Systemone’s Similarity Engine goes alpha

Semantic Web is the buzzword these days. TextDigger, Radar Networks, Boorah, PowerSet, Adaptive Blue are just some of the startups working on semantic web variations. Another semantic web contender Systemone today launched their first consumer facing Similarity Engine named Infolust. The initial launch is very basic. Right now all you can do is feed in the URL of a page for which you want more details, and Infolust will quickly analyze to come back with 10 related pages from Wikipedia. As I mentioned before the this is an alpha product so it is not yet ready for prime time. From the few search queries I ran, most of the results were way off track. So lot of work needs to done in that aspect, which even Systemone openly admits. Besides this, Systemone will be adding more content sources beyond the Wikipedia domain. Even if Systemone can get Wikipedia part working right, it will be big improvement over the standard Wikipedia search which is no one’s favorite.

Another similar product that I came across few days back is Watson from Intellext whose search results are completely and totally lost.

Links:
Infolust

Advertisements

4 Responses to “Systemone’s Similarity Engine goes alpha”

  1. Bruno Says:

    Hi Vivek,

    thanks for the posting! While the similarity itself works very well, we do have a hard time separating the actual content on a page from all the navigation / footers (thus the RSS links in your Reqall posting) etc. surrounding it. While Adwords do have a narrow vector, this screws the context when working with long text vectors on both sides. Currently there’s an, of course fully anonymous, analysis service running in the background and thus the more searches are conducted, the more it helps us improving the fetching process.

  2. Vivek Puri Says:

    Bruno,

    Apart from your bigger strategy that i have no clue about, you might have to keep in mind how common pages are built. Take the case if wordpress blogs, where the post content will “usually” be contained in a particular metatag. Similar might be the case for Six Apart’s blogging softwares. This can easily avoid the pains of removing “navigation / footers” from millions of blogs.

  3. derek Says:

    For better wikipedia search results check out http://futef.com – We offer basic keyword and advanced facetted navigation.

  4. Bruno Says:

    Definitely right, but in the current release we’re intentionally holding this back because it’s mostly about improving the general “content determination” capabilities. This stuff has to get smarter before we can customize it for specific types of pages like blogs, news etc.. A small factor are customizations and updates of blogging platforms, which could worsen things for a percentage of blogs and disrupt the functionality at all if changes are made which we don’t react to fast enough.

    But we’ll definitely add more specific content handling routines for blogs in the future, including microformat detection and other features.

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


%d bloggers like this: