Don't Call it a Comeback: Semantic Technology and Search

Aug 11, 2009 - 4:45 pm 0 by

Below is live coverage of the Don't Call it a Comeback: Semantic Technology and Search from the SES San Jose 2009 conference.

This coverage is provided by Chris Boggs of Rosetta.

We are using a live blogging tool to provide the real time coverage. You can interact with us and while we are live blogging, so feel free to ask us questions as we blog. We will publish the archive below after the session is completed.

Don't Call it a Comeback: Semantic Technology and Search(08/11/2009) 
1:52 Chris Boggs:  Dana Todd from SEMPO is moderating
1:52 Chris Boggs:  Search and the Semantic Web - Nova Spivack from radar Networks is up first
1:52 Chris Boggs:  Semantics helps you get more precise searches and more precise targeting of ads
1:53 Chris Boggs:  Semantic is supposed to mena "meaning" what d
1:53 Chris Boggs:  oes this page mean, what does it "worry about"
1:53 Chris Boggs:  keyword search is "losing its edge"
1:53 Chris Boggs:  at a certain point it is not scaling, or growing up with the complexity of the results available
1:54 Chris Boggs:  shows graph that shows that as amount of data has increased,
the productivity of searching has decreased
1:55 Chris Boggs:  when web was smaller, google's approach could give you manageable results
1:55 Chris Boggs:  now there are way too many results to make it manageable
1:55 Chris Boggs:  "how do we save search"
1:55 Chris Boggs:  there is so much great information - how do we make sure it gets seen?
1:55 Chris Boggs:  one method: tagging - like meta data
1:56 Chris Boggs:   added by consumers or editors when someone mentions a jaguar do they mean a car or an animal...
1:56 Chris Boggs:  in higher semantic search, also trying to understand the logic of the page (not just the language)
1:57 Chris Boggs:  "find me used japanases cars with 300hp engines within 300 miles from my house for less than 30 lk""
1:57 Chris Boggs:  the search above would require "really hardcore semantic search"
1:58 Chris Boggs:  tags and definitions were set up by W3C
1:58 Chris Boggs:  this pushes the work away from AI
1:58 Chris Boggs:   to who makes the tags
1:58 Chris Boggs:  problem was people are lazy and dont want to put tags on the content
1:59 Chris Boggs:  technology is now enabling automation of this content. this makes the content more findable by Search Engines now paying more attention to these tags
1:59 Chris Boggs:  if you think about it, Sp
1:59 Chris Boggs:  ivack says, it's kind of like SEO
2:00 Chris Boggs:  he shows another timeline going from PC era to Web 1.0 to where we are as of March 2010: web 3.0
2:00 Chris Boggs:  will now briefly intrduce product: "Twine"
2:01 Chris Boggs:  "interest tracking" they add tags, and create little pages which then gets "SEO'd" with these tags (lol who knew it could be that easy ;)
2:01 Chris Boggs:  Twine 2.0 has new technology under devel
2:01 Chris Boggs:  opment, which will allow it to do true semantic search of the whole web
2:02 Chris Boggs:  system will analyze the site and crunch the data and create tags
2:02 Chris Boggs:  this will alow for a "new type of SEO"
2:03 Chris Boggs:  will allow people to search the web like a database.
2:03 Chris Boggs:  Where is semantics all going?
they view the web like a graph
2:04 Chris Boggs:  they combine the semantic graph with social graph, whcih gives a cool view of the world
2:04 Chris Boggs:  this view of the world provides a new way to look at
2:04 Chris Boggs:  

search, which Google, Yahoo, and others are also moving towards

2:05 Chris Boggs:  next speaker being intrduced by Dana:
2:05 Chris Boggs:  Lane Soelberg C
2:05 Chris Boggs:  

Coop Media Ven

2:05 Chris Boggs:  tures
2:05 Chris Boggs:  they are bringing wolfram alpha to market
2:06 Chris Boggs:  Talks about how Tim Berners Lee dreamed of th web becoming capabale of analyzing all the data
2:06 Chris Boggs:  on the web, and coming up with an answer for what you are looking for
2:08 Chris Boggs:  

He talks about W|A trying to do this

2:09 Chris Boggs:  The big assumption is that all systematic data known to man CAN be collected (and I missd the rest of the bullet point on the slide)
2:09 Chris Boggs:  talks about how W|A is "strong in math"
2:10 Chris Boggs:  40 million queries per month
2:10 Chris Boggs:  what is coming? he asked if anyone has seen the wood block mirror?
2:11 Chris Boggs:  it used cameras on blocks to compute "on the fly" the color blocks to retunr an image of the person walking up to it
2:11 Chris Boggs:  that is what they are trying to do with semantic search
2:11 Chris Boggs:  context and defining what a user means and tying it to data
2:12 Chris Boggs:  will ultimatel
2:12 Chris Boggs:  

will ultimately be defined by the "person walking up to th mirror"

2:12 Chris Boggs:  A
2:12 Chris Boggs:  PI release is coming, with more inputs and outputs
2:12 Chris Boggs:  will continue aggresive data
2:12 Chris Boggs:   curation process
2:13 Chris Boggs:  if semantic web is what we call the interfac
2:13 Chris Boggs:  e that gives us the answers we are seeking, then yes this is what we want
2:14 Chris Boggs:  he goes through a couple Wolfram Alpha type of searches combining data together
2:16 Chris Boggs:  wow there are 7 panelists on the stage...only three will be formally presenting
2:16 Chris Boggs:  room is about half full
2:16 Chris Boggs:  this is the first time I have tried this live blogging technology from coveritlive...pretty cool
2:17 Chris Boggs:  Albert Einstein said that imagination is more important than knowledge, says lane
2:17 Chris Boggs:  developers need to dream about the questions that can now be asked of this type of system
2:17 Chris Boggs:  in order to help it grow
2:18 Chris Boggs:  stopping for a few minutes for questions
2:18 Chris Boggs:  as a publisher being encouraged to do the data
2:19 Chris Boggs:  

