By Barry Levine / CIO Today. Updated April 24, 2013.
Google is making an even bigger bet on semantic searching, where a search engine understands meaning and context. This week, the technology giant acquired personalized news feed start-up Wavii for a reported $30 million, which follows by a few weeks Yahoo's purchase of a similar start-up for the same price.
According to news reports, Google had been bidding against Apple for the company. Based in Seattle, the company has a couple of dozen employees and will be relocating to become part of Google's Knowledge Graph division at the company's headquarters in Mountain View, Calif.
'Make Facebook Out of Google'
Knowledge Graph is a Google project that is attempting to take another step toward efficient semantic searching by mapping various contexts for hundreds of millions of terms, many of which have multiple meanings that are usually understood by humans in conversation or communications but not as easily by search engines. Knowledge Graph is being rolled out in certain markets, and expects to build its knowledge base with feedback from search engine users on the desired meaning of given terms, such as which "chips" you're looking for.
Wavii's technology offers the ability to summarize stories, not unlike news aggregator Summly, which Yahoo recently bought for $30 million.
On its Web site, Wavii describes its product as allowing a user to "keep up with everything you care about in a personalized news feed." It automatically creates status updates for specific subjects, such as politicians, celebrities or gadgets, and posts the latest news, including a new app release or a political victory.
The company said it wanted to present news of the world in the same fashion that Facebook presents a news feed updating the status of friends. In other words, it said, its intention was to "make Facebook out of Google."
Facebook's status-updating news feed draws on literally hundreds of millions of unpaid workers who are manually entering updates -- that is, the social site's members. On the other hand, Wavii said it had to create software that taught computers to "read everything that is reported or shared" on the Net and to "automatically produce interesting social content about it," based on natural language research.
For instance, the company said its system intelligence has been designed to figure out which "named entity" a user wants to follow, such as the difference between George W. Bush and Bush, the band. To do that, it utilizes machine learning, where, instead of humans creating rules, a machine-learning algorithm is fed hand-labeled examples from which to learn the rules.
Yahoo received a great deal of attention for its purchase last month of Summly, which was headed by a 17-year-old in London. Summly and Wavii were seen as competitors, and both Yahoo and Google maintain heavily followed news sites.