May 19, 2009
Thanks for Asking! Ask.com is a Webware 100 Award Winner Three Years Running
Today Ask.com was recognized as a 2009 Webware 100 award recipient in the “Search and Reference” category, which specifically gives a nod to our innovative search tools and our technology that finds and organizes online data.
We’re grateful to the 630,000 consumers who cast their votes in this year’s competition, and to our engineers, product developers, and designers … the talented men and women behind the search innovations and products that have made Ask a Webware 100 winner for three years in a row.
We also want to thank Webware and CNET for putting on this annual quest to find the best of Web 2.0. We’re proud to be a part of it and to be in such stellar company. Congratulations to all the winners in this year’s competition!
--The Ask.com Team
Ask.com Searches Smarter
Ask.com’s very own Tomasz Imielinski, EVP of global search and answers, recently sat down with Jennifer Zaino over at SemanticWeb.com to chat about the newest technologies that search engines are using to find information on the web. As of late, one of the most popular discussion topics has been around semantic search and data extraction. With last year’s introduction of Ask.com’s DADS (Direct Answers from Databases), DAFS (Direct Answers from Search) and Answer Farm technologies, Ask.com is focused on improving how we search structured data in order to provide consumers with the best answers.
Check out this article titled “Ask.com Answers the Data Extraction Questions” from SemanticWeb.com. Tomasz shares his insights about semantic search today, and how the industry has only begun to tap into the potential this amazing technology holds for us. And if you find this article interesting, you’ll love the Web 3.0 panel titled Searching Smarter that Tomasz is speaking on this Wednesday, May 20th, at 10AM ET. Web 3.0 is being held at The New Yorker Hotel, 481 Eighth Ave., in New York City.
-- The Ask.com Team
Ask.com Answers the Data Extraction Question
By: Jennifer Zaino, SemanticWeb.com
When it comes to online search, you don’t know how bad you’ve got it. Or at least, how much better it can potentially be, one expert says.
But that’s just par for the course, though -- after all, there was a time when people thought the height of search efficiency was the Dewey Decimal System. Search engines’ heavy reliance primarily on the tagging employed by web site creators in anticipation of what they think people will be searching under leaves users in the position of having to do a lot of work before they -- or if they -- get the results they really want.
“What’s wrong about search today is that it is essentially driven by keywords. It’s an interesting situation, because when you ask people if they are happy with search, many say they are. A lot of people don’t realize what search could be,” says Dr. Tomasz Imielinski, EVP of global search and answers at Ask.com, who will be speaking at the upcoming Web 3.0 Conference in New York on the topic of smarter searching.
When it comes to more specific questions -- especially around subjects that might not be popular (almost anything other than, say, a celebrity’s name or these days, swine flu) -- people can spend hours trying to hit on the exact information they need. Imielinski says they’ve played a little game at Ask.com, wherein people have to type in as many different ways of phrasing a question that they can think of and hope to hit on the specific hidden phrase the game’s leaders have chosen. In fifteen minutes, around ten people generated 1,000 different legitimate ways of phrasing a particular question.
“Imagine a situation where today in search maybe a few of those phrases will actually get you a result,” he says. “The content is there, but it’s not ranked properly.”
Even the advent of semantic search engines hasn’t yet changed the productivity picture dramatically. Things still break down when you move from asking, for example, a fairly straightforward question such as, “what is the population of China?,” to skewing the query to something a little less keyword-friendly, such as, “What is the number of people currently living in China?”
“Nobody’s nailing it,” Imielinski says. “My main criteria is that any search engine that claims to be semantic should be invariant to phrasing.”
In fact, it’s veteran Google that Imielinski points to as continuing to lead in the unstructured data search market. “Google did a great job in pushing keyword search to unbelievable limits, with the massiveness of the web they have gone farther than anyone would have guessed,” he says, noting that the search engine often presents results that don’t contain a specific keyword but rather variations on it. It can catch results across rephrasings that the new search engines in the game still can’t, he contends, noting that they don’t do it through linguistics -- but they do it. “It’s more than blind matching, but to go beyond that is the next step.”
Ask.com is putting a focus on the structured data search problem, helping searches extract web data that is often not in text but in database tables and XML feeds where keyword searches don’t cut it. For example, a table might have data points around the words Toyota, Prius, and hybrid, and price, but if you ask most search engines to what is the price of a 2009 Toyota hybrid Prius that table won’t come up because those keywords aren’t together in the table.
“This is essentially a hidden database,” he says. “There are millions of these tables – it’s a goldmine of information. But it’s basically inaccessible.” Even if you somehow get to the page it’s up to you to navigate through what could be a lot of data rather than get pointed directly to it.
Ask.com has products that work on that data extraction problem in a couple of areas, based on its proprietary DADS (Direct Answers from Databases) and DAFS (Direct Answers from Search) technology. Ask TV Listings, for instance, captures time-based data to show searchers the answers to queries such as what football games are on tv tomorrow.
“This is something that is rooted to the understanding of time semantically,” Imielinski says. “If you ask a query which has the word ‘tomorrow’ in it in Google, you will get a key word match against ‘tomorrow,’” he says – it doesn’t understand that you are using tomorrow to put a time context around the query. Ask.com, which is the official search engine of NASCAR, also has launched a NASCAR search engine as part of its Super Verticals strategy to get people using its engine to search for facts on particular topics.
Searching structured data around a particular topic -- movies, tv, restaurants, etc. -- has the advantage of working within a defined channel and of being somewhat more predictable. “If you recognize the intention of the user you can provide a limited form of semantic web in that narrow application,” he notes. “We at Ask are doing a great job on structured data and we’re probably ahead of everybody.”
Opinions expressed here and in any corresponding comments are the personal opinions of the original authors, not of IAC Search & Media and may not have been reviewed in advance.
Blog Search from: Bloglines