The history of the word "algorithm" is from the Arabic and it refers to numerical methods. When we are talking about numerical computations I tend to use the word "algorithm" in a certain way to express the core of the subprogram that makes a certain kind of calculation. However, when talking about general programs my usage of the word is different.
Although a program might be imagined as continuing forever, most algorithms would be finite of course. However... We can think of an AGI program as a kind of algorithm because it will be solving problems. It will not solve all problems just as human beings cannot solve all problems. But it would be solving a great variety of problems. And we would need those problems to be solved in a finite amount of time! (Of course.) Another part of the current definition of the word includes the concept that an algorithm is a solution to a particular problem. However, even in mathematics we generally have algorithms that solve a class of problems. In AGI we are hoping to be able to write an algorithm that would solve a much greater variety of problems including problems that were not predefined. If I was working on an AGI program I would use the term algorithm to refer to a subprogram. Since as I pointed out, we cannot predetermine what a subprogram of an algorithm that is interacting with IO will do, and we cannot predict how an algorithm that is to be derived from interacting with the IO will be written by the program, the current Webster's or Wikipedia's definition is not relevant to our discussion. The modern usage of the term and the underlying realities of how computer programs can produce unexpected sequences of operations demands a more sophisticated definition of the word. One cannot assume that the participants of an AGI group like this will use the term "algorithm" only to refer to predefined numerical calculations. Jim Bromer On Thu, Oct 11, 2012 at 10:16 PM, Jim Bromer <[email protected]> wrote: > On Thu, Oct 11, 2012 at 8:42 PM, Charles Hixson < > [email protected]> wrote: > >> It's not absolutely clear that the brain operates on an algorithm. Not >> given that algorithms are by standard definition composed of a finite >> number of steps which are all determinate. >> > > > That is (part of) one definition of algorithm but it is not the only > definition and it is not a historically established "standard definition". > Another definition is that of a subprogram which does not, by definition, > need to be composed of a finite number of steps. A better definition is > something that is general but is based on a sequential combination of > "steps". > > Furthermore, even though the next step of a program may be determinate, it > is determinate given the input conditionals. Since input of a device with > extensive IO cannot be predicted before hand, the nature of a subprogram is > not strictly determinate in the sense that it is undetermined until it > occurs. > > This kind of analysis is not rocket science, but it does require > some effort. > > Jim Bromer > ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968 Powered by Listbox: http://www.listbox.com
