From: Jeff Bone (jbone@jump.net)
Date: Tue Jan 08 2002 - 08:47:29 MST
Ben Goertzel wrote:
> The degree of metadata/data asynchrony is going to be (roughly, perhaps
> nonlinearly)proportional to the degree of understanding
> embedded-in/attached-to the storage substrate
Fair enough.
> Until you can identify the details of this proportionality, you can't show
> that an incremental improvement in "AI" won't lead to a significant
> improvement in data/metadata asynchrony
Also fair.
> I think this is where your intuition, Jeff, differs from that of those who
> are really bullish on the Semantic Web.
(BTW, it's not *entirely* intuition. It's somewhat informed by an exercise in
"intelligent" news processing that consumed me for part of '99 and most of '00.
It's a somewhat informed opinion, admittedly based on a relatively small sample
of practical experience in trying to apply some of the concepts we're talking
about.)
> As simple examples of success, automatically marking up
> document by *topic area* can be gotten to work with 95%+ precision, and
> automatically marking up *sentences* by topic area can be gotten to work
> with 80% precision or so, assuming the sentences are grammatical (news
> articles, not message board text...). We could also automatically mark up
> addresses, phone numbers, prices and so forth with 95%+ precision. So a
> part of the Semantic Web vision is certainly achievable right now, should
> someone with appropriate power and $$ choose to commission the appropriate
> people to write the appropriate software.
All good points. None of this particularly concerns me, btw. What I'm more
concerned with is the idea that we're going to turn the Web into some huge and
fine-grained CycL KB. ;-) IMO, effort to do so is, modulo some high-level stuff
of interest as mentioned above, mostly wasted effort. (I.e., the Web (e.g. its
content) doesn't need a general concept of "man," something else ---
applications that use the Web --- needs specific notions of "man.")
Put another way, type is a slippery slope. Strong and weak typing don't coexist
with each other happily. If a critical mass of applications begins to expect a
certain level of structure embedded in Web content, then there will be a
negative incentive for content that does *not* embed such structure. That is
equivalent to *increasing* the friction of publication on the Web, and
undermines the very thing that made the Web so useful (and drove adoption) in
the first place. (Exercise: go back and cast a historical eye on hypertext
research before the Web. It seems amazing to me that we're rehashing some of
the same arguments. Ask yourself why the Web succeeded.)
> This partial success could be seen as evidence that the nonlinear
> proportionality mentioned above is not as severe as you're suggesting, Jeff.
Fair enough.
> > There
> > is no evidence to
> > suggest that information is stored / accessed / used in a
> > structured way in biological
> > intelligence.
>
> Actually, there IS a lot of evidence that information is *used* and *stored*
> in a structured way in the human mind. However, the structures involved are
> not the same as a simplistic semantic network.
True. By "structured" I meant the simplistic, reference-oriented semantic
network. The "structures" that the human mind seems to work with seem to be
much more fuzzy, abstract, and of a higher order than "structure" as its
typically dealt with in information systems.
> but it certainly
> demonstrates the existence of *memory structures*,
Sigh. Please let's not have a semantic argument about the meaning of
"structure." I'm specifically talking about relational arrangements, reference-
and graph-based structures, semantic nets, etc. as they are commonly known and
used in building information systems today. Information for which particular
metadata (often of fixed schema) is known or can be directly deduced from the
stored form of the information.
We probably agree on one thing: the kind of structure that the human brain uses
to store and access information is not anything that can straightforwardly /
efficiently be represented in a textual medium. ;-) I.e., the Semantic Web is a
graph structure because text markup is good for representing graph structures,
not because semantic networks are a particularly good model of human information
storage. :-)
jb
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT