[Search-l] Delta

[Nathan Braun]

On 5/9/07, qiip at freesurf.ch <qiip at freesurf.ch> wrote:
>
> Some time ago I designed and implemented a prototype system (Delta)
> which allows users to rate search engine results[...]:
>
> 1) Ratings can be used to generate recommendations for certain queries.
> 2) Ratings can be used to learn a relevance function which takes a
> (user,query,document)-tuple as input and returns a relevance value.
> 3) Ratings can be used to cluster users into groups, a prerequisite for
> group-specific (personalized) search.


Would this function somewhat like a "Digg" of search (with up/down
ratings?)?  If so, I believe this would be a phenomenally effective
approach.

If so, this could also effectively address Fred Bauder's concerns:

-----Fred Bauder [mailto:fredbaud at waterwiki.info] wrote-----
"Let's think a little bit about how users fit into the scheme. I use google
quite a bit; I'm not going to use a start-up since I'm there to get
information, not fool around. I find good sources, and totally inappropriate
sources on most searches. I often click on useless sources and have to
backtrack. That information is not worth much as the source might be useful,
if I wanted the information that was in it. Then usually, I find a source
that has got the dope, perhaps it is so good that I bookmark it."

(ie, this could also work similar to Del.icio.us)

Such a feature -- a simple up/down voting mechanism --, if implemented
effectively with regard to search, could evolve into a nearly ideal hybrid
between Digg and Google... ie, a Digg-like up-down voting-mechanism for the
entire Internet and accessible to "everyone" (not just registered users like
on Digg??).  Such a voting capability, however, would necessarily need to be
a) personalized (so that your voting only directly applied to you and
influenced your own search results) and/or b) restricted access (to prevent
abuse).

I think that personalized search results along these lines would probably be
better since if we want our search to be effective we need to include
"everyone" and exclude "no one."  Spam-type results could effectively be
removed from the system by a community approach (similar to Wikipedia) where
only "recognized and accredited contributors" contribute to the overall
group's total datapool, in terms of ratings.   (In other words, people could
then vote on other users rankings/ratings, and if their contribution should
contribute to the "master algorithm" on the ENTIRE dataset)

I assume that there would "rise to the top" about 500-2000 "search gods"
(similar to the hierarchy of power and distribution already established on
Wikipedia and Digg, for example)-- who would effectively control the
majority of prioritized content, in an open source and transparent way---
again, very similar to Wikipedia in the grand scheme of things.... Ideally,
however, the "search gods" influence/control would be moderated -- more
effectively perhaps, than on Wikipedia -- in a truly democratic function -
where complaints against administrators would be truly logged and noted and,
if necessary, they could effectively be denied and barred from power if
their actions are deemed inappropriate by the masses.

I'm probably going too far down this road, in my humble attempt at
elaboration, but I can finally see fleshed out more fully in my mind, how a
ratings-based search system such as this would truly end up resembling
Wikipedia (vis-a-vis Digg and possibly Del.icio.us) in several key
respects.  We just want to make sure--whatever we do-- that we pre-correct
and counterbalance any of the excesses and problems of Wikipedia to the
greatest possible extent in our new search system...

If I find some time, I will write a "white paper"  containing more details.


In other words, I for one would greatly appreciate a white paper  fleshing
out the details of your proposal,  Joel. I think your proposal may be
exactly along the right lines for Wikia Search.

What does everyone else think?

Regards,

Nathan

Here's why I think this approach might be of interest for Wikia:
>
> 1) Such a system can serve as a complement to existing search engine (at
> least initially), rather than as a replacement. Hence gaining a critcal
> mass of users is less difficult.
> 2) The approach doesn't require the same amount of infrastructure as a
> full-scale search engine, since it suffices to store only ratings.
> Storing whole documents is not strictly necessary.
>



-- 
"Seriously, though, the Web is what we make of it. We have a powerful,
widely-deployed, largely uncontrolled communication network. It's up to us
to decide where to go next."
http://blog.outer-court.com/archive/2007-05-07-n78.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.wikia.com/pipermail/search-l/attachments/20070509/e06ab6c7/attachment.html