Can you guess the source.

[Phil Henshaw-2]

I guess the thing the brain does, when it has the experience, is suggest
avenues for exploration & you suggest a better way for a computer to do
some version of that than I was thinking of.  If your search engine was
linked to a database of possible associations like you describe it could
prompt a little pull down menu, as if a spell check prompt, of other
word associations you might be looking for.  It could get those
associations from Wiki sources, dictionaries, encyclopedias and things.
It might filter trade name publications for commercial word
associations, and scientific papers for scientific ones, and the
blogisphere for others, and perhaps give you a selection of alternate
association groupings derived from each source.  

That solves the problem that the computer has no way to have any of the
ideas any word refers to for a person, or that it might suggest to a
knowledgeable person, by intelligently focusing the computer's sources
of information.  In a sense it's the opposite of 'garbage in & filtered
garbage out' you get 'wisdom in & filtered wisdom out'.   So I'd put
that in the 'very smart' category for the programmer, but not
necessarily for the computer.  Are you saying that the 'contextualize'
function extracts the meanings of the words though, and looks for other
associations with those meanings, or does it just reflect things like
frequency and proximity of occurrence.  

I'd be very interested in what advanced 'contextualize' functions
actually do, of course, in that some programmer may have been very smart
in thinking of other tell-tale syntax structures in addition to
proximity that a computer might discern, but I'm not aware of any that
would correspond to 'meaning' yet.


Phil Henshaw                       ????.?? ? `?.????
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
680 Ft. Washington Ave
NY NY 10040                      
tel: [phone]                
e-mail: pfh at synapse9.com          
explorations: www.synapse9.com    

> -----Original Message-----
> From: friam-bounces at redfish.com
> [mailto:friam-bounces at redfish.com] On Behalf Of Marcus G. Daniels
> Sent: Sunday, April 22, 2007 10:43 AM
> To: The Friday Morning Applied Complexity Coffee Group
> Subject: Re: [FRIAM] Can you guess the source. - on Earth Day
>
>
> Phil Henshaw wrote:
>
> > How would a computer be able to suggest that
> > when you search for 'Corian' you might actually be looking
> for 'solid
> > surface'.   You might assume that the original discussion that
> > associated the terms was not coded, and only the gradual
> change in usage
> > can be documented (e.g. as for punctuated equilibrium).   I
> can see some
> > assistance, but not a lot, being provided by a computer
> able to mark
> > the growth dynamics of word uses, giving a specific date to
> when a new
> > phrase began to mature (first turning point ending the first growth
> > period).  The poor computer is just never going to be coding the
> > 'idea' the terms convey for people, and won't it always be
> making word
> > associations a different way?
> >  
> Wikipedia gives both definitions:  solid surface and acrylic
> polymer +
> alumina trihydrate.
> In R, for example:
>
>  > dict <- new.env()
>  > dict[["corian"]] <- c(old="polymer",new="solid surface")
>  > contextualize <- function (entry,context) if (!is.na(context))
> entry[context] else entry
>  > meaningOf <- function (word,context=NA) {
> contextualize(dict[[word]],context) }
>
> Then:
>
>  > meaningOf("corian","old")
>       old
> "polymer"
>  > meaningOf("corian","new")
>             new
> "solid surface"
>  > meaningOf("corian")
>             old             new
>       "polymer" "solid surface"
>
> So, given some context, it's a simple matter to grab a subset of
> possible meanings or all of them.   Computers are especially good at
> combinatorics, and can report a confidence interval on any
> conclusions
> by logically extrapolating each outcome from ambiguity in
> each word.    
> Some conclusions might defy common sense though, and that's where I'd
> see something like Cyc coming in to play.   A customer asked for
> granite, or pointed at a polymer-like countertop that wasn't from
> Dupont, and so the inference could be made that they really
> meant "solid
> surface".
>  
> If you like redefine the contextualize function to be smart
> by looking
> at a dynamic description of the immediate environment, or index into
> different dictionaries as a function of time or whatever.
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at http://www.friam.org
>
>

[Marcus G. Daniels]

Phil Henshaw wrote:
> Are you saying that the 'contextualize'
> function extracts the meanings of the words though, and looks for other
> associations with those meanings, or does it just reflect things like
> frequency and proximity of occurrence.  
>  
The purpose of the contextualize function in that ad hoc example is to
get the meaning of a word based on the situation in which it is used.  
In the example it was a simple lookup, but there's no reason it couldn't
do backtracking, on-the-fly simulations/forecasts, or whatever.  
Modeling the context is, worst case, itself a hard problem.  But then I
think people frequently misunderstand one another too, and there is
always the option of the computer asking for clarification.

I'm a little concerned we're going to get bogged down in a
round-and-round discussion about whether artificial intelligence is
possible and whether meaning is something specific to human brains.    
My answer to those questions are yes and no, respectively.   Digital
computation can simulate physical processes as needed (e.g. of brain
neurochemistry) and where digital simulation isn't efficient enough,
there's always the possibility of measuring analog circuits, or even
quantum ones.  I personally expect pure logic will be the best tool for
realizing computer intelligence.

[Phil Henshaw-2]

You're quite right that arguing whether computer intelligence will be
like human intelligence is not the point, but how we can make computer
intelligence useful.  The ways you've been pointing out for how
computers can mine the 'Rosetta stone' type archives that people of all
kinds put together, are wonderful.  It potentially offers useful
extracts from different points of view of the same things.

It seem to me to represent computers being used to refer to ideas, but
it does something possibly more useful in my reckoning.   It offers
another way to attack the wall of subjectivity, buy exposing people to
independent constructs relating to the same things, adding perspective
to information.  Lack of perspective is one of the great barriers to
meaningful thought and any help on the would be great.  I don't expect
machines with 'separate intelligence' except by growing them in the
usual way, but new formats for information serving as 'extensions of
intelligence' that enable a new levels of thought, that may well be
possible!


Phil Henshaw                       ????.?? ? `?.????
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
680 Ft. Washington Ave
NY NY 10040                      
tel: [phone]                
e-mail: pfh at synapse9.com          
explorations: www.synapse9.com    

 

> Phil Henshaw wrote:
> > Are you saying that the 'contextualize'
> > function extracts the meanings of the words though, and looks for
> > other associations with those meanings, or does it just
> reflect things
> > like frequency and proximity of occurrence.
> >  
> The purpose of the contextualize function in that ad hoc
> example is to
> get the meaning of a word based on the situation in which it
> is used.  
> In the example it was a simple lookup, but there's no reason
> it couldn't
> do backtracking, on-the-fly simulations/forecasts, or whatever.  
> Modeling the context is, worst case, itself a hard problem.  
> But then I
> think people frequently misunderstand one another too, and there is
> always the option of the computer asking for clarification.
>
> I'm a little concerned we're going to get bogged down in a
> round-and-round discussion about whether artificial intelligence is
> possible and whether meaning is something specific to human
> brains.    
> My answer to those questions are yes and no, respectively.   Digital
> computation can simulate physical processes as needed (e.g. of brain
> neurochemistry) and where digital simulation isn't efficient enough,
> there's always the possibility of measuring analog circuits, or even
> quantum ones.  I personally expect pure logic will be the
> best tool for
> realizing computer intelligence.
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at http://www.friam.org
>
>