|
|
|
|
|
|
(e) successful behavior by a scientist working in a given, possible reality |
|
|
|
|
|
|
|
|
One important question about a given paradigm is this: For what classes of possible realities do there exist scientists who are guaranteed to succeed within any reality drawn from the class? |
|
|
|
|
|
|
|
|
In order to apply the resources of computational theory to the problem of inductive inference, possible realities are often conceived as sets of integers. In turn, the integers can be conceived as codes for complex objects such as sentences or experimental data. Hypotheses within our theory are usually taken to be computer programs operating over integer inputs. A scientist is any system that converts the finite data sets generated by an environment into hypotheses about the totality of that environment. We shall mainly be concerned with scientists whose behavior is simulable by computer. Much of our inquiry will be devoted to scientists who possess special properties, such as efficiency, ability to cope with noisy data, etc. |
|
|
|
|
|
|
|
|
Through a simple game, we illustrated one criterion of successful scientific behavior; subsequent chapters will investigate a variety of alternatives. It was noted that our theory distinguishes scientific success from the confidence that a scientist might feel about such success. Only the former will be at issue here. |
|
|
|
|
