Re: Other "tested" techniques for detecting bias

On Aug 3, 5:10 pm, fropome <monk...@xxxxxxxxxxxxxxxxxxx> wrote:

< snip >

By 'some' of the cases, you mean 'those cases where I know the
production algorithm or the sequence is so simple I think I can guess
it'. Do you have any idea how useless that is?
1) If you know that an algorithm was used then you don't need to
perform your test.
2) You will only be able to guess the algorithm is an infinitely small
number of cases (when you don't get to choose the sequence) anyway.

You can't "know that an algorithm was used" until you've actually
carried out a test to support that hypothesis. It is only after the
prediction that such and such algorithm was actually used is
successful that the hypothesis gains predictive value - never reaching
perfection.

Here's a string:
100100100100

If you guess the next two digits as '10' then are you right? How do
you know?

You don't "know" until you test your prediction against what actually
comes next. Predictions are about the future of the unknown - a
future that cannot be perfectly known. That is what science is all
about - predicting the future with better odds of success than a
random guess. If your hypothesis about what will happen in the future
actually happens, it gains "predictive value". Ever heard of it?

How does this help you predict the two after that?

Through inductive reasoning. Every successful prediction gives the
hypothesis, the prediction algorithm, greater "predictive value".
Through inductive reasoning one assumes that what has been consistent
in the past will continue to be consistent in the future. This
reasoning becomes more and more valuable with each successful
prediction. This is the very basis of science itself.

You seem
to be assuming that the more guesses you get right, the more likely
the string is to be created by the algorithm that you've guessed, but:

That's correct . . . Again, this is what inductive reasoning is all
about. It isn't about perfection. Could I be wrong? Yes, but the
odds that I will be wrong decrease with each predictive success.

Also, don't underestimate simple algorithms. They can be very useful
in discovering quite a few biased processes. Sure, the vast majority
of potential biases are beyond our detection, but that's the nature of
science. Science is very limited. This is the restriction that SETI
scientists have to work with. Even if there were intelligent aliens
all over the place in the universe sending signals all around, if they
didn't want us to discover them, there would be an infinite number of
ways to conceal their signals so that we could not decipher them. It
is basically on the extremely remote chance that aliens will actually
use some sort of signal that we could decipher that SETI scientists
have even the smallest glimmer of hope of ever finding any clear
evidence at all of deliberate artifactual extraterrestrial radiowave
generation.

This is nonsense. SETI is looking for signals where the creator
deliberately designed the signal in such a way that a receiver would
see the signal as designed.

Oh come on now! SETI scientists are *not* simply looking for signals
that are sent with the intent that human beings be able to detect the
signal as artifact. That's utter nonsense. SETI is looking for
signals of alien intelligences that most likely have no idea that we
humans even exist and have no clue as to what we would or would not be
able to detect as "artifactual". You really need to read up just a
little bit on the assumptions behind SETI.

This would be the sole useful property of
the received signal. As such the signal would be chosen in order to be
obvious- thus being from a repetitious algorithmically produced
string.

This is again, way way off base - and obviously so to anyone who has
done even a bit of reading into what SETI scientists are actually
looking for.

< snip a lot more nonsense >

To which tested technique are you referring? Please do explain to me
any tested technique you know of by which you can detect and measure
bias, to a useful degree of confidence, without some fundamental use
of the technique I've just described here . . .

Why don't you do a google search? The following search terms hit a
number of results you can start your research on:
random number test
random number detection
random number cryptography

Why don't you describe here how one of these methods actually works?
How do scientists or mathematicians "detect" a random number? I don't
think you have the first clue.

I'm sure you can find search terms of your own.

I don't know if you're aware (of course, now I've said it you'll claim
you were all along whether you were or not), but random numbers are
very important to cryptography (a master- pardon my sarcasm- of
mathematics such as you will know why of course). As such, some of the
world's best mathematicians have done an awful lot of work in this
area, including those working for big software companies or banking
institutions or government or... well, anyone for whom cryptography is
important, which is mostly everyone big and important as well as a lot
of people who aren't. Claiming that you intuitively came up with a
formula which they didn't think of isn't just unfeasible- it's hugely
unlikely and arrogant... but then, having lurked here for a while it
doesn't surprise me that you've made this claim.

