Then I suppose one could then say that there is another class of
"algorithmic information theory" that does not assume unbounded memory.
Also, if one wants to incorporate time, Levin Search/Schmidhuber's speed
prior are less rigorous than DCG NOR for the same reason that the
instruction set is not specified for the UTM.  DCG NOR at least has a clear
unit of time:  A state transition.

On Sat, Nov 20, 2021 at 1:14 PM Matt Mahoney <[email protected]>
wrote:

> Wolfram's 2 state 3 color Turing machine is the simplest, although it
> requires an infinite non repeating input and does not have a halting
> state. It is also very difficult to program.
> https://en.wikipedia.org/wiki/Wolfram%27s_2-state_3-symbol_Turing_machine
>
> Rule 110 cellular automata is also Turing complete and very simple.
> https://en.wikipedia.org/wiki/Rule_110
>
> We could argue whether n-NOR or 2-NOR state machine is simpler, or
> whether these are universal since you have a fixed sized memory. It
> would still be an interesting challenge to find the smallest state
> machine that outputs the bits of enwik9 in sequence.
>
> On Sat, Nov 20, 2021 at 1:41 PM James Bowery <[email protected]> wrote:
> >
> > Algorithmic Randomness is very clear:
> >
> > A random string of bits cannot be represented as a program in fewer bits.
> >
> > On Sat, Nov 20, 2021 at 12:19 PM Jim Bromer <[email protected]> wrote:
> >>
> >> So if I see a string of counting numbers with a length greater than 3
> or 4, I would conclude that those numbers are not "random" based on my
> experiences or samplings of strings of numbers, the psychology of
> elementary number theory and my awareness from thinking about this stuff.
> But that is psychology  - sociology - and the familiarity of the
> sociological use of mathematics. It would not be based on pure mathematical
> analysis or something Platonic like that.
> >
> > Artificial General Intelligence List / AGI / see discussions +
> participants + delivery options Permalink
> 
> --
> -- Matt Mahoney, [email protected]

------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T5ff6237e11d945fb-M4151a7cb2239546fcf5f2fa7
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to