Home > Method, Monitoring > Semantic search engine – not your "average Joe" search task

Semantic search engine – not your "average Joe" search task

I still have a lot to learn about semantic search engines and the semantic web but I have few early observations about its direction.

I belive that Semantic search engines will do a lot of work in the back-end to help people with simple queries to find deep meanings.

I don’t see my dad going online and typing something like “where did US president’s kids went to school?”. Not because of the remote subject but because of the complexity. Most people type one or two words the most inside the search edit box. Maybe in the future he will ask such question using his own voice (see Nuance).

So, my assumption is that semantic search engines have different roles:

  • In the front-end
    • Research – A tool to to find patterns, trends and sets.  This tool should provide multiple visual ways to present the results (map, heat map, timeline, bubbles, tag cloud, 3D). Use case: I want to expose these and show them to the world in a clear visual way (example). I like to present them on my blog or web site (iFrame/Widget it). This is a great tool for researchers, news reporter, and bloggers.  TechCrunch, ReadWriteWeb, and others do present such data on their blog periodically – example. Again, this is not for the average user. 
    • Time saving: Operator is a Firefox extension that adds the ability to interact with semantic data on web pages, including Microformat, RDFa and eRDF.  You can use it to extract and export contact from a web page to your contact list, and event information to your calendar. It is still hard to find sites that supports microformats today (examples that do: Technorati – search page results, Google Map, LinkedIn – contact) but maybe Dapper will change this. You can now take an HTML page and automate adding microformats classes using this tool – let the tool find recurring elements in the origin page. Dapper does the “semantic work” for you. For more about this tasks see this smartly titled blog post: Does Tim Burners Lee’s Blog support Microformats? By the way Operator is a great way to check which web-sites supports microformats (use the Operator sidebar: from Firefox menu bar choose View->Sidebar->Operator).
  • As a back-end operation
    • Organize – I want my data to be linked using semantic techniques – see Twine .
    • Recommendation (discovery and sharing) – based on mining my data (and others) what you can tell me that: I don’t know that I don’t know.
    • Time saving: people have trouble with tags. In my opinion, the main reason is that it is hard to come up with forward thinking useful tags that will help to find data later, to match other tags and to help search engines to drive more traffic to the article. I can spend few good minutes thinking, should I tag it as: social network, social networks, or social-network. So, automating this process will help in many ways: more tagging, quicker tagging, more common tagging=more links and association. Twine today suggests tags that I can pick from, yet I think that this is just the beginnings. 
    • Money – in a way Google does this in AdSense – matching content, and target readers with ads. If you manage to do better job in this area there is a great potential. There are more places to add ads: people profile pages (not just in Facebook – I have a ton of profile pages across many social services, sigh), maybe in comments. 
    • Passive search – Alerts – at this point Google allow you to set alerts for keywords and Delicious for Tags. This is another way (as Chris Brogan says) to listen to the web. Maybe there is a way to improve exact keyword matching with associated content – same as semantic recommendation engines works. This could be also useful to organize the alerts results – I get alerts about some keywords in one long list. I’m not sure how many people are using alerts but I think that this is a great tool. This is by the way another marketing channel  – if I’m telling you what I’m looking for why don’t you “help me find it”? Alerts can become more sophisticated looking for patterns – example: more than 10 mentions of a word/phrase/product/company/my competitor in a day/week/month. I know that you can check that today on Google Trends but only for high volume terms, and it is not set as Alert – if you don’t look for it, it will not come to you.

Finally. Google today does a lot of “Microformat work” and some semantic discovery too, all behind the scene. You don’t ask for it explicitly but Google will still going to deliver it. When you search for ReadWriteWeb for example you’ll get along with the site link also the Contact, About, and Products information. If you’ll search for “movie near Lexington, MA” you’ll get what you need. This is too, support my assumption that you’ll not going to see sophisticated queries submitted by most people but the semantic web will be able to come up with better and more relevant answers to simple queries with complex meanings.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: