This summer, I went to the London School of Economics Summer School for an introductory course in data science and machine learning.
“What?” gushed my colleagues and classmates in unison.
Classmates because, well, I’m going to the LSE not to study economics. Colleagues because, as a home to a Google Deepmind lab and the founder of reinforcement learning, my home university could have afforded me better training in ML. Europe is seen as an also-run in the artificial intelligence race, barely keeping its breath behind America, Canada, and China. And of course, why machine learning in a social science institute?
But the decision is only natural. Firstly, machine learning is so popular in my university that they have to stack 3 nested prerequisites in front of it to defend it from overwhelming interest. My course plan is crammed full of economics, there is no chance that it can accommodate anything other than essential math. Secondly, I’d have only core economics courses in my final year. I wouldn’t want to have these core learnings from other institutions, where the reflections are cut short to only a few weeks. Thirdly, they are doing text analysis in the course! Behind this LSE course are the authors of quanteda, a text analysis package in R. No other standard ML offering will put 1/3 of instruction time in natural language processing. While text analysis is not used a lot in economics, an introductory background in this is very helpful.
Indeed, as it turned out, many of my coursemates are not in economics, but statistics, government (as political science is regarded here), junior engineering, math, and other sciences. All of my lecturers have their roots in political science, and nowhere in course materials mentioned economic theories (albeit some econometrics). During my time at the LSE, I busied myself with this course and another on alternative investments, but did no formal economics in general.
I’d later realise that my most important takeaways are still mainly economics. (I should get a life.) But that aside, I learnt to not be presumptuous about the workings of the society; but to observe and adapt.
Some background of what led to this:
University learning
The course put us through 4 different lecturers in 3 weeks. We began with a week with composite materials, a week based on a published textbook and a textbook under public review, then the rest based on lecture notes and a research paper. We were also required to arrive having learnt the core language and concepts by ourselves. It is interesting seeing the shape of learnings under these arrangements.
University culture
I never thought I’d experience culture shock in the United Kingdom. As Robbie Williams once said to an American audience: once you get over the language barrier, it’d be just fine.
Well he lied. I certainly had a lot of trouble adapting to university culture here. The “social science vibe” was palpable since orientation day and I still don’t know how to fit in. At University of Alberta, no one cares about correct word use, let alone choosing words and tones that amplify one’s presence. One always has to sift through a number of inaccurate terms to take jibe at what a fellow really wishes to say. Now accustomed to these monotonic mumblings, the eloquence in LSE is startling to me. Every morning before class, there is a master seminar on the art of flirting (then silence for 3 hours after lecture starts;) everyone seemed so gregarious and so purposeful. “There’s breeze even in their footsteps,” as I told my mother over the phone.
The university social scene is vibrant and full of beer burps. People practically spill over from the campus bar to the next street on weekday afternoons, and you can even clink (plastic) beer glasses with your own teaching assistants. This is somewhat illicit back in Canada, something my cheery TAs Tom, Selina and Sarah wholeheartedly condemn. It turned out that many of my learnings come from talking to these current LSE scholars……including the advice to stay in North America for my research direction. If I had stuck to my old way, I would have had a bad time.
University assessment
I had the biggest trouble with this. I knew that grading is different in Europe, that a third or more students graduate with a first class honours from a UK university, but experiencing it was still a shock. For example, we were (implicitly) allowed to discuss and work with classmates for our graded assignments. I found out about this after the midterm, after I nearly fell off the “good” grading band. At that time, final assignment/exam was just released, and I heard classmates arranging that they’d work on one question each and pool efforts later. It took me a whole day and some clarifications to believe that it is not academic dishonesty.
Truly, I have believed for all these while that everyone sticks to a “in your head” rule — largely a practice used in computer classes in the U of A, where we can talk about problems with classmates, but if the discussions involve any jotting down things like ideas or equations, it is considered cheating. No way we could work on assignments together, let alone final exam.
Colluding is evidently not a big matter here. Viva la freedom?
I later did share my hiccups with friends, and learnt so much from those discussions. Though my stubborn head remained indignant about sharing work efforts on graded assessments till the end, I’m glad to have encountered this juncture. It is easy to afford a few more days of work as my classmates finished bright and early to roam Europe…. but much more stand to be lost in the future, if I had not been shaken awake from my warped reality.
In short, LSE taught an economics student that society has rules, and they evolve. Observe. How shameful to say, but what an awakening! I think this alone is worth the trip.
Written by Natasha. Last edited:2019-08-22 19:04:43