Search as an application, not a bolt-on

I saw a couple of interesting talks at the SUPA last week

The first was by David Harper at RGU: an academic project to show the advantages of using ‘piles’ as a metaphor for saving search results – each pile representing an aspect of the search topic. The piles could then be used by the search engine to find more, similar, documents, and the user could explore tangential regions without fear of losing existing search results. I use multiple tabs on Firefox to achieve the same result, so I can see how this is meeting a genuine need. I guess these piles could easily integrate with del.icio.us style tagged bookmarking as a way of saving and remembering clusters of search results (and sharing them too, but that’s another story).

The best description I could find of the work was here . Strangely, Google didn’t come up with any written-up papers I could find.

InQuira: intra-site search engine

The second talk was a corporate-level professional search technology presented by Bryan Sivak of InQuira. The search engine is explicitly designed for use within sites, so can take advantage of:

  • A known ontology combined with NLP techniques (InQuira restricts itself to a small number of vertical markets such as vehicles, mobile phones, banking)
  • Knowledge of the probable intent of the person doing the search which can be combined with the search term actually used to deliver relevant results. Intents are known because market research will have identified, say, the top 7 motivations behind visits to that site
  • Any known customer details, if the user has signed on to the site (eg search results can vary according to account status, allowing targetted marketing)
  • Direct access to underlying data and web components

Wherever possible, the search engine returns the answer, not just a list of links to possible answers – saving users having to click through a list. For instance, the response to “how do I change my password” would be the password-changing screen itself. The answer may be a result of a database query or pulling in a subscreen from elsewhere in the site – ie the results aren’t just based on document searches.

Questions I should have but didn’t ask included:

  • How easy is it to add another domain, e-government for instance
  • How easy is it to add another language. Or is this another inherantly english-only application?

There are a number of white papers (natch) on the InQuira site which I think will be worth a read when I get the chance.

The InQuira presentation really shows how NLP is being taken forward by commercial applications as part of the move from search being a semi-useful add-on to a corporate website, to being at the the core of understanding and meeting the users requirements.

Advertisements

About Peter Cruickshank

Lecturer in the School of Computing and a member of the Centre for Social Informatics at Edinburgh Napier University, Scotland. Interested in information systems, learning, politics, society, security and where they intersect. My attempts at rounding out my character include food, cinema, running, history and, together with my lovely wife, bringing up a cat and a couple of kids.
This entry was posted in information retrieval. Bookmark the permalink.

One Response to Search as an application, not a bolt-on

  1. Pingback: Entropy, search results and finding answers « Spartakan

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s