On Thu, Sep 05, 2019 at 10:46:54AM -0700, Adam Golding wrote:
> I want to try automating programming as search where I have various methods
> to enumerate the set of all programs in different orders (fastest to halt
> first? shortest source code first? etc.) and filter out certain programs
>
It's as they say: if you want an answer on the internet, just say the wrong
thing and wait for someone to correct you. :)
Original Message
On Sep 5, 2019, 3:45 PM, Laurent wrote:
> Adam,
>
> I strongly recommend you take a look at Levin Search (aka Universal Search),
> which
Adam,
I strongly recommend you take a look at Levin Search (aka Universal
Search), which enumerates and runs programs by dovetailing in a smart way:
http://people.idsia.ch/~juergen/mljssalevin/node4.html
http://www.scholarpedia.org/article/Universal_search#Universal_search
This avoids the
In all honesty, I think you are asking for something so broad that it would be
more practical if you ran experiments and simulations against the specific
space you are curious about. Once you do that enough times you'll be able to
pick out the patterns. I'm not convinced there is an off-shelf
It almost sounds like you want a cleaner interface for defining a neural net.
~slg
‐‐‐ Original Message ‐‐‐
On Thursday, September 5, 2019 2:07 PM, Adam Golding
wrote:
> At this point I don't need the search (through the space of all programs) to
> be efficient I just need it to be
At this point I don't need the search (through the space of all programs)
to be efficient I just need it to be 'total' in that every program would be
reached given a countable infinity of time. Then I could alter this search
to search the same space in a different order (still total) and
Thanks, that helps.
In the way I read this, it sounds like you want a program that computes the
most efficient search algorithm regardless of the context in which said
algorithm is used.
Is that accurate?
~slg
‐‐‐ Original Message ‐‐‐
On Thursday, September 5, 2019 1:48 PM, Adam
The relevant part of the fb thread is where I remark "Suppose every voter
is a program--then we can randomly sample programs and voting methods on
objective questions and see which methods the epistemic argument for
democracy is likely to be correct for" -- the word 'randomly' can probably
be
Basically I want to enumerate programs with as few assumptions as possible
aka enumerate the largest set of programs I can--I want to be able to
enumerate them in a variety of different orders to compare search
strategies.
On Thursday, 5 September 2019 13:46:54 UTC-4, Adam Golding wrote:
>
> I
I want to try automating programming as search where I have various methods
to enumerate the set of all programs in different orders (fastest to halt
first? shortest source code first? etc.) and filter out certain programs
almost like evolutionary programming. I don't have a specific
This question can be read a couple of different ways too. What are you trying
to do once you have the answer you are looking for?
Original Message
On Sep 5, 2019, 1:13 PM, Adam Golding wrote:
> What is the shortest program listing the largest list of programs that can be
>
What is the shortest program listing the largest list of programs that can
be listed without looping?
On Thursday, 5 September 2019 11:10:59 UTC-4, dvanhorn wrote:
>
> How about this: a stream of strings which can be be parsed and
> compiled. (Note that this will loop when it gets to the first
This program works in 7.3, but not 7.4, which complains about the use
of strings before it's definition. Swapping the order of valid-progs
and strings fixes that, but the program then loops, although I don't
understand why.
On Thu, Sep 5, 2019 at 11:10 AM David Van Horn wrote:
>
> How about
How about this: a stream of strings which can be be parsed and
compiled. (Note that this will loop when it gets to the first program
that makes the compiler loop; luckily it's inefficient enough that
you'll never actually get there.)
#lang racket
(define valid-progs
(for/stream ([p strings]
It's okay if the program never halts.
--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to racket-users+unsubscr...@googlegroups.com.
To view this discussion on the
On Thursday, September 5, 2019 at 3:05:48 PM UTC+2, Adam Golding wrote:
> What is the shortest/smallest racket program (ithat enumerates all and
only valid racket programs?
Given that "valid" means "a Racket program that compiles correctly".
As the Racket compiler is Turing Complete and can be
*that
On Thursday, 5 September 2019 09:05:48 UTC-4, Adam Golding wrote:
>
> What is the shortest/smallest racket program (ithat enumerates all and
> only valid racket programs?
>
--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe
17 matches
Mail list logo
