Florian Cramer on Sat, 13 Dec 2008 16:33:59 +0100 (CET)


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: <nettime> Google cranks up its Engines of Consensus


On Saturday, December 13 2008, 01:34 (+0000), Andrew Orlowski wrote:
 
> Google this week admitted that its staff will pick and choose what  
> appears in its search results. It's a historic statement - and nobody  
> has yet grasped its significance.

Even the original PageRank algorithm, whose patent is owned by
Stanford University, is "pick and choose" in the end since it
implements a human choice of how to rank importance of information;
this is true for any such algorithm, be it Bayesian filters, Markov
chains or what have you. Google's breakthrough as a search engine was
based on the fact that a great number of users found its choice of
ranking more fit to their own subjective choices of ranking. Google's
claims that PageRank was somehow objective were nonsense from the
beginning on (and even more dubious for news.google.com) no matter
whether additional human intervention interfered with the ranking or
not. It's the old cybernetic pipe dream, and the old fallacy of meta-
and object level: a bot or a program is never "objective" only because
human - often political or economical - choices have been made on
the meta level of its design rather than on the object level of its
execution. The politics of computing more often than not lies in these
false beliefs (and make-beliefs) of objectivity. [And with the debates
on software culture, among others on this list, of the past decade,
one would think that we're beyond having to discuss this any further,
aside from the fact that it can be read up among others in Joseph
Weizenbaum's writings from the 1970s.]

Apart from that, Google "admitted" nothing new. The company
officially states that next to the PageRank algorithm, it uses
200 secret factors and methods for ranking search results
<http://www.google.com/corporate/tech.html>. Two months ago, Google
employee Scott Huffman wrote on Google's staff blog that since 2005,
he has worked "as an engineering director responsible for leading
search evaluation". He makes no secret about the following: "Human
evaluators. Google makes use of evaluators in many countries and
languages. These evaluators are carefully trained and are asked to
evaluate the quality of search results in several different ways. We
sometimes show evaluators whole result sets by themselves or 'side by
side' with alternatives; in other cases, we show evaluators a single
result at a time for a query and ask them to rate its quality along
various dimensions."

In the same posting he states that... 

   "today's search-engine users expect more than just relevance. Are the
   results fresh and timely? Are they from authoritative sources? Are
   they comprehensive? Are they free of spam? Are their titles and
   snippets descriptive enough? Do they include additional UI elements
   a user might find helpful for the query (maps, images, query
   suggestions, etc.)? Our evaluations attempt to cover each of these
   dimensions where appropriate." "Fourth, evaluating Google search
   quality requires covering an enormous breadth. We cover over a
   hundred locales (country/language pairs) with in-depth evaluation.
   Beyond locales, we support search quality teams working on many
   different kinds of queries and features. For example, we explicitly
   measure the quality of Google's spelling suggestions, universal
   search results, image and video searches, related query suggestions,
   stock oneboxes, and many, many more."

More interesting stuff can be found in the blog of another Google
employee who reveals how "[o]ur work on interpreting user intent is aimed
at returning results people really want, not just what they said in
their query".
[http://googleblog.blogspot.com/2008/07/technologies-behind-google-ranking.html]

Florian

-- 
http://cramer.plaintext.cc:70
gopher://cramer.plaintext.cc





#  distributed via <nettime>: no commercial use without permission
#  <nettime>  is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: http://mail.kein.org/mailman/listinfo/nettime-l
#  archive: http://www.nettime.org contact: nettime@kein.org