Media, Internet

Taking on Google: Is Semantic Technology the Answer?


A growing number of entrepreneurs and investors promise a next-generation Internet that is organized, simple to use and makes it easy for people to find things.

The so-called semantic web will be based on search technology that will enable people to type in their questions using everyday language. Developers of next generation search engines say their technology will “understand” the language within those queries—much more precisely than Internet giant Google’s technology.

The new search engines will then scour databases that have already “read” documents from across the web to find the best matches. Specific answers can be found for queries such as “Which is the best oil company in the world?” or differentiate between “movies by women,” and “movies about women.”

That’s the goal of a handful of semantic search startups that hope to change the way people look for information online. But their greatest challenge will be taking on Google, which vaulted to the forefront of the Internet search market by creating the simple but sophisticated engine that now processes roughly half of all U.S. searches.

The challenge may be great, but the rewards are worth it because search results these days go hand in hand with advertising dollars. Don Dodge, who heads business development for Microsoft's emerging business team, estimates that 1 percent of the U.S. search market is worth more than $100 million in annual revenue and $1 billion in market cap.

New York-based hakia has already launched its semantic search engine, with positive reviews for its ability to quickly find very specific information. The company also released a browser plug-in that lets users find exact sections of documents that contain the answers to searchers’ queries.

Search engines like Google rank search results mainly based on popularity without much understanding of what is on each page, according to Melek Pulatkonak, hakia president and COO. At hakia, "a lot more analysis has been done offline to extract meaning of what each page is about," she said.

The company is also building a "chat box" that will allow users to ask a question in an instant message format that hakia believes is more in tune with young people. hakia currently uses Ask.com's advertising system, but it is expected to launch its own ad platform next year.

The most hyped of this emerging group is San Francisco-based PowerSet, even though it has not yet launched its search engine yet.

It is backed by prominent investors such as Esther Dyson and the Founders Fund, it has earned a reputation for poaching Google engineers, and it has licensed sophisticated language technology from Xerox. Still, getting a computer to understand natural language—and find the appropriate search result—is a difficult process.

Skeptics say most people are happy with the average search query of about two words. “Generally, users are extremely lazy,” said Jakob Nielsen, a consultant on Web site usability. “The less work they have to do, the more successful Web sites tend to be.”

Barney Pell, CEO of PowerSet, conceded that semantic search engines must convince people to change their search behavior. But web surfers have been trained to use keywords on Google, even though that is not the way people think. So that behavior can change, he argued.

PowerSet this week launched PowerLabs, an invite-only program for users to test a demo built to search online encyclopedia Wikipedia. In one example, the program extracts facts about people and things in Wikipedia, such as Hulk Hogan. It also compares PowerSet’s results side by side with competitors to show the company’s progress.

Some companies are taking a more pragmatic approach by creating semantic products for more specific tasks. Startup Radar Networks—backed by Paul Allen’s Vulcan Capital and Leapfrog Ventures—has developed an information sharing and collaboration service that allows people to find and share web sites, photos, products, and other information with others. It is expected to launch later this year.

And already-launched AdaptiveBlue, backed by Union Square Ventures, has a widget that automatically searches for books, movies or music related to any web page that a person is viewing. (See “Semantic Technology Now.”)

For its part, Google has a number of engineers working on natural language, including noted expert Ramanathan Guha. And Google’s technology is already savvy enough to answer simple questions like “What is the population of Japan?”

However, the company’s philosophy has been: the simpler the better. “For the majority of questions, that's not what people want,” Peter Norvig, Google's director of research, recently told Technology Review. “They don't want the burden of having to express it as a full sentence.”

The startups have one key advantage: Google is rapidly pushing into new markets such as word processing, online payment systems, and mobile devices. These new markets provide higher growth—and more satisfaction for Wall Street—than rebuilding its existing search engine would.

That leaves an opening for upstarts – if they can provide users with a good enough reason to switch from Google’s powerful simplicity, said Greg Sterling of Sterling Market Intelligence.

“These engines need to create incentives to change and reward people for their behavioral change,” he said. “If (semantic search engines) deliver, people will likely respond.”

Comments

avatar
Semantic web is not only going to improve search, it's going to improve wikis. One of the largest semantic wikis (a directory of articles about individuals and corporations) is found at Centiare.com.
avatar
Finally the semantic web is becoming mainstream. You may also have a look at www.quintura.com. Best,
avatar
Thanks, Steve. What a Quintura.com user sees as a dynamic tag cloud, it is actually a semantic neural net. Please check how a Quintura search cloud works for a particular blog at http://www.seobook.com/quinturas-seo-book-search-cloud
avatar
At www.AltSearchEngines.com, we believe that equal attention should be given to the growing number and strength of the vertical search engines: Jobs, People, Health, Video, Blog, et al. What hinders them is the fact that nobody is or can be familiar with all of them; there are too many relatively unknown choices, and so they peck away endlessly at Google. What they lack is a common Interface which will guide users who are unfamiliar with these choices effortlessly and automatically. Badda Bing! The Next Generation of Search.
avatar
Semantic Technologies have a role far beyond 'simply' challenging Google's dominance in search, and although helpful in the examples you cite above, VC backing isn't a prerequisite. At Talis, for example, we're leveraging the benefit of over 30 years in the software business to pour those experiences into building a whole new technology Platform with semantic technologies at its heart. The point isn't 'simply' to improve traditional search. The point is to leverage all of those connections, inferences and contexts implicit in so much that people do online, and to make them manifest and actionable in delivering better applications. The applications won't necessarily be ours, but they'll call upon the Platform to enrich much that they do. And real, pragmatic, practical Semantic Web technologies lie at the heart of all of that... with not a venture capitalist to be seen.
avatar
http://www.askwiki.com is the first thing I've seen that truly has semantic search technology. It's a collaboration between AskMeNow and WikiMedia (the group that runs Wikipedia). It's still early but apparently it's built to 'learn' based on usage. I played with Powerset's closed beta integration to Wikipedia and this is way ahead.