As a kid, you are told that scientists follow the scientific method. They come up with an hypothesis, and they try to falsify it. You also learn about engineers who solve practical problems using science. Later, you learn about induction and deduction. Some researchers collect evidence and derive general rules (induction); other researchers start from the laws of nature and compute models to explain specific phenomena (deduction). Both of these dichotomies (science/engineering, induction/deduction) fall short of classifying the researchers I know.

Apparently, I am not alone. According to Julian Togelius, Machine Learning is neither science nor engineering:

Most of what I do is pretty far from being useful or even reliable. Instead I think of myself as an inventor, practicing blue-sky invention of algorithms and toy applications without direct economic pressure. (Julian Togelius)

There are many research models and no satisfying taxonomy. Nevertheless, I made one up:

• The specialist: he has invented or perfected one method.
• The perfectionnist: he improves existing methods.
• The practioner: he solves real problems.
• The algorithm designer: he invents a new and better algorithm.
• The system designer: he builds new prototypes.
• The hardcore theoretician: he solves hard mathematical problems, for their own sake.
• The practical theoretician: he picks up on what is found useful and analyzes it.
• The advocate: he promotes a particular technology.
• The politician: he networks.
• The educator: he writes readable research papers.
• The librarian: he writes reviews, citing 150 papers.

What do you think? Which researcher are you?

I am a practical theoretician. When I come up with a new algorithm, it is almost always by accident. I also do not care to solve problems for their own sake. I also hate to build systems for their own sake. I am not even interested in coming up with something new! But I am interested in understanding why things work. I build prototypes, I study them, try to find interesting and strange patterns.

A colleague of mine, Sébastien Paquet, wrote his Computer Science Ph.D. thesis on the benefits of social software for scientists. Sébastien determined through self-experimentation that using social software could improve substantially your network as a scientist. After years as a blogger, I can testify that blogging did improve substantially my network. I also believe it made me a better writer.

• Better networking: My list of readers shows that I am read by a wide range of people, from famous scientists to entrepreneurs. Considering that, as a scientist, I rarely attend conferences or participate in large-scale projects, I am vastly more connected than I should be.
• Better writing: When I write research papers or lecture notes, writing well is important, even essential, but the focus is on the scientific results. I find that blogging forces me to focus on writing better. Not to mention that some people criticize my writing publicly: that is a strong incentive to get better!

Secondary benefits include better knowledge management: often I can’t remember something, but I remember writing about it on my blog! However, I find that I am spending more and more time on sites such as facebook, friendfeed and twitter. I am slightly worried that these sites may have lesser benefits for scientists.

• No better writing: Most of the social software sites require little prose. A lot of the interaction involves no writing at all! And the writting is simply not sophisticated enough to require much effort.
• No better networking: It is much harder to build long-lasting content on these social sites. Many of the 1000+ subscribers to my blog subscribed because they found at least one blog post interesting. On facebook, I am only a stream of slightly incoherent bits of news.

Conclusion: I will not give up on this blog in the near future, no matter how good twitter or facebook get.

Reference: Aimeur, E. Brassard, G. Paquet, S., Personal knowledge publishing: fostering interdisciplinary communication, 2005.

Universities, research centers and grant agencies often make the implicit assumption that research productivity depends on the innate abilities of a few key individuals. These people are research heroes. By putting these people in charge and giving them all resources, research productivity is thus maximized.

In Research productivity: some paths less travelled, Martin takes a contrarian viewpoint on research productivity. He believes that given the right nurturing, many people can become highly productive researchers. Research excellence comes by relentlessly training yourself in the right conditions. With this model, it could far more economical to grow many different highly productive researchers (see Research productivity versus funding received). I suggest that Ph.D. programs should have explicit training on productivity!

The following table presents the difference between the traditional viewpoint and the proposed alternative:

Further reading:

• The Heroic Theory of Scientific Development by Peter Turney
• David Dobbs, How to be a genius, New Scientist, 2006.
• Geoff Colvin, Why talent is overrated, Fortune, 2008
• Thinking intelligence is innate makes you stupid
• Do you think because you write, or write because you think?

I love gardening. I love research.

These activities are analogous:

Life is multidimensional. Research papers should be multidimensional too! We should ask several interesting questions. We should give several nuanced answers. We should expect more from the reviewers and the readers!

Yet, in Computer Science, research papers are expected to tell the following story:

• We consider problem X;
• other people solved problem X with solution Y;
• we propose solution Z;
• we show that solution Z is better than solution Y.

It makes the job of the reviewer easy: (1) the contribution is clear (2) we can quickly quantify the value of the contribution. The more competitive the conference or journal, the more often we see this Z-is-better-than-Y story.

I submit to you that these research papers are the equivalent to the movies Hollywood  producers like so much:

• Bad guy creates a problem;
• good guy comes in, beats the bad guy, fixes the problem;
• good guy gets the girl.

Similar B movies can be repeatedly produced without any new insight. These movies are easy to follow. They are also quickly forgotten.