Clueray invites you to their beta project:

Welcome! The purpose of Clueray’s recommendation engine demo is to show how our patent-pending IntentMatch technology can take users beyond the limitations of keywords – beyond the large collections of mixed-intent, mixed-quality results delivered by conventional keyword-driven search – to focused, high quality results which meet a specific information need.

 

Figure 1. Who hasn’t slogged through page after page of mixed quality, mixed purpose results returned by keyword-based search? Clueray’s IntentMatch technology allows users to focus on those documents which meet a specific information need.

Clueray’s recommendation engine is easy to use – just enter your keywords like you would with any other search engine, hit the button, and you’re off! The “focus elements” below the search box will help you narrow down the result set to meet your specific information need.

Have fun, click around, and let us know what you think! You can contact us at feedback@clueray.com.

Also, make sure to come back often, because new features are being constantly added to the beta.

First Some Disclaimers (because we know you’re going to ask)

Yes, we know it’s currently kinda slow. Remember this is early beta – we’re looking for feedback on the content and interaction (expect it to change significantly over the next weeks and months) plus we’re building our index on the fly. This means we’re actually going out to retrieve each of the 20 (or whatever) documents and doing the analysis in real-time. For now, you should expect to see response times between 15 and 30 seconds for a brand new query, a few seconds (max) for a repeat query.

Yes, it can scale and will be (much) faster in the future. Really. The sum total of all our analyses currently takes a few hundred milliseconds on a single processor. The wait time you experience in the demo is almost entirely the network time required to retrieve the documents.

What it does:

In the first phase of our public beta, the user is able to focus the results based on the kind of document they are looking for, something we refer to as the document intent. Future versions will allow the user to add additional focus elements to more completely express their intent.

How it works:

Clueray analyzes each document and automatically identifies the “interesting” content on the page. Information (key features, specific content, etc) is extracted from the areas identified in the document and this is used to determine whether or not the document meets the specified information need and/or passes specifiec quality criteria. See our Technology Overview for a more detailed description of our unique technology.

What you’ll see:

Clueray’s results display highlights the ability of our IntentMatch technology to make sense out of large, unstructured collections of documents (like found on the web). It also highlights some of the analysis capabilities which are enabled by the kinds of things IntentMatch does. For example, the word cloud shown to the right of each search result (in the list view) shows the distribution of words in the “core content” area of the web page (eg. that part that isn’t the header, navigation bar or ads – the “interesting” stuff). It’s interesting to note that this word cloud doesn’t always correlate with the keywords chosen (because conventional search techniques don’t distinguish between “core” and “non-core” regions in a web page).

Of course we’ve also injected our opinions as to how we think a search interface should look (if you have opinions on this, we’d love to hear ‘em), flavored by our HCI mantra – there has to be a way for search to provide a heck of a lot more context before a user clicks on a link.

Finally, a sample results screen (using the default “List” view) is shown below. A sample results screen (using the default “List” view) is shown below. A description of the different results views can be found here.

 

This entry was posted on Monday, March 10th, 2008 at 11:45 am and is filed under Alts, Newcomers, Reviews. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.