• Posted by Konstantin 19.12.2008

    The day before I've accidentally stumbled upon an old textbook on pattern analysis written in Russian (the second edition of a book originally published in 1977, which is more-or-less the time of the classics). A brief review of its contents was enormously enlightening.

    It was both fun and sad to see how smallishly incremental the progress in pattern analysis has been for the last 30 years. If I wasn't told that the book was first published in 1977, I wouldn't be able to tell it from any contemporary textbook. I'm not complaining that the general approaches and techniques haven't changed much, these shouldn't have. What disturbs me is that the vision of the future 30 years ago was not significantly different from what we have today.

    Pattern recognition systems nowadays are getting more and more widespread and it is difficult to name a scientific field or an area of industry where these are not used or won't be used in the nearest future...

    Further on the text briefly describes the application areas for pattern analysis that range from medicine to agriculture to "intellectual fifth-generation computing machines" and robots that were supposed to be here somewhere around nineties already. And although machines did get somewhat more intelligent, we have clearly failed our past expectations. Our current vision of the future is not significantly different from the one we had 30 years ago. It has probably become somewhat more modest, in fact.

    Interesing, is this situation specific to pattern analysis or is it like that in most areas of computer science?

    Posted by Konstantin @ 12:01 pm

    Tags: , ,


    1. Ando on 14.01.2009 at 16:36 (Reply)

      I don't think this is specific to pattern analysis. Obviously this has been very pronounced in AI (the "AI winter). Also, I've found the same to be true in semantics. Progress has been barely incremental, and its impact on practical computing has been rather marginal. C is still one of the most popular languages (35+ years after it was developed), with popular "new" languages being mostly syntactic sugar on the "old" ones. Things such as program verification are pretty much as far from mainstream as they were 30 years ago. Of course there have been improvements, but I think most of the goals set 30 years ago are as far away now as they were back then. If anything, predictions are more conservative.

      1. Konstantin on 15.01.2009 at 17:36 (Reply)

        Well, at least the entertainment industry is doing better. It certainly lives up to, if not even exceeds the expectations that were put on it 30 years ago. Note that this somewhat motivates even stronger expectations for the future. I, personally, am pretty sure we shall all be playing pure virtual reality or something like that in 20 years already. This leads to a (nearly obvious) observation that we, humans, tend to base our expectations on the linearized trend of the near past. AI progressed a whole lot in the 70-s hence everyone was waiting for it to skyrocket further. Seems logical.

        Another observation is that most of the progress in the game industry should actually be attributed to better hardware and a lot of hard-working coders and graphical designers. What concerns the more "theoretical" research, it is as incremental as everywhere else. Now if we generalize and look into science as a whole, it seems to me that most "progressing" areas (biotechnology & medicine or material physics, say) are always more about hard manual work rather than deep theoretical braining.

        Or, thirdly, maybe it's the other way around with money on the top:
        Commercial » Good theoretical progress » Lots of manual work » Visible progress?

    2. swen on 19.01.2009 at 12:29 (Reply)

      This is a natural consequence of two things.

      1. There are relatively few things in the world with low Kolmogorov complexity, that is manageable by a single aspired individual. As a result, there will be fewer and fewer radical breakthroughs in any discipline that has grown beyond infancy.

      2. Science as a whole is a commerical project instead of hobby project. Since the amount of work is so big to achieve new things, science must get a constant income. For that, a good manager must show constant progress and therefore all near-trivial things must be published. Only a few people can act as Plotkin and drop the whole research unless there has been a radical breakthrough.

      More specific to your initial complaint. I honestly think that all sub-blocks that generate "intelligent" (read humanlike) behaviour must be simple enough or they have not emerged in evolution. That is "intelligence" is a truckload of boringly simple things connected together and the "bling" emerges on the system with so high Kolmogorov complexity that we cannot grasp.

    Leave a comment

    Please note: Comment moderation is enabled and may delay your comment. There is no need to resubmit your comment.