Mailund on the Internet

If only I could teach like this, then I wouldn’t feel so bad about my teaching career.

It seems that Google is finally getting around to synchronising Google Calendar with other tools (or at least Outlook, for now).

I have missed this feature for ages. I hacked my way around it twice on windows, but on Linux and OS X I can still only synchronise one way, so I can see my Google Calendar on my desktop, but I cannot add entries except through the web interface.

I hope this means we will get an API to interact with Google Calendar, and that synchronisation will eventually be possible on Mac and Linux as well.

A friend sent me the link for this video:

I must say, I loved it.

This week I finished a class on Teaching at the University Level. This is a class that is mandatory for junior researchers here at Uni Aarhus, if they have any plans on moving onto a tenure track. This is probably a good thing. We are expected to teach at least 20% of our time (at least at the faculty of science, I don’t know the rules elsewhere), and there is no formal introduction to teaching except for this class.

It is not a major effort to take this class. It consists of one week in the summer, away at the University’s mansion Sandbjerg, then some group work with coaching and peer-evaluation of our teaching, and two follow-up classes. The last of those were Tuesday this week.

I was very sceptical about this class. I hate long group discussions about this and that and the course description sounded like that was exactly what I had to look forward to.

It was, and I hated it.

Still, I actually learned a lot from just being forced to think about my teaching. In that way, the goal justifies the group discussions.

I hate teaching

Now the only problem is, that learning how to teach properly has taken all the joy out of it. I’ll explain…

Teaching vs lecturing

This is how it usually works. You take an auditorium and fill it up with students. Then you put a lecturer in front of the students and let him talk for an hour or two. After that, the session is over and everyone can go home.

The keen students might have read ahead in the text book and have a question or two that the lecture clears up, but 80% of the students haven’t bothered and are just attending the lecture because it is expected of them. If you are really lucky they are listening to the lecture, but not really thinking much about it. They will learn the material from the text book just before the exam.

We are wasting our time

The lecture is basically a waste of time for everyone involved.

The few questions that the lecture might clear up could be better addressed by actually talking about those specific questions. It would probably take less time, and it would be focused on the essential problems.

We are wasting our time. Why?

I think we are just doing it out of tradition. Before we had the wealth of textbooks we currently use, we needed the lectures. We had no other source for the material. You followed the lecture, took plenty of notes, and your notes were all you had.

I’ve actually had a class like that, but those are few and far between these days, and not something I would recommend.

If you have a textbook, you have all the material there. Read it. Repeating it all at a lecture is a complete waste of time. If you want to teach, why not spend the time more sensibly? When are you teaching and when are you just lecturing (i.e. wasting everyone’s time)?

You want your students thinking about the material. You want them active. In the teaching class I just finished they called this deep learning (contrasted with surface learning), but really it is all just about activating the students.

Students are, just like the rest of us, basically lazy. Anything they consider dull, they will try to avoid. Who can blame them? We would do exactly the same.

Two options

It seems to me that you have two options, then: 1) you motivate them to a degree where everything is exciting and nothing is dull, or 2) you force them to do the dull work so they are forced to be active.

You’d like to go through door number one, but that is not really realistic. I don’t believe that anything can be all excitement. There will always be some boring parts, even if you try to minimise them. They are part of what you are teaching, and if you avoid them or ignore them, you are simply lying to your students.

Let me make this concrete with an example.

I teach Machine Learning at the Department of Computer Science. Now, in Aarhus at least, computer science students hate statistics. There is a proud tradition of hating statistics, and my students are very conservative when it comes to this tradition. The problem is, that machine learning is essentially applied statistics.

This gives me a choice: I can leave out the statistics and just teach them the various algorithms in machine learning, or I can drag them through material that they really, I mean really, do not want to learn.

If I leave out the statistics, they will learn some recipes for solving specific problems, but they will never understand why they work or be able to extrapolate. I have to include the statistics. I need to take option two and force them to learn statistics.

Forced teaching

How do you apply force in teaching, then?

You can yell at the students all you want, but that has no effect. You are (sadly) not allowed to beat them. This leaves only exercises and projects.

If the students are required to complete a number of projects before they can take the exam, then there is no way to avoid the dull material. No honest way, in any case, but let’s ignore actual cheating here.

So when are you teaching, and not only lecturing? When the students are working on projects. This is where they learn. The lectures are just where they are wasting hours and getting older.

The good news is that option two — forcing some work on your students — actually motivates them to learn more than just the minimum required. You are getting some of option one for free, by choosing option two. Once they get excited about a project, they do much more than you could force them to.

Teaching vs popularity

After each class, the university makes a student evaluation of the class. The way the questions are phrased, it quickly turns into a popularity contest instead of a proper evaluation. At least, all the quantitative measures in the evaluation turn out that way — the qualitative measures, the student comments, can be quite useful.

The problem with making the evaluation a popularity contest is, of course, that good teaching and popular teaching is not necessarily the same.

If you want to be a popular teacher, you teach easy classes and you are a bit of a show-man during lectures. You will get excellent evaluations, but you won’t necessarily be a good teacher.

