. It was fun (if not moderately unproductive) to sip at a nice red while hacking a bit. Anyway, eventually they came over and we hung out for a while. Rox passed out at around 1am or so but Liv, Tatiana and I stayed up chatting and finishing a bottle of Wolf Blass South Australia Shiraz 2001 until 5am. Sadly, I don't remember much of that one as I'd already had quite a bit of the merlot but what I do recall was quite pleasant. Not as smooth as the merlot but a little lighter. It was nice chatting with Tatiana. Liv and I are really looking forward to getting to know her and Adam (her boyfriend) better. They are likely to be good couple friends, although Tatiana is a little weird.

Now, back to Saturday. We woke up, went to get the tickets and then stopped to help finish Rox load the truck and see her off. It was a tearful event for both Rox and Liv. I hope everything makes it there alright, although I'm not entirely convinced. The one thing I definitely realized was that not everyone knows how to pack or move. I do. Rox doesn't. I really feel bad for the poor saps she recruited to help. They'd think they had everything and then she'd say, "oh, yeah, this mirror goes too." Or, "oh, and that plant, it has to come. I'll be so sad if it dies." Ah well. She got there and, as far as I know, there weren't any major disasters. Anyway, after she left we returned home (although we were up four plants, a frozen chicken, two tuna steaks and a marble pastry board which has been abused as a cutting board) hot, sweaty and with me getting a little sick from the heat and exertion. Of course, we had no time to relax, as Liv was to be at the studio preparing for her first belly dancing performance in an hour and a half. So we took a hurried shower and frantically gathered all the bits of costume that were around the house. I was feeling ill so I wasn't on top of things as I usually am and after getting on the subway we realized that neither of us knew where we were going. After much panicking we guessed and managed to find the place. The performance went great. You can see some pictures of it here. My Livvy was wonderful.

After that we got a bite to eat at Green Mango and we went to see Collateral which was alright. A bit better than I was expecting but still, more or less, an action for the sake of action flick. I don't know but I've been getting somewhat bored of action movies of late. Maybe I've seen too many or what but they're just dull these days. *shrug*

Sunday was a chilax day. We did a small amount of grocery shopping during the day, talked to my parents and just relaxed. In the evening I did some work and that was that.

Today I, hopefully, handed off the project that I've been doing for Sven off to an undergrad doing a project course in the fall. I wrote up a huge email talking about what's to be done, how to setup and build everything and that kind of thing. I'm hoping this will be, more or less, the last that I have to see of the project. After doing that I went out to Spring Rolls on Yonge (the one on Dundas, ironically enough) with Liv and some people from her work and then we (well, me and the three girls, the two other guys went to see Fahrenheit 9/11) went to see A Home at the End of the World which was just wonderful. I strongly recommend to to everyone. Honestly, compared with Collateral and a lot of the crap I've seen recently, it was a really nice change. It was a great story with wonderful characters played beautifully by some talented actors and actresses.

So that pretty much brings us up to now. Anyway, time to go fill out some loan paper work and have some coffee with Liv.

Oy, today hurt. Instead of holding my score steady I actually added 8 strokes over 18 holes. Sigh. I was just painfully inconsistent the entire day. One hole I would putt wonderfully, the next I would drive well but 5 putt it (or something similarly stupid) and then after that I would botch both. About the only bright spot is that I finally got par on a hole which felt nice. Mind you, it didn't last long, but for those brief moments after the ball dropped in the hole until my next drive I was pretty happy. It was, in my memory, the first time I ever made par on a hole this season and probably my first time ever.

Yesterday I met with David for a little bit. He gave me some good papers and we get a laptop ordered for me which is awesome. It should be coming in the next weeks. I'm getting a Dell Inspiron 8600 which is a pretty sweet laptop. Not the flashiest around but I think I'll be happy with it.

After that I scooted off to meet with Peter, an undergrad who may be taking over the project I'm working on for Sven. I think I may have scared him off a bit with the size and complexity of the project but I hope he takes it. I really want to be done with this thing.

