Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

Re: Semantic Searches - Knowledge Engines

__/ [www.1-script.com] on Monday 10 October 2005 16:09 \__

> Roy Schestowitz wrote:
>> Given all the data which is contained in our E-mail box(es), our files
>> (photos, documents and sounds to name a few), wouldn't inference be the
>> natural direction to follow? Extracting the semantics from our data and
>> forming a network of knowledge will enable us to search for answers
>> rather than text that resembles our query. As for personal search
>> (limited scope due to privacy), the implications can be particularly
>> revolutionary.
> So, you think an ?intellectual thumbprint? of a person can be useful?

To /ourselves/ it may be, not to others (_NOT_ public). Data that is
contained among E-mails, for example, has been read already so no new
information would crop up in our minds. Rather, pieces would be put
together more cohesively. It's like a memory aide to oneself. 

> To
> whom, if I may ask? If you *are* the person your system is profiling, you
> already have pretty good knowledge of who/what you are (leave pathology
> aside for the sake of the argument). Privacy implications are immediate
> and obvious, even the actual ?limit? that you are referring to, is
> debatable. There may be VERY good reasons to leave some (most?) parts of
> your life outside of public domain of knowledge.

Not public domain. Think about tools like Google Desktop, putting aside the
security hacks that toy had in the past.

> Who gets to decide what?s
> indexed and what?s not? What about people not familiar with the system
> that accidentally let it index things not intended to be indexed?

I think that's an entirely different issue. You are referring to people who
participated in UseNet, for example, oblivious to the fact that text would
often 'leak'.

>> Rather than "search engines" we will be talking about
>> "knowledge engines"
>> (Googlism is maybe worth citing). There is currently no barrier that
>> stands
>> in the face of implementation apart from computer power and code. With
>> Open
>> Source software, I think it will soon be achieved.
> Well, on the other hand, you make it sound like it?s a mega-reminder
> system you?ve conceived. Like, you?ve already had that piece of
> information, but it eludes you, so you need an aid to find it. Well, in
> this case it could be useful.

The only daunting aspect as this, as often we see in life, is that so-called
'robots' become more capable than humans, which can shatter self-worth. I
bet this is how a group people felt when Deep Blue became a Chess champion.
It was inconceivable in the past that anything could perform than a wise
man. What about photography and painting...?

>>Just imagine a neural network out there which rather than contain text
> your name >as got complex knowledge about who you are.
> Umm? once again, myself and the closest family pretty much make up the
> complete list of parties I would ever want to have knowledge of who I am.
> I don?t need no stinking network having my ?intellectual thumbprint?
> available to advertising agencies, political parties, law enforcement,
> IRS, Greenpeace, well, you go ahead and complete the list of who?d want to
> know not only my name but my way of life, too.

This, in fact, is why inference based on the World Wide Web (in its present
form) can be dangerous. Some things on the Web are public although they
should not be. Some would list Google Earth's high-resolution maps and vile
political content as examples.

> I have enough trouble with
> too many people knowing my social security number and my home phone number
> as it is.
>>Moreover, it can answer questions
>> that involve you and is too complex to be explored by a human.
> If it?s too complex to be explored by a human(?), I do not want to be
> involved because I would not be able to understand what it means to me.
> Therefore I would not want to donate any personal data into this project.
>>To many
>> bodies including governments this would be invaluable, so be sure it's
>> headed that way.
> Yeah, way to go, Roy! Things are already headed that way, anyways, so bend
> over before it?s too late!

When I wrote this I was merely making a prediction. Sooner or later somebody
might attempt to do that because the financial incentive is there. I loathe
the idea just as much as you do, but what I thought would be valuable is:

-Public domain: a collection scientific facts (think of chemistry for
instance) where you can form a knowledge pool that's easily accessible
using queries.

-Private (personal scope): a "life manager" as Jeffrey Hawkins called it.

> You are conceiving a system that may, eventually, become more dangerous
> than Google. How?s that for a statement? Google is already indexing much
> more data than is required to provide ?knowledge? in its academic sense.
> It can already be used for any kind of site penetration or at least doing
> an extensive research plotting a penetration attack. Credit card numbers
> accidentally indexed, social security numbers, account numbers ? you name
> it, Google has it. Personally I don?t think Google has developed a set of
> morals yet that would lead them through managing these problems. They are
> pretty much letting it be this way: you left it visible to our bot ? you
> are responsible. This leaves out a basic argument: if there were no bot,
> the data won?t be compromised.

Speaking of which,


Look at that worrying smile...

> So, if you did not think hard about implications, think again. You still
> have time to do it right (if of course there is such possibility).
> Good luck!

When I think of the public domain, I still have this mental picture of
someone looking for a way to solve a mechanical problem with some
electronic device. A query can solve this just as well as a proficient
person in a call centre.

On a personal (non-public level), I am thinking about the ability to ask
questions like "when was my grand-grandmother born"?

I don't see anything evil in the scenarios above apart from the fact that
our human mind become futile and interaction with people less necessary.


Roy S. Schestowitz      | Warning 0x12C: ispell feels tired
http://Schestowitz.com  |    SuSE Linux    |     PGP-Key: 74572E8E
  2:00am  up 46 days 14:14,  2 users,  load average: 0.22, 0.55, 0.64

[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index