Jonah asks about "what's in it for the publisher"

2:20 Chris Boggs:  Hansson from Google says that their objective in search quality in Google is to not try to give the answer on the search results page, rather give you the relevant pages
2:20 Chris Boggs:  Jonah was conceredn that f everything gets marked up then the publisher could lose traffic
2:21 Chris Boggs:  Lane says that the publishers will keep their data
2:21 Chris Boggs:  proprietary
2:22 Chris Boggs:  

next up will be Jason Meneyan from YieldBuild

2:23 Chris Boggs:  Jonah says that it is "va
2:23 Chris Boggs:  


2:23 Chris Boggs:  

stupid keyboard

2:23 Chris Boggs:  is the only answer that we forfeit our data
2:24 Chris Boggs:  in exchange for the possibility they will get traffic
2:24 Chris Boggs:  Jonah had said it was "vaguely reassuring
2:25 Chris Boggs:  Hackett feels that ultimately the publisher will have better control
2:26 Chris Boggs:  Hackett is from ask.,com
2:26 Chris Boggs:  they are trying to focus more on the semantic user query
2:27 Chris Boggs:  it is easier for a search egnine to understand the word "population" than the words "how many people live in"
2:27 Chris Boggs:  they call "semantic invariance" the problem of figuring out what the meaning is
2:28 Chris Boggs:  Haas is from Bing
2:28 Chris Boggs:  "Sea
2:28 Chris Boggs:  

"Search monkey" is trying to work on this semantic problem

2:29 Chris Boggs:  sorry Haas is from Yahoo
2:30 Chris Boggs:  the more data you can give us, the less we have to rely on our internal processes
2:30 Chris Boggs:  If Bing wanst to be a semantic engine, then there is lots of work we need to do on the semantic side, as well as the document understanding
2:30 Chris Boggs:  the focus of Powerset is to understand the documents better
2:31 Chris Boggs:  Johnson is from Bing and said above
2:31 Chris Boggs:  Google's Hansson focuses on "r
2:31 Chris Boggs:  

"Rich Snippets" effort going on the improve local search results

2:33 Chris Boggs:  Google also focuses on query semantics, but they "don't dwell on it"
2:33 Chris Boggs:  Jason Menayan is coming up to present now
2:34 Chris Boggs:  will be talking about a couple of tests they did on semantic seo, on a proeprty called "Hub pages"
2:34 Chris Boggs:  allows people to write articles and get paid for their distribution via revenue share
2:34 Chris Boggs:  why are they interested in semantic seo?
2:35 Chris Boggs:  to raise the content quality of their user/author base
2:35 Chris Boggs:  to get guidance on the types of informatoi
2:35 Chris Boggs:  information found in natural, high quality copy on a topic
2:36 Chris Boggs:  they use tools including adwrods tool versus other tools whihc provide more semantic variations, including "D
2:36 Chris Boggs:  


2:36 Chris Boggs:  

"Digger" and "Relevad"

2:37 Chris Boggs:  these show some different content that may be mentioned in high quality articles about a particular drug
2:38 Chris Boggs:  since the words refered to early research or other obscure references
2:38 Chris Boggs:  they tried taking some currently indexed pages not ranking well
2:39 Chris Boggs:  and adding the specific words. when the words were added to the content, all three pages tested ranked wihtin 2 weeks wihtin first 2 pages of Google
2:40 Chris Boggs:  they also tested with tagging using both Re
2:40 Chris Boggs:  


2:40 Chris Boggs:  Relevad and other internally generated keywords, and saw improvement with both
2:41 Chris Boggs:  semantic seo strengths:
2:41 Chris Boggs:  broadens content, which search engines reward
2:41 Chris Boggs:  Da
2:41 Chris Boggs:  na asks what does this do to make advertising better
2:42 Chris Boggs:  Soelberg says that if you are a holder of interesting structured data that is prorpitary and are willing to make pieces available, you could get great return in terms of traffic and interest
2:43 Chris Boggs:  Meanaya
2:43 Chris Boggs:  Meneyan sees this as a refinement of targeting
2:44 Chris Boggs:  this type of technology could help avoid seeing an ad for mrs Fields when someone searches for "clearing cookies from cpu"
2:45 Chris Boggs:  Noah mentions that Google already bought a company named Applied Semantics and is using this in ther AdWords system
2:46 Chris Boggs:  there are multiple standards currently for semantic markup or microformating, which makes it currently a highly fragmented space



Popular Categories

The Pulse of the search community


Search Video Recaps

Video Details More Videos Subscribe to Videos

Most Recent Articles

Search Forum Recap

Daily Search Forum Recap: May 29, 2024

May 29, 2024 - 10:00 am

Google Discover Feed Not Loading For Some Searchers

May 29, 2024 - 7:51 am
Google Maps

Google Business Profile Call History & Chat Going Away

May 29, 2024 - 7:41 am
Google Search Engine Optimization

Google Search Working To Surface Site When Site Name Is Referenced In Query

May 29, 2024 - 7:31 am

Two Conflicting Google AI Overview CTR Studies

May 29, 2024 - 7:21 am
Google Maps

Google Business Profiles Shows Product Rejection Details

May 29, 2024 - 7:11 am
Previous Story: How to Turn Your Web Analytics into a Money Making Machine