Lassoing relevant data with semantic search

"Most of the reasoner systems I tried died without even being able to produce a single answer. Now we've built a Scalable Highly Expressive Reasoner that, as far as we know, is the best reasoner on the market."

Most popular search engines do a decent job of turning up straightforward results to your query. Type “colon cancer,” for example, into Google and you will find references to definitions, symptoms and treatments. But what if you’re a clinician who wants to target a specific drug regimen at a patient with colon cancer in a discrete part of the colon? Google will not get you the information you need.

IBM researcher Kavitha Srinivas and her team have developed an allusive search system that uses reasoning instead of simple word matches to aggregate refined search data. This technology is playing an important role in clinical trials – and is slated for use in other data-heavy industries, such as financial services.

In this episode of Computer Science Spotlight, Dr. Srinivas gives an overview of her team's semantic search work at IBM.

Subscribe to Computer Science Spotlight.

Download the mp3 file (7 min., 20 sec.)	Download the transcript (Word Document)

Series producer: Barbara Finkelstein
Music: No Frills Salsa by Kevin MacLeod. Licensed under Creative Commons.

Last updated on July 17, 2008