Powerset Launches Semantic Search Tools for Wikipedia
Posted on: Monday, 12 May 2008, 15:00 CDT
Silicon Valley start-up Powerset announced Sunday the launch of new tools that gives users a way to search millions of entries in Wikipedia's online encyclopedia, providing detailed answers to questions rather than isolated links requiring further research.
The move is seen as the first step towards challenging the Web search capabilities of industry leader Google.
The technology behind Powerset’s tools extracts the meaning of words and phrases into related concepts, which means that users do not have to type in the exact words they want to find. The tools analyze each sentence and entire documents, looking beyond words to comprehend conceptual relationships that to more accurately determine what a user is searching for.
For example, a person searching for how many wives King Henry VIII had (six, or two, depending on how marriage is defined) can find an answer through Powerset's service at tinyurl.com/5qpcr9/.
The 60-employee company is looking to leapfrog services such as those provided by Google Inc, Yahoo Inc, Microsoft Corp and IAC InterActiveCorp's Ask.com, whose current generation of services relies on keyword searches.
"The Wikipedia is becoming a microcosm of the most useful parts of the Web," said Greg Sterling, an Internet analyst with Sterling Market Intelligence, told Reuters.
"This offers a powerful way to find what you are looking for against this subset of the Web."
Powerset is trialing its Wikipedia search services to demonstrate how the underlying technology can be used to search other Web sites using natural questions or phrases. The company plans to partner with other sites, such as financial or patent filings, the CIA Factbook or Wikipedia-inspired clones, where information can be organized in a question and answer format that fits well with Powerset’s search techniques.
Along with its search services, Powerset plans to eventually generate revenue through advertising. For now, however, the company consists almost entirely of computer scientists and linguists, with only a handful of marketing and support personnel and no advertising staff.
According to Sterling, it would likely be years before the two-year-old startup could search the Web on the scale of Google, which uses statistical ranking techniques to find relevant Web links.
"What I don't know is how Powerset will perform on the wide open Web. In a sense, this is a massive prototype using the relatively structured information of Wikipedia. It is difficult to compare to what Google has built," he said.
Sterling said a bigger threat to Google would be if Microsoft were to acquire Powerset and integrate its capabilities into Microsoft’s search technologies. Microsoft recently withdrew a $44 billion bid for Yahoo. Had it succeeded, the merger would have created a rival to Google in Web search and online advertising.
"This could become the basis of a Google-killer," Sterling said. "Someone like Microsoft might want to buy Powerset."
Spokesmen for Microsoft and Powerset declined to comment on rumors of a potential relationship between the two companies.
Powerset offers fully annotated methods for searching Wikipedia entries for related concepts. Known as "Factz", these concepts generate summaries, outlines and automated answers to users' questions.
"Our system is a little more forgiving," Scott Prevost, Powerset’s general manager, told Reuters.
"It is not looking for hard-word matches. We are not searching for exact words, but concepts," he said.
The company licensed natural language processing technology and related machine processing methods developed over 30 years at Silicon Valley’s Xerox PARC research center to create new consumer Web search services.
With implicit approval of the non-profit Wikimedia Foundation, officials at Powerset said they are hosting a copy of Wikipedia's 2.5 million English-language entries on its own computers, something that will allow Powerset to make links across the full range of Wikipedia data.
"What Powerset is doing is offering readers a natural-language search interface, and we think that is an interesting experiment," said Mike Godwin, Wikimedia Foundation's general counsel, told Reuters, referring to how the two organizations would work together.
In addition to Wikipedia, Powerset's tools also search a related database by MetaWeb, another Web search startup, called Freebase. Barney Pell, Powerset’s co-founder and chief technology officer, said that natural language processing is at last going mainstream, after years of debate and research.
"2008 is the year that semantic and linguistic technologies cross over into widespread consumer use," he said.
---
On the Net:
Source: redOrbit Staff & Wire Reports
 |
 |
User Comments (0)
RSS Feeds