Phoenix.pm: Meeting Announcement: Thursday, June 5th, 7:00pm

[Scott Walters]

Hi Jason,

Sorry for the slow reply.

I'm going to answer your questions out of order. Let me know if I miss one.

The presentation assumes no prior knowledge - atleast the first presentation.
I'm not sure about the next two. I've commited myself to a series. Just
call me Captain Kirk.

Fuzzy Logic does work well for expert systems, but is popular in two other
areas: control (embedded applications/robotics/etc), and working with 
approximate ("imprecise") data. Actually, you can have it two ways - reasoning 
with approximate data, and approximate reasoning with data.

This is subject to revision, but as it stands, each presentation will
cover one of these three aspects of Fuzzy Logic. The first one will be a simple 
membership test used as a search engine. This illustrates processing approximate
data. 

I'm trying to remove the last kinks from a modified version of 
http://www.slowass.net/geekdate/ . I wrote that as a quick hack then later 
realized it would good a good Fuzzy Logic example.  It is kind of a joke - an 
online dating service for geeks.  "Geek Date" crawls Google's search results 
for "BEGIN GEEK CODE BLOCK", downloads all matching pages, and converts each 
Geek Code block to a set-of-sets. When a search is peformed, a Fuzzy membership 
function is performed between a set-of-sets built from the search criteria and
each set-of-sets built from each Geek Code block, and results are returned 
sorted by the degree of Fuzzy membership. The Geek Code lends itself well to 
building Fuzzy Sets. For example, P++(--) means something like "I like Perl and 
use it write most things, but sometimes I don't like Perl very much". Search
attributes can't be represented as a scalar value. This person might show up
in searches for people who like Perl, in searches for people who don't like
Perl, and to some lesser degree, people who are neutral about Perl. Fuzzy
Membership tests used in this way also find "near matches". A geek has
interests besides Perl, and a complex search might specify any number
of attributes, so Fuzzy Sets are wrapped up into sets-of-sets. My 
FuzzyLogic module overloads the normal operators in Perl (& is the Fuzzy
membership test, for example), so working with sets-of-sets is syntactically
the same as working on normal scalars.

Thats my example of using Fuzzy Logic for its membership tests. Actually,
I might have one more minor example to illustrate Fuzzy Logic's ability
to deal with non-linear data (a common and humorus failing of most control
computers/industrial robots). Time will tell.

I haven't started my examples for industrial control and the expert
system yet, but I think I know what I'm going to do.

Through these examples, I'm trying to explain "what is Fuzzy Logic?"
(no one can be told what Fuzzy Logic is, you must be shown for yourself),
and how to use it in practical applications using my particular 
implementation of it. I'm always trying to do everything at once =)
Understanding the Perl should be easy; understanding the Fuzzy examples
should be easy; understanding how to apply Fuzzy primitives to solve 
problems that have yet to have been thought of will be more difficult,
but I'm going to try.

As whether or not you would find it worth while, I don't know you well
enough to say. From helping people on #perlhelp on EFNET, I've noticed that
people latch violently onto the first solutation that presents itself
(a tendency that can be used to great amusement). Most people are confident
that they can tackle any given and find a working solution. Most people
are also confident that there are far better solutions than what they
could cook up, and that they would spend more time than they want in the
process. Depending on what you do with Perl, Fuzzy Logic might be a life
saver or it might be a complete waste of time. I have no doubt that
AI has wasted far more time than it has saved, to date. To me, modern
life is advanced form of wasting time, and computers doubly so. We
all know damn well that we should be busy hunting and gathering right now,
and a lot of us deserve to be eaten by bears.

-scott

On  0, Jason Riedel <jasonriedel at jasonriedel.com> wrote:
> 
> Scott,
> 
>     I know nothing about Fuzzy Logic except that I have heard the term. I
> did some what is and FAQ look ups on it and seems it is only used on fuzzy
> expert systems and very few other real world implementations. Do you think I
> would find this topic worth while ? For example, is the presentation geared
> towards what is fuzzy logic then into more depth or do you assume knowledge
> of what is fuzzy logic and it's uses and will just be discussing your module
> theory, implementation and use ?
> 
> Thanks,
> Jason Riedel
> 
> ----- Original Message ----- 
> From: "Scott Walters" <scott at illogics.org>
> To: <phoenix-pm-list at happyfunball.pm.org>
> Sent: Monday, June 02, 2003 1:17 PM
> Subject: Phoenix.pm: Meeting Announcement: Thursday, June 5th, 7:00pm
> 
> 
> > We'll have a Phoenix.pm meeting Thursday, June 5th, at 7:00PM.
> >
> > It will be held at Bowne, which is located at 1500 N. Central Avenue,
> > which is on the Southwest corner of Central and McDowell.  The parking
> > lot is gated, so just press the button on the intercom, and tell the
> > receptionist that you are there for the Perl meeting.  Park in the lot
> > that is straight ahead from the entrance on the South side of McDowell.
> > Park in any uncovered, non-reserved space.  Proceed to the main lobby,
> > which is on the Northeast side of the parking lot.
> >
> > Scott Walters will be presenting the first part in a three part
> > series on his FuzzyLogic module:
> > Theory of Fuzzy Logic, a proposed CPAN module implementing it via
> > overloaded operators. An example will be presented that performs
> > fuzzy classification.
> >
> >
>