[Top] [All Lists]

Re: [ontolog-forum] Semantic Systems

To: "'[ontolog-forum] '" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Rich Cooper" <rich@xxxxxxxxxxxxxxxxxxxxxx>
Date: Sun, 28 Jun 2009 16:42:37 -0700
Message-id: <20090628234320.A484D138CFD@xxxxxxxxxxxxxxxxx>
John, continuing...    (01)

JS>If you don't believe me, just try to imagine any possible case
in which you could directly observe any operator of FOL other
than existence and conjunction.    (02)

RC>I can more sincerely imagine directly experiencing existence and
disjunction, and the negation of inferred lack of evidence, which also make
a mathematical logic taken as a set of three basis functions.      (03)

RC>Enumerating, or equivalently, distinguishing persistent patterns as
repeats instead of brand new things, is not included in the algorithm I
attributed to SERA.  (The S must stand for something else, but I haven't
come up with a word to substitute if you don't want me to use "Sowa", but I
will come up with a new name if you prefer.)      (04)

RC> Here is a sample parse from the LGP...
JS>You can generate conceptual graphs with a phrase-structure grammar,
but the trees generated by link grammars and dependency grammars
are almost in a one-to-one correspondence with the nodes and links
of a conceptual graph.  Those are the grammars we use at VivoMind.    (05)

JS>John    (06)

RC>The LG nodes are particularly useful because there already exist very
precise and detailed descriptions of the meaning of each symbol in the LG
dictionary for English.  The developers were excellent in detail, in
engineering and in documentation - an amazing feat for academic
organizations.  The logic of how the graphs are constructed is properly
explained in clear English.  The algorithm for parsing, while not so well
explained, is understandable if given a few graphical support tools.      (07)

RC>So the LGP looks very leverageable.  I was thinking of having it take
English sentences in, along with the LGP parse word art, and put out
statements in a second, machine oriented language (SQL, CGIF, IDEF0, JSON,
or something simple and cellular like that).  The second language (CG,
IDEF0, JSON, CGIF, ...) is the one that the components of this English
sentence processor will mine for predicates that are true or false of this
situation/experience e[t0].  English is the only language processed by the
LGP in my imagined scheme.      (08)

To curate a database of LGP parsed sentences, this database should present
the best matching templates to a stimulus sentence against a database of
string templates using some distance measure.  For example, I will start
with the Levenshtein distance measure applied to parser categories and words
in sentences instead of applied to characters in strings.     (09)

Which brings me to the ultimate design issue; how do you enumerate the set
of all possible sentence templates?  The potential number of ways to vary
canonical sentences is enormous.   It quickly rises inconveniently past the
64 bit level if you do it the obvious way.  Like origami, the number of ways
in which English can be folded to say the same thing is enormous.  So the
problem as I see it is to fit <the world of English sentences likely to be
spoken in an application> within <the controlled English vocabulary> which
you espouse.      (010)

I want to represent my diagram of the sample sentence to embed this
discussion in specifics a little better, so I have repasted it below:    (011)

the quick brown fox jumped over the lazy dog
++++Time 0.04 seconds (0.04 total)
Found 2 linkages (2 had no P.P. violations)
  Linkage 1, cost vector = (UNUSED=0 DIS=0 AND=0 LEN=18)    (012)

 +---------Ds---------+              +-------Js------+
 |     +-------A------+              |   +-----Ds----+
 |     |       +---A--+---Ss--+--MVp-+   |     +--A--+
 |     |       |      |       |      |   |     |     |
the quick.a brown.a fox.n jumped.v over the lazy.a dog.n    (013)

By collecting all the arc annotation terms (Ds,Js,A,Ss,MVp,...) and node
annotation expression:    (014)

 ( jumped.v  fox.n (Ds(the,A(quick.a,A(brown.a,fox.n)))), 
            dog.n (Js(over,Ds(the(A(lazy.z,dog.n)))))
)    (015)

-Rich    (016)

Rich Cooper
Rich AT EnglishLogicKernel DOT com    (017)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (018)

<Prev in Thread] Current Thread [Next in Thread>