Daniel Lemire's blog

, 7 min read

Lack of steady trajectories and failure

10 thoughts on “Lack of steady trajectories and failure”

  1. The expression “you are the leading authority on one or two topics” implies that the topic must be narrow enough. Nobody is the “leading authority on Computer Science” or the “leading authority on Information Retrieval”. You might be “the leading authority on the application of Bayes Networks to ranking problems in Information Retrieval”.

    Presumably, the granularity level depends on the popularity of the field. If you are one of the last two remaining Physicists in the world, then maybe you could be the leading authority on Quantum Mechanics.

    For very popular fields, you need to be very narrow to be an authority.

  2. jeremy says:

    What’s your level of granularity on topics?

  3. @jeremy It is awfully difficult to define “diversity”. I know, I tried.

    There is nothing wrong with doing Information Retrieval for 30 years.

    What I’m saying is that if you can both contribute something to Information Retrieval, and also to a different field, you might have a better appreciation of both fields.

    Can you do Philosophy and AI? Medicine and Psychology? Physics and Algorithms? I think that true scholars can contribute to multiple fields, not just multiple topics.

  4. Anam says:

    Some topic -> more specific of the topic -> “narrow” topic. When you go along the way, how would you filter out topics in second stage, which not necessarily is a single step?

  5. jeremy says:

    My question was more along the lines of encouraging non-static research topics. What does it mean to be diverse?

    For example, I’ve gone fairly deep in both text (recall-oriented) information retrieval, as well as music information retrieval. Not to mention explicitly collaborative information retrieval. On the one hand, it’s not diverse at all: It’s still just information retrieval. On the other hand, each facet involves some fairly specialized knowledge.

    So are all my eggs in one basket (IR)? Or in multiple baskets (text IR, music IR, collaborative IR)? And how does that relate to others that I know who have their eggs both in the physics and text IR baskets — is that more diverse or less diverse?

  6. Kevembuangga says:

    Do you mean that since there are no “heroes” in Science anybody can match Von Neumann?

  7. Srikanta says:

    Excellent idea – but it seems to me in practice diversifying one’s research is possible only for tenured faculty like yourself!

    I don’t see how a young researcher (untenured yet) can afford to diversify, at the same time keep up with the need to publish, teach, obtain grants, serve on committees, advise graduate students, do administrative duties,…etc.

    I am not complaining about the ideal – it is indeed something I would aspire for, but I am questioning the practicality of it.

  8. @Kevembuangga von Neumann was probably a workaholic. He worked continuously. And he started his research career quite young. Also, he got access to extraordinary ressources. So it may not be practical for most people to match von Neumann.

    @Srikanta I think that the incentive to stick with what you know may become even stronger as you advance in your career. But I don’t deny that diversifying is costly.

    I suspect, however, that a diversified research expertise may offer more opportunities later on. But I’m not sure.

  9. @jeremy Yes, collaboration is a good way to reach out to other domains, but it has its own overhead.

  10. jeremy says:

    @Srikanta @Daniel How about synthesis research? Achieve multidisciplinarity through collaboration?