[Sigia-l] Findability

[Peter Morville]

Thanks Karl.  

I agree that Google's tremendous success came as a big shock to many of
us with roots in the LIS community.

However, I'm not (yet) convinced by your attempt to link Google to
genetic (evolutionary) algorithms.

My understanding is that Google's algorithms have been created by very
smart humans.  Also, these algorithms are leveraging manually-created
structural and descriptive metadata (e.g., structure of each web site,
domain names, title tags, links between sites) in addition to full-text
indexing.

Do you have evidence Google is using genetic algorithms to enable
fast-forward software evolution and emergence?  If so, please share!

* As an aside, I'm curious if you agree with Ray Kurzweil
(http://www.amazon.com/exec/obidos/tg/detail/-/0670882178/) that through
rapid developments in evolutionary algorithms, "computers (will)
routinely pass apparently valid forms of the Turing Test" by 2029.

Cheers!


Peter Morville
President, Semantic Studios
www.semanticstudios.com


-----Original Message-----
From: sigia-l-admin at asis.org [mailto:sigia-l-admin at asis.org] On Behalf
Of Karl Fast
Sent: Monday, July 21, 2003 9:49 AM
To: 'Sigia-l'
Subject: Re: [Sigia-l] Findability



> > My opinion is that (a) a balance is required and (b) that balance 
> > will keep shifting in favor of algorithms because algorithms will 
> > evolve at a much faster pace than we will.

> I agree with your overall perspective here, but could you clarify your

> theory on the evolution of algorithms?


Simon pretty much covered it, but let me respond from a slightly
different angle.

Consider the evolution of information retrieval over the last few
decades where we have gone from the foundations laid down by Salton to
developments like Google and Teoma.

Imagine it's 1983, Reagan is still in his first term, and I said that in
twenty years it would be commonplace to search through a three billion
document set in less a second and return results that most people were
satisfied with most of the time. Not perfect mind you, but good enough
that the company offering this service would find their name being
turned into a verb and being suggested as a new word in the OED.

If I said this you'd think I was bonkers. But you'd be wrong.

Google and Teoma represent significant evolutionary advances in
algorithm development. And that's just information retrieval.

In the same time frame (two decades), how far have human beings advanced
in our abilities to classify and organize information? 

There are astounding developments happening in areas like genetic
algorithms. Consider this recent story in Discover:

   http://www.discover.com/aug_03/gthere.html?article=feattech.html

And the related video of a software biped learning how to walk based on
an evolutionary genetic algorithm:

   http://www.naturalmotion.com/pages/technology_hiw.htm


I am not saying that humans do not evolve. Nor am I saying that
algorithms will replace people, there is no need for humans in terms of
organizing information, or that we better start preparing for a
Terminator/Matrix scenario.

To say that would be like sitting in 19th century England, watching the
rapid development of the steam engine and other mechanical devices, and
concluding that human beings were doomed because we weren't evolving as
fast as the machines were (however, if you've read "The Muse in the
Machine," by David Gelernter, a book about artificial intelligence, you
may remember that he argues that humans have evolved significantly in
the last few thousand years, and I think his point is excellent).

I'm not saying all that doom-n-gloom stuff.

But what I am saying is that algorithms are developing quickly and that
this rapid development will continue to shift the balance between what
humans do and machines do. This shift is accelerating because algorithms
are only getting better. And they are getting better really fast.

Meanwhile, humans are pretty much the same as we've always been. If we
are evolving (and I think we are), it's happening at a much slower pace
and in different directions than, for example, algorithms.

As Simon notes, humans don't scale well. So we are abdicating the job of
sifting through huge piles of information to algorithms. This makes
sense.

But computers don't understand well. Humans do. Google doesn't
understand what it finds. Humans do.

It's been noted that the purpose of computing is insight, not numbers.
But it's important to understand that humans are the ones who develop
the insight. The computer still just sees numbers, no matter the
algorithm. Humans develop insight. Computers just run the algorithm.
 

Of course, if someone makes a cyber-librarian like in Neal Stephenson's
"Snow Crash" then it's game over.


--karl   
------------
When replying, please *trim your post* as much as possible. *Plain text,
please; NO Attachments

Searchable list archive:   http://www.info-arch.org/lists/sigia-l/
________________________________________
Sigia-l mailing list -- post to: Sigia-l at asis.org
Changes to subscription: http://mail.asis.org/mailman/listinfo/sigia-l