By Tom Dunlap and Jennifer ZainoSAN JOSE, Calif. -- The fifth annual
SemTech 2009 Conference kicked off here yesterday, with a slew of companies set to make announcements at the five-day show.
The conference is off to a good start, judging by the sheer number of companies and people crowding the halls and the lunch tables at the spacious Fairmont Hotel. Many product announcements or upgrades are due this week, and Wednesday's Search Day is a much-buzzed-about portion of the conference.
Also on Wednesday, a last-minute panel was added, featuring two of the biggest applications in the semantic web: Twine's Nova Spivack will interview Wolfram Alpha's business development guru, Russell Foltz-Smith. They'll discuss the recent launch, what Wolfram Alpha is (and isn't), and where the project is headed from here.
Unlike other semantic web conferences, "here we have everyone," said Sally Khudairi, who heads up communications for SemTech 2009. "We have the academics, the researchers, the entrepreneurs, the technologists, the evangelists, and the venture capitalists. They're all here. It's nice we can do that."
Among the announcements so far is news from Thomson Reuters, which has upgraded its OpenCalais service. The list of new features in 4.1 includes "The Recession Pack of Facts & Events" -- a catchy name for tuning that has been done to extract a new set of facts and events related to company performance and company actions in a down economy, including accounting changes, labor issues, layoffs, earnings restatements, delayed filings and more.
"It's the bad news pack," says Thomas Tague, Calais Initiative lead, Thomson Reