Re: consciousness
- From: Stephen Harris <cyberguard-1048@xxxxxxxxx>
- Date: Sun, 01 Apr 2007 10:59:18 GMT
Don Geddis wrote:
Stephen Harris <cyberguard-1048@xxxxxxxxx> wrote on Sat, 31 Mar 2007:I am unsure of the status, can a Turing machine also incorporate sensory
data while processing; I don't think so according to the original
description of a TM, which has a purpose which is fulfilled solely by a
finite predetermined input being computed (is computable)-> IMO.
One quibble: you describe the TM's input as "finite", which is not the case.
Formal Turing Machine have an infinite tape, and their input could be
infinite as well.
As for "predetermined" ... that is, in essence, correct. But algorithms are
deterministic. Imagine you had some different device which had sensors and
effectors in the real world. Over the course of its computations, say years
or centuries, it would examine some finite amount of data. Clearly it's the
case that you could (in principle) record all of that data, and then start
the machine all over again with exactly the data that it had on its first
run (of years or centuries), "preloaded" onto its input for the second run.
The deterministic machine would necessarily go through the same states and
computations the second time as it did the first time.
predetermine: b : to determine beforehand
2 : to impose a direction or tendency on beforehand
I'm fairly persistent and have kept digging. The quote uses pre-given
for predetermined. I notice the paper reads, "they claim". I'm using
what the quote says about Turing Machines, not Interaction Machines.
www.macs.hw.ac.uk/~greg/publications/cm.cj05.pdf.gz
"Wegner and Eberbach draw heavily on the idea of an algorithm as
an essentially closed activity. That is, while the TM realising an algorithm may manipulate an unbounded memory, the initial memory configuration is pre-given and may only be changed by the action
of the machine itself. Furthermore, an effective computation may
only consume a finite amount of the unbounded memory and of time,
the implication being that an algorithm must terminate to be
effective. Wegner and Eberbach refer to Interaction Machines as a
class of computer that is more powerful than the Turing Machine.
The latter [TM], they claim, is restricted by having to have all
its inputs appear on the tape prior to the start of computation.
Interaction machines onthe contrary can perform input output
operations to the environment in which they are situated.
The difference between Interaction Machines and Turing Machines,
they claim, corresponds to the technology shift from mainframes
to workstations. Interaction Machines, whose cannonical model is
the Persistent Turing Machine(PTM) of Goldin [Gol01], are not
limited to a pre-given finite input tape [TM], but can handle
potentially infinite input streams." [Interaction Machines]
http://www.earlham.edu/~peters/courses/logsys/turing.htm
"If a Turing machine starts with a non-blank tape, we may
consider that the marks already on the tape are the "input"
to the program that will then execute." [If the marks are
already on the tape, then they must be finite in number.]
.
- References:
- Re: consciousness
- From: Lawson English
- Re: consciousness
- From: Stephen Harris
- Re: consciousness
- From: Don Geddis
- Re: consciousness
- Prev by Date: Re: consciousness
- Next by Date: Re: consciousness
- Previous by thread: Re: consciousness
- Next by thread: Re: consciousness
- Index(es):
Relevant Pages
|