Then, by complete chance and whim, I met with Rich Zemel a bit who is someone I've been wanting to talk to for a while now. I saw him in the hall and remembered that I meant to let him know that I'd like to take his neuroscience topics course if he offers it again. So I stopped in his office, let him know and (somewhat inevitably I guess) started talking about what I'm looking at doing for my thesis. It was a good conversation. He had several interesting points. Having to justify (or at least explain) my thesis ideas to a ML person was really helpful in elucidating some things for myself. He recommended a few papers and I'm going to read them and then go talk to him again. I'm looking forward to it.

Tomorrow I need to register and get those things going. I also need to apply for the Stafford loans so that I can repay CEFCU and close my account with them.

It's been a busy couple of days. Thursday night we went to see Sarah McLachlan at the Molson Amphetheatre. The show was just amazing. Her voice was perfect and she performed for around two hours. She talked a bit and just seemed really happy to be there. Plus, the venue was just wonderful. I love seeing shows there. It's an outdoor venue and we had lawn seats. We were a ways out but seeing shows outside at dusk is really a spectacular experience.

Last night we went out to see Garden State which was a nice movie. It was cute but somewhat demented. The perfect kind of movie. On the way back from the theatre Liv and I got into a long discussion on life in general. I think we've both been feeling a malaise recently and we're not really sure why or what to do about it. One of the conclusions that we came to is that we both haven't been appreciative enough of the things that we have gotten, instead always assuming that it wasn't significant and that we should do better.

The big example for us was school. When we got in to university neither of us made a big deal about it. It was like "of course, we were going to get in, it's nothing special." But we both have friends for whom getting into university was a huge deal. Their parents threw parties, their families showered them with presents and congratulations and they proudly told everyone they knew. Liv said something which pretty much summed the whole problem up: "When I would get marks I was never happy or excited to see an A but I was always depressed, upset or annoyed when I saw a B."

The thing is, we just need to be happier with what we have. We're both very fortunate and have managed to do a lot more than most people yet somehow we're not happy about it. The problem, of course, is how to do that. I'm not really sure.

Well, I played 18 with Rolf this morning. I played pretty well. I'd put what I shot but that would just be too embarassing. Suffice to say that I did alright but not exceptional. My drives were pretty good except that I had a cramp just below my right shoulder blade which was causing me to drop my shoulder at the top of my backswing and, consequently, screw up the shot. Once I realized what I was doing I was able to correct for it but it really started to show on the back 9 where I had some of my worst holes.

My irons were alright, not good...just alright but my putting is what made my score so bad. I'd be on in two on almost every hole (with the occasional 3 or 4) but then I'd 4 putt it or something ridiculous like that. Not sure what was wrong with me. I could never get it quite dead on. Always a bit to the right or a bit to the left and usually just a little too much power. Who knows. It was definitely the weakest part of my game.

Anyway, afterwards we sat down, had some burgers and then, when we were leaving, he gave me his sons old computer which, if it still is functional, will become our new MP3 server in the living room. It will be nice to be able to listen to music again.

Oh, and on the way there I found some exciting retinal models which take into account temporal dynamics as well as spatial dynamics. It was in a book I just got on interlibrary loan from the Mississauga campus library called The Neural Basis of Visual Function which is part of the Vision and Visual Dysfunction series which has been an excellent source of information in my research. Once I figure the model out in more detail I'll post more about it.

Oh, I forgot to mention that I met with Rolf on Sunday and spoke with Art today. They were my business partners. I was pretty certain that they would be ok with everything but after I hadn't heard back about the email after two weeks I was slightly concerned. Well, they're cool.

Rolf called on Sunday and we met in the afternoon. He and I chatted and he's getting back into tech things and may call on be for limited contract work, time permitting. So we're cool and he's going to try to make the business work, which is good. We're meeting Wednesday morning to play 18 holes, which should be nice.

Art called me today and he's been swamped with work. He seemed cool and he's digging more into Linux again. It sounds like he's still planning on quitting once he's back from Latvia, which is probably the right decision for him. We'll see though. He wants me to tutor/teach/whatever him sometime so that's cool. Might be some money too, you never know.

Anyway, bed time.