|
According to the Cambridge dictionary, to compute means to calculate an answer or amount by using a machine. Let us consider the following three related questions:
|
|
- Can nature be understood as a computer?
- If so, what is the aim of its computations?
- What are the methods and algorithms being applied?
Nature, here understood as the whole universe, is no doubt dynamic in the sense that
it continuously changes with time. Large rocks are exposed to weather and reduced to sand; the waves in the
ocean beat incessantly; and people keep accessing the WWW. While we have no hint on why nature exhists or
changes, each small transformation of physical entities seems to be governed by a simple principle, that of
parsimony
. Indeed, changes in the universe are performed in order to
achieve maximum economy of energy, an idea which is nicely expressed by the mathematical concept of variational
calculus. The film in a soap bubble, for instance, is the solution of a minimization process regarding
gravity and surface tension. Many principles in physics, from classical to quantum mechanics, are effectively
formulated in terms of such an interesting theory. In such a sense, nature seems to be a huge variational
computer, except that we can not pull its plug out (nor we should).
|
|
But, if nature is indeed a computer, what are the answers being sought? Although we have no access to the final configuration of the universe, we can gather a good idea of what nature is computing in the same way as we can look at the execution of a program: by analysing its intermediate results. In this sense, nature has been computing stars, planets, continents, plants, animals, and, yes, you bet: humans and you. As a matter of fact, and with little poetry left, we can think of ourselves as the intermediate result of nature computations, and a very special one in the sense that we have been the first (arguably) to realize that. And we have also created machines capable of enhancing computations in our world.
|
|
Thus, regarding the abouve questions (a) and (b), the universe seems to be a huge variational computer of which the we are intermediate results. Ballard's new book, titled "An Introduction to Natural Computation", is aimed at providing an introductory overview of answers to question (c), namely some of the main techniques possibly underlying natural computation. The reason why we can not be limited to variational calculus, which by the way is not covered in Ballard's book, is that we have not yet bridged the gap between several known and useful techniques, such as those covered in that book, and the variational approach. Such a unification no doubt provide one of the great challenges for the next millenium.
|
|