Powerset’s search engine is expected to change people’s idea about search engines as it is designed to understand the meaning of Web pages, instead of using the conventional keyword technology. Powerset will offer a smarter and efficient experience and its beta version delivered good results, however with limited scope and index.
“We’re changing the way information is searched by doing a much deeper analysis of the pages we index,” said Scott Prevost, Powerset’s product director.
Usually keyword engines consider pages as word bags; only indexing their content without taking into account its meaning. Powerset’s engine uses technology developed in-house and licensed from Xerox’s PARC subsidiary. It creates a semantic representation by parsing each sentence and extracting its meaning. According to Prevost, meanings are indexed by Powerset instead of just words. Powerset assembles a collection of facts for each query and also summarizes the information found besides providing direct answers to factual questions.
Powerset’s index is quite limited now compared to the millions of pages provided from Wikepedia and Metaweb technologies’ Freebase. However, the volume is expected to grow in the following months.
Greg Sterling, an industry analyst at Sterling Market Intelligence, said in an e-mail interview, “Powerset’s capabilities are impressive. What they’ve created is both a better search engine for Wikipedia and a massive ‘proof of concept’ for their algorithm and technology.”
Powerset has to scale and deliver for several billion Web pages in order to serve millions of concurrent users. What Powerset has offered Wikepedia must also be offered to the entire Internet; an enormous task requiring resources and time. Although it needs more computational power, the indexing and retrieving of information pages will not be that difficult.
Although the company is a start-up technology firm facing several challenges, it still has investors providing the resources necessary to scale up the search engine to enable indexing of up to 20 billion pages.
The business model of Powerset is based on advertising, but the company plans not to serve ads in the beginning. The search engine will limit itself to Web search for now, and will be adding specialty engine for images, video, and other important topics like health, product reviews, and travel later on.
Shamila Janakiraman is a contributing editor for TMCnet. To read more of Shamila’s articles, please visit her columnist page.
