Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon. Entire thread

Non-computability.

Name: Anonymous 2010-03-04 21:21

According to Roger Penrose, humans can perform non-computable feats, such as dealing with Gödel questions. He uses this as a foundation to claim that the human mind cannot be expressed in terms of classical processes, and as such must be party to the only other (known) game in town: Quantum Mechanics.

Now, I haven't had the patience to sit through all of his arguments yet, though I slowly make progress. My understanding is that a large part of his stance is that an algorithm cannot usefully deal with a Gödel question, or equivalently, with the halting problem, while a human can.

My objection to this is that such problems always demand a certain quality of response when asked of UTMs: failing to respond forever is not acceptable as correct, nor is providing any response other than one that yields a truth when taken in combination with the question. This much is fine, however, when it is time for the human to answer, he is permitted the liberty of rejecting the question on the grounds that it is inherently unanswerable.

Obviously I am interested in artificial intelligence, and also find his assertion to be simply a self-serving one with a contrived philosophical backdrop for foundation. If anyone knows of, or can think of, a more sophisticated argument than the one above (or expose my flaws in my assessment of it) I would like to hear it.

Apologies for bringing up a largely philosophical question, my only excuse is that I cannot trust any other board with the question.

Name: Anonymous 2010-03-11 6:46

>>77
Well it was high up enough on the page really. Other than that, if I'd realized the 'optimizing the bbcode' thread was threadstopped I would have bumped it, but I didn't notice 'til after. Now it wouldn't matter because the old thread got bumped and we even have a new one too!

>>79
To be fair it wasn't the real halting problem and it can't be proved impossible to solve in the standard way... and I don't think anyone was really trolling.

Name: Anonymous 2010-03-11 18:26

I am confused by the assertion that the Halting Problem is unsolvable for Turing Machines with a finite tape. Surely you could enumerate all the possible "super-states" (i.e. a subset of TM states × tape × alphabet) of such a bounded TM and thereby determine if the TM ever entered a loop?

Name: Anonymous 2010-03-11 18:36

>>82
Yes, you could. I'm not sure why such an obviously false assertation would confuse you.

Name: Anonymous 2010-03-11 18:50

>>82
But Turing Machines do not have finite tape.  Having an "infinite tape" means that any size program of any complexity can be emulated on it.  You can argue that it isn't possible to write a program that has an infinite number of states, which makes more sense.  So, let us assume a situation where the program does not enter a state more than twice.

Instead of a program, a thread performs some task but enters a situation where it needs to perform more tasks before it can proceed.  It creates and runs a recursive thread on what is left over and has to wait for that new thread to "join."  We shall assume that this is the top level.  Our Turing Machine will prove whether this top level original thread ever finishes.  So now you're in a situation where the Turing Machine has to simulate itself.

Name: Anonymous 2010-03-11 19:05

>>84
Sorry, I should have added: I know my admittedly hasty experiment violates a major rule of the Halting problem - that the Turing Machine automatically replies "stops" or "loops" - and any actually running of the program P on input I is considered hypothetically; but, the standard definition of the Halting problem and of a Turing Machine are not a part of your concern anyway.

Name: Anonymous 2010-03-11 19:05

>>82,83
You must use the very same tape for both TMs.

Name: Anonymous 2010-03-11 20:09

>>83
Wasn't this obviously false assertion made upthread, and taken seriously?

Name: Anonymous 2010-03-11 20:14

Not only that, it wasn't (and isn't) false.

Name: Anonymous 2010-03-11 20:25

>>88
0/10

Name: Anonymous 2010-03-11 20:36

>>89
Ah, imageboard scum. No wonder you can't count.

Name: Anonymous 2010-03-11 21:08

lol wut

Name: Anonymous 2010-03-12 3:31

>>86
No.

Name: Anonymous 2010-03-12 8:54

>>92
Oh really? Anything else is meaningless. Separating the two TMs means we're no longer speaking about any kind of universe of computation, but some arbitrary setup that says nothing at all about general computability--the issue at hand.

Name: Anonymous 2010-03-12 9:05

>>81
and I don't think anyone was really trolling.
I can confirm that I have in fact been trolling in this thread. Have a nice day.

Name: Anonymous 2010-03-12 13:44

>>94
Well you are now. Whether or not you were in the past is unverifiable.

Name: Anonymous 2010-03-12 18:22

>>95
I have MD5'd "YHBT" into the email field of one of my previous posts. I hope this verification is sufficient.

Name: Anonymous 2010-03-12 19:23

Not really, but I'd buy it anyway. I'm not a verificationist except as a defensive manoeuvre.

Name: Anonymous 2010-03-12 21:10

>>96
fuck IHBT

Name: Anonymous 2010-03-12 21:35

>>98
Oh dear, did you go checking all of the email fields?

Name: Anonymous 2010-03-12 22:04

>>99
Yes, then I fucked >>96-san's mother.

Name: Anonymous 2010-03-12 22:09

>>100
I'm quite amused by this because I took >>96-san at his word and didn't bother checking because... well, who really cares?

Name: Anonymous 2010-03-12 23:12

>>99
I paged through the source briefly, because I had nothing better to do than BT, and I didn't see any MD5sums, but what the hell is with >>27?

Name: Anonymous 2010-03-13 1:28

>>102
You could have applied a little grep.

Anyway, >>27 is trying to enumerate states. Infodynamics applies to this kind of procedure, and the jig is up once entropy is maximal and resource pressure forces an infinite loop or termination.

This is why >>92 doesn't want to run both TMs from the same tape: solving the problem in finite space is only possible by removing a class containing all of the principally insoluble cases (including the candidate TM itself!) Recast to the infinite case, this would have the effect of removing Turing Completeness, so it's easy to see that we're no longer talking about Turing Machines with limited resources, we're talking about something else.

At this point the problem is no longer interesting for any of the same reasons. The machines just aren't TC any longer, and the problem space isn't related to completeness in any form.

Name: Anonymous 2010-03-13 9:35

>>103
At this point the problem is no longer interesting for any of the same reasons. The machines just aren't TC any longer
You must find DFAs, PDAs, and LBAs a real bore then.

Name: Anonymous 2010-03-13 9:43

>>103
view-source:http://dis.4chan.org/read/prog/1267755680/27

grep wouldn't have worked if there was some unexpected difference in the input - letter case, extra punctuation, spacing. But that's neither here nor there, I didn't care enough to actually look for it.

Name: Anonymous 2010-03-13 9:47

>>105
That's not an md5 sum. It's a bunch of [m][o][/m] tags

Name: Anonymous 2010-03-13 11:05

>>106
Holy shit you're a genius.

Name: Anonymous 2010-03-13 13:13

>>104
for any of the same reasons
Those reasons all depend on a notion of TC being preserved.

>105
I was thinking something along the lines of grep mailto, or better yet: grep -Po "mailto:.*?\"" |grep -v :sage\" VALID PERL REGEX

Name: Anonymous 2010-03-13 15:45

>>108
Why would you do that?
mailto:(?!sage)

Name: Anonymous 2010-03-13 16:08

>>109
Three reasons:
1. grep doesn't implement look-ahead in PCRE mode
2. Your pattern only makes post-pattern assertions about the mailto, and (if it worked) would not display the contents with -o (without -o, you get a bunch of garbage you don't care about)
3. There is a chance that the MD5 sum begins (or ends) with "sage" (hence the \" in my pattern (and the : in the grep -v))

Of course the real solution is to compute the hash and grep for "mailto:$hash" but that isn't nearly ENTERPRISE enough.

Name: Anonymous 2010-03-13 18:56

1. Use perl -ne 'print if /mailto:(?!sage)/' and do that all in one step. Why fork two processes? grep's PCRE mode sucks anyway.
2. Point irrelevant if not using grep. Aside from that, your solution is broken, because the existence of the word "sage" anywhere will cause the line to fail to display.
3. The canonical form of an md5sum is 32 hexadecimal characters, but good point anyway, you're thinking outside the box and that's a good thing; sure, maybe it was base 32 or base 64 encoded instead of base 16. /mailto:(?!sage")/ it is, then.

And of course the real solution is to extract the e-mail fields using an HTML parser, and pipe through sort -u. As already stated, searching for a precomputed hash is a terrible! solution because there is a good chance that the input you use is different than that used to create the md5sum you're searching for. Capitalization chance is possible, although unlikely, that this might be different than advertised. Punctuation might be different -- are the quotation marks included? Line endings could certainly be different. Did you use echo or echo -n? If there is a newline, is it \r\n or \n?

Name: Anonymous 2010-03-13 19:04

And here is that real solution, which I just hacked together and verified:


import urllib2, BeautifulSoup, re
set(tag['href']
    for tag in (BeautifulSoup.BeautifulSoup
                (urllib2.urlopen('http://dis.4chan.org/read/prog/1267755680/').read())
                .findAll('a', href=re.compile(r'mailto:(?!sage$)(.*)'))))

Name: Anonymous 2010-03-13 19:11

>>111
perl -ne 'print if /mailto:(?!sage)/'
Fails point 2. Also invokes perl.

the existence of the word "sage" anywhere
No, the existence of :sage just before the first ". I pointed this out already. : isn't an alphanum, so it won't appear in the hash.

there is a good chance that the input you use is different than that used to create the md5sum you're searching for
``YHBT'' only has one md5sum... that's kind of the point.

Name: Anonymous 2010-03-13 21:04

>>113
Your and idiot!

Newer Posts
Don't change these.
Name: Email:
Entire Thread Thread List