Explain to me how my formula is wrong then. Provide just one example
where it doesn't work. So far, all I've seen coming from your end is
bald claims like this with absolutely nothing to back yourself up.

It's been interesting to argue with you- unfortunately not because
you've had valid points, but because I've discovered a lot about this
news-group. It's made me appreciate how patient some of the posters
here are... not being a biology or geology expert, this exchange has
made me realise quite how ready you are to lie, obfuscate and
generally make shit up in order to appear to have a valid argument. I
don't know whether this is because you actually believe what you're
saying or not- unlike some I think you probably do believe what you
say, but I'm not sure whether this makes me more or less pessimistic
than those who think otherwise.

Regardless, my brief involvement here has given me an insight into
creationism that I would otherwise not have gained and I thank you for
that- if I ever had any doubts then you've removed them
comprehensively. I'm sad that there are people who are as self-
deluded/ dishonest as you are (delete as applicable, iff [sic]
applicable) but glad that since there are I'm more aware of what they
are like.

Yeah - - guys like you are a dime a dozen in this forum. You come in
with all kinds of bravado but very little substance. When asked for
any sort of real evidence or counter evidence to back up your bald
claims, you make the classic argument from authority, "How can you be
right and everyone else be wrong?" Pardon me, but that common
argument has absolutely no explanatory power. You're going to have to
do better if you want to convince anyone with anything more
substantive than "other people know more than you do".

I've spent the last couple of weeks watching this thread like a hawk-
waiting for a valid responses to my posts, checking my arguments,
worrying about the way I put things, hoping I'm easily understood- and
it's only now the elucidating light of strong beer is with me that I
see I'm wasting my time. You don't have a valid argument and... well,
to put it frankly, you don't have a friggin' clue what you're talking
about. I wish you did, because quite honestly it'd make the whole
thing more entertaining.

My thoughts exactly . . . It's like you've been drunk this entire
time given the arguments you've presented thus far.

I'm going to back out here before I become obsessed with rebutting
ridiculous arguments. I'm glad that the regulars here have the
patience to correct your ignorant rubbish but I'm afraid I don't. I've
got admiration for all those who have the time to respond correctly
and more of an appreciation of why people can be so insulting at
times.

A creationist friend of mine told me once that he thought himself a
'signpost' to Christianity and God. He's a good guy (as much as we
argue) and I consider him a good friend, but you're more of a signpost
to atheism and secularism than 100 Richard Dawkings could hope to be.

Wow! Not bad . . . better than 100 Dawkins's you say? Thanks! ; )

I don't know if you are capable of reconsidering what you're doing in
posting the crap that you do, but I hope you at least think about it.

Ditto . . .

Apologies again for the rambling nature of this post, I did warn you
at the start.

It is a common theme in this forum for those who have no real argument
to resort to the only thing they have left - empty personal
pejoratives.

Thanks,

Any time . . .

fropome

Sean Pitman
www.DetectingDesign.com


.

Relevant Pages

  • Re: Is the Internet Intelligent?
    ... believe we are a reinforcement learning engine at the core, ... this fails to understand that the algorithm running in is us not like a ... shapes behavior using reward prediction - using a value prediction ... High values of future reward prediction, ...
    (comp.ai.philosophy)
  • Re: Other "tested" techniques for detecting bias
    ... If you know that an algorithm was used then you don't need to ... ways to conceal their signals so that we could not decipher them. ... coding theory and cryptography as ... any tested technique you know of by which you can detect and measure ...
    (talk.origins)
  • Re: Is the Curt net a kind of decision tree?
    ... only prediction it needs to make. ... maximizing function on it's own, and as a result, there was no need for ... output of the amplifier. ... input signals, to a different set of output signals. ...
    (comp.ai.philosophy)
  • Re: Chex Wat: Pi is "random" and "not predictable"?
    ... the sequence appears to be part of a "normal" number or not. ... This compression algorithm can then be used to predict ... If this prediction succeeds, it gains predictive ... predictive value as the string in question increases in size with each ...
    (talk.origins)
  • Re: implementing K Nearest Neighbour algorithm using Managed PLUG-IN question
    ... public class Algorithm: AlgorithmBase ... Prediction happens based on in-memory patterns ... Tokenized form of the same case is a set of attribute/value pairs: ... training set is required for each predicition. ...
    (microsoft.public.sqlserver.datamining)