Why Jesse's Jazzed About New Search Technologies

Search stinks. But if you're a regular, you've heard me say that before.

Search stinks. But if you're a regular, you've heard me say that before. Click for more.

So today I've got a heads-up about new search technology I think will be important. Potentially very important.

There's a move afoot to automate the search process. That means a computer figures out what a document is about -- "reading" the data and making sense of it. Thus doing a better job of categorizing and retrieving the information you're searching for.

There are a number of companies working in this field. Microsoft has devoted significant R&D resources to it. Here are three companies working on these new search technologies. Each takes a slightly different approach.

Autonomy software relies on a pattern-matching technology to identify and encode a "signature" of key concepts within free-form text documents. Then it searches for similar concepts in other text. This makes it easier on users, who can type what they're looking for in plain language. Alternatively, Autonomy can understand what users are interested in based on the type of content they read. Autonomy's new Kenjin service analyzes the content of what's being viewed, then surfaces related data from the Web or from info on your hard drive. Click for more.

Primus search software focuses on customer service functions, working with structured text -- or text in fields. So in some sense it is complementary to the Autonomy engine. It also lets users ask questions or describe problems using natural language -- capturing knowledge as questions are answered and problems solved. The information gleaned is used to solve new and recurring issues.

Dolphin Search utilizes research on how dolphins use echolocation to build complex information patterns, as Alex Lash explains in The Standard. The result is a search engine that employs neural networks to learn a specific bias or point of view. Click for more.

The IT folks in my office are most excited about the potential for comprehension based on neural network technology. Which means in a few years, computers may read and understand documents almost as well as we do.

Is that a good thing? Use the TalkBack button and tell me what you think. Or jump to my Berst Alerts forum and talk it over with other readers.


You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All