Around the Net

Giving Search A "Semantic" Understanding

Search startup Hakia is creating what it likes to call "semantic" search, which means that Hakia's engine will try to understand user search queries rather than relying on statistical methods like linking, keyword analysis and user history to return relevant search results. Here's the idea: Search for "headache," and Hakia will presumably know enough about what you want to return "Tylenol," "Advil" or another brand name.

Right now, the only way for brands to get in front of "headache" is by buying keywords. However, this kind of understanding is a massive undertaking for a search company--it may be the key to the future of search.

Hakia's technology is so complex, its creators have spent years researching its foundation. The group has been at it for 36 months, having raised $16 million--all from institutional investors. If the technology works,it would open the door to whole worlds of esoteric long-tail content left to rot by Google and other engines' search crawlers, which rely heavily on counting and weighing links to content to compile results. Queries of three words or more would be particularly well-served by such a technology.



Read the whole story at Ars Technica »

Next story loading loading..