If the students have fun and feel that they are learning, they will love the class. But having fun and feeling you learning something is secondary to actually learning something. We are only evaluating the first, not the second.

Don’t get me wrong. I don’t want teaching to be boring. I think it should be fun, and I think it can be fun. It is just not the main objective to have fun while learning, and this ties in with the necessary evils you sometimes have to go through before you can start having fun. Like statistics in machine learning.

The evaluation for machine learning is very clear about this. The students hate statistics and want it out of the class. They love the rest of the class. They have fun with it; they have fun with the projects; they feel they learn a lot. They just don’t like the statistics. They think it is too hard.

Sorry guys. The parts you have fun with are only possible because of the parts you hate.

Challenges vs perceived challenges

Students (and people in general) like to feel challenged, but not actually to be challenged. We all love the rush, the thrill, when we solve a difficult problem. We likewise hate struggling with a difficult problem we cannot quite get a hold on.

Try this: at a lecture, ask a difficult question, then spend the rest of the lecture solving it, using all the cleverness you can muster, but in steps small enough that everyone can follow them. At the end of the lecture, everyone will feel that they have understood a very difficult problem. They have a war story to tell.

But there is no actual struggle with the problem. We learn from the struggle, not the rush of solving a problem. It has just been a waste of time.

I have tried this a lot (but I am wisening up to it now). I explain a problem during the lecture and go through all the steps needed to solve the problem. Everyone nods and agree that they have understood the problem and the solution. Then I ask them to solve very similar problems, and suddenly no one have any clue about how to approach the problem.

This is frustrating for everyone.

Lying through PowerPoint

A related issue is the use of PowerPoint (Impress, Keynote, whatever) slides. When I started at the university, lectures would be on blackboards and possibly using a few overheads. Now PowerPoint is ubiquitous. I think this is a problem, and I don’t think I am just being conservative here.

It is so easy to hide the true complexity of the material in colourful PowerPoint slides.

PowerPoint is a very powerful tool for visualising complex issues, but this power is very tricky. Some problems are just very difficult to understand, and clever visualisation is not going to change that. Instead, there is a risk that it hides the complexity.

There are times where you want to hide complexity, and there are times when you do not. Hiding the complexity of a problem in a lecture is just dishonest. The complexity is there, whether you hide it or not, and students need to be aware of it.

You might not require that they understand it, but they really need to know that it is there.

Of course, you could also hide complexity before PowerPoint, but now you end up doing it without even thinking about it. At least, I do, from time to time…

Optimising evaluations

We are all, as humans, pretty good at optimising whatever objective we are being measured against.

If the teaching evaluations are popularity contests — and they have any influence on our chances of getting tenure at all, which they currently do not (I am not sure why we make these evaluations in the first place) — then we will optimise towards that.

To get good evaluations, all you have to do is this: Explain difficult, but interesting, problems with lots and lots of colourful power point slides. Leave out the most difficult parts (perhaps with a reference that the students will never look up, but which makes them feel that they could figure it all out if they tried). Never, ever, ask questions that tests if the students actually have learned anything from these lectures. The point is to make them feel good, and exposing their lack of knowledge (and learning) runs counter to this.

Why do I hate teaching now?

After taking this teaching class — and thinking more about my teaching over the last six months — I realise that I am doing everything wrong!

I have just been lecturing, essentially repeating the kind of teaching I went through myself. I’m giving exercises and mandatory projects, but again just because I went through the same as a student, not because I thought much about how to integrate this into my teaching.

Now that I’ve realised that I am wasting everyone’s time with lecturing, I want to do something else. I’ve made some experiments trying to activate my students, but it feels so futile.

My students really prefer the old lecturing style. They are used to it, and do not see the point in changing it. When I try to activate them, it feels like I’m fighting them all the way.

The only thing I’ve gained from my efforts so far, is poorer evaluations…

Regaining my love of teaching

Something has got to change. I cannot continue teaching now that I hate what I am doing.

I just have no idea about what to do. I need some substitute for plain old lecturing. I’m thinking more project oriented classes, but I am not sure how to go about doing that.

This teaching term is ending next week, and the next term is just around the corner, so I guess I need to think quick, now.

Neil Saunders asks: Can every workflow be automate?

Workflows is something I’ve been thinking about myself, especially in the context of grid computing.

My “grid computing” collaborators are working on ways of running workflows on grid resources. This is a good idea, but I am worried about figuring out the workflows in the first place.

Quoting Saunders:

To me a workflow is rather like a scientific paper: an artificial summary of your work that you put together at the end, describing an imaginary path from starting point to destination that you couldn’t know you were going to follow when you set out. Useful for others who want to follow the same path, less so for the person blazing the trail.

I agree completely on this.

I spend much more time on figuring out how to analyse my data, then I ever spend on the actual data analysis.

Of course, I am still working with workflows when I am doing this, but I am fiddling with it all the time, and I have to go in and look at intermediate results in each step in the workflow to make sure everything is running the way it is supposed to.

Giving me tools to efficiently run finished workflows is not going to help me much. Better tools for experimenting with workflows, on the other hand, would win you a beer from me.