Tuesday, October 03, 2006

$1M is a lot of movies

Netflix is giving up to $1 million to whoever can write the best movie recommendation system. And if you do it using NLP techniques, Hal Daume will sub you an extra $10. Get your statistical machine learning on!

Sunday, October 01, 2006

Random News

Some of the events of last week:

  • Auguste Anberree was born, six weeks early. He is fine, but will be staying in hospital for observation.
  • I did much work on the robotics course which starts next week.
  • I did some experiments and got some encouraging results.
  • I had an eye test and discovered I have mild farsightedness, and moderate astigmatism. I thought I had shortsightedness, as distant objects appear blurry. Turns out that is the astigmatism.

Monday, September 25, 2006

La France

Zut alors! The blogging just can't keep up with life. Now several weeks ago I went to France with Messieurs Anberee, Gurnell, Swinbank, and the lovely Mademoiselle Pounds-Cornish. Photos are online here and here.

I've only been to France twice before, when Bree and I went to Paris, and when we went to Lille. The trip to Paris was the first trip we made after moving to the UK. We hadn't yet started working so we were mostly broke. We walked all over the city, and subsisted on picnics of cheese and bread. I remember we had left over cheese one night, which we wanted to keep, but it smelled so bad we had to keep it on the window sill outside our hotel room. Good times! This time I was in small town France, first with Thomas' brother in Ile de Yeu, and then at Richards Dad's holiday home near Niort. Quite a different experience. I won't go into what we got up to. Take a look at the photos if you want to know. A few points of difference though: the French care much less about gardens and the appearance of their houses. It is normal for the rendering of a house to be falling off. Food is cheaper and better quality. The stereotype really is true! I'd like to know why that is. France is so close that I should think the same produce would be available.

Tuesday, September 19, 2006

The Grad Student's Pain: TeX

Ezra's things I learned about TeX today resonates strongly with me. In many ways (particularly it's layout algorithms) TeX is a thing of beauty, but in terms of interface it is a complete pile of junk. Most incredible is that after some 20 years no-one has done better.

Thursday, July 27, 2006

Why is the UK so expensive?

Why is the UK so expensive? The ripple down effect of high property prices seems to be the consensus opinion. The UK town planners have made a conscious decision to avoid sprawl, at the cost of raising the price of land. I can't complain, as the higher population density is one of the reasons I live in this country instead of Australia

Tuesday, July 25, 2006

Thursday, July 20, 2006

Gettting Dumber

Brain size has decreased by 10% in the last 30'000 years. Oh dear. The most common theory is a relationship to diet.

Monday, April 24, 2006

CS and Religion

From a student project:

The best way to design a system...should include...Polytheism

You read it here first!

Saturday, April 15, 2006

Stop, Easter Time!

Today being part of the Easter holidays Bree and I did what is traditional, and worked on the garden. The plan is to reshape the lawn. Apparently this plan has been in gestation since we moved in, but as my gardening role is to provide manual labour, brute force, and ignorance, I was, of course, largely ignorant of this plan. Anyway, I dug up lawn, dug up garden beds, shifted lawn into garden beds etc. I am told It Is Good.

Earlier in the day we kicked out Matt who'd broken his nightmare Glasgow to Canterbury train journey with a couple of nights in Brum. Much geeking was done, and at one stage we had 42 inches of Powerbook in the house! A side-effect of kicking out Matt was I went in to town and shopped at the markets for the first time since, well, before going to Australia. And the result of that is:

Basil and Black Pepper Stir Fry

This is inspired by a meal I had at Han's Cafe in Perth

Ingredients for the stir-fry

  • Some beefy protein (e.g. beef, seitan)
  • A carrot
  • A red capsicum
  • An onion
  • 2 cloves of garlic

Ingredients for the sauce

  • About 12 leaves Thai (sweet) basil
  • 1 tbsp cornflour
  • 1 tbsp soy sauce
  • Dried mushrooms
  • Cracked black pepper


Get your rice on! Works best with rice, but noodles would work as well.
Cut up everything.
Mix the cornflour to a paste with some water
Rehydrate the mushrooms in boiling water. When the water has cooled, add the cornflour, soy and basil. If using seitan, add some of the juice from the can.
Get your wok on!
Cook in the usual way: meat (if using) first, then set aside. Then onion, garlic, carrot, capsicum.
Add your protein.
Add the sauce, boil till thickened.
Get your snack on!

Tuesday, February 28, 2006

The Economics of War

This week's File on Four covers the war in Afghanistan. Our stereotype of the Taleban tends towards the religous fanatic. An interview with a Taleban fighter reveals a different story:

“I'm fighting with a Kalashnikov and an RPG - a rocket propelled grenade launcher,” he said. “I'm not trying to take over the country. I am just trying to earn my salary.”

Compared to his previous work in the coal mines, fighting is easy and the pay is much better. He goes on to explain the Taleban even offer bonuses for killing American soldiers. An interview with an opium farmer reveals a similar economic motivations. He'd happily grow vegetables, but the market for opium is just so much better. This highlights the core economic problems behind the current problems in Afghanistan (and many other places in the world). Instead of the Wars on Terror and Drugs, perhaps Dubya should launch a War on Farm Subsidies. It ought to play well at home given that France is one of the main proponents of the EU's Common Agricultural Policy.

Thursday, February 23, 2006

Control of creativity

For a few years now the media companies have been going nuts about copyright protection. This article compares the differences in control over creativity in the fashion and the media. Why is it that Sony, for example, goes nuts if you sample a few seconds from one of their artist's tracks, but if you copy someone's clothing designs (as Topshop, H&M, etc. routinely do) you get, if not applause, at least no negative repercussions. I don't have a good answer, but I know which model I prefer.

Friday, February 10, 2006

Winter and Tech Reports

I walked in today along the canal. It's a clear cold day (-3C right now), a relief after all the dreary overcast days of late. The canal had frozen and I could clearly see a hexagonal structure in the ice, like enormous frozen snowflakes.

I've spent most of this week working on a techinical report. I'm glad it's finished and I feel good about the work. Doing the analysis, and getting statistically significant results, was nice (when I reviewed other work for the Related Work section I was rather surprised how few authors do any proper analysis of their results). I hope to turn it into a publication soon.

Wednesday, January 25, 2006

A Doctor is made

My mate Dave has just got out of his viva, with minor modifications. And that, folks, is what it is all about.

Thursday, January 05, 2006

The Failure of Research

There are is no Journal of Failures, an unfortunate state of affairs as what doesn't work can provide just as much insight as what does. Of course, I have a personal interest in salvaging knowledge from failure, as at this moment my simulated annealing process is failing to converge (example results, till I delete them). I was hoping to make ICML'06 but it is unlikely I'll have results by then. Well, that's the way it goes. Research is a high risk activity. If it wasn't, it wouldn't be research! A lot of the structure of our research institutions are designed to reduce the risk of failure. Even if your experiments don't work, you'll still get paid. You might not get a promotion, but that is preferrable to being destitute. This structure has been very successful. Thousands of people now spend their careers as researchers, compared to handfuls in previous centuries. The resultant benefits to society have been vast. All up it's a pretty good deal. However, decreasing risks necessarily involves also decreasing potential gains. I wonder if Universities lose out on the best and brightest due to this policy?