Picture Credit to New York TimesFor the past several years the biggest name in search has been Google. Everything about them seems charmed: from innovating search, improving e-mail, to buying out Youtube. The number of innovations coming out of the Googleplex doesn't seem to be stopping anytime soon; Google has recently introduced Google checkout, and is committed to developing web versions of desktop software programs. But with so many things on its plate, will it neglect the core technology that made the company a household name and a verb?
Now, out from Silicon Valley, a new group of dreamers, flush with cash from venture capitalists, are setting out to do the seemingly impossible; beating Google at its own game. Not only that, they are also trying to achieve one of the holy grails of computing: getting it to understand natural language.
It's been over 50 years since, Professor Alan Turing, developed his eponymous test of machine intelligence which goes as follows; if a human judge cannot distinguish between who is the machine and who is the human in a natural language conversation with the both of them, then the machine has passed the test. So far, not one program has managed to fool a human.
The usefulness of a program that could understand queries in plain English was not lost on the founders of Ask Jeeves (now just Ask.com). With more and more information available on the web, there was a need for quick answers, which would be provided by an user-friendly search engine that could understand everyday language. In practice, Ask Jeeves didn't quickly make the cut, but it was a start at least. A new crop of search engines are trying to pick up the mantle, a decade after the launch of Jeeves. Since 2004, nearly $350 million of venture capital money has gone toward at least 79 start-ups that have something to do with search. Not one has gained more than 1% of internet search share, and still others are still in beta or pre-beta, but that hasn't diluted the optimism of some people.
You can't try out the search features just yet (the company figures that it won't be ready for the public till at least the end of the year), but that hasn't stopped it from raising $12.5 million dollars from prominent venture capitalists and capturing a lot of buzz. It just made a deal with Xerox Corporation's research spin off, PARC (Palo Alto Research Center) to license its technology which is the result of 30 years of research in the field of natural language technology. Barney Pell, co-founder and chief executive of Powerset is very bullish about his prospects. He seems like a smart guy, his Ph.D is in machine learning and games, and he helped develop the AI for NASA's Deep Space 1 spacecraft. Unlike Craig Silverstein, a technology director for Google, who believes that natural language search will be a long time coming, Bell thinks that within five to ten years people will say: "remember when we used to search using keywords?" Another Google head, Marissa Meyer, vice president of search and and user experience doesn't expect it to happen in the next five years. I guess we'll just have to wait till Powerset lifts the veil of secrecy around its program, till then its anyone's guess. You can read Pell's blog here.
Hakia recently received $16 million in funding; however, it is currently undergoing beta testing so that means anyone can test it out. I tried asking, "How can I graduate college without any debt?" but couldn't get a satisfactory answer, though it delivered good answers to the question, "Why is the sky blue?" There is also a Hakia blog accessible here. Other Contestants
Another recent entrant is Lexxe (pronounced leksi, like lexicon), which is based out of Australia. It also clusters search
results like another search engine, Clusty. Clusty is choosing a different way to make searching more efficient by grouping together search results into categories. So far it hasn't generated as much buzz, but the fact that it's in beta, and working is promising.References and Related Links
0 comments:
Post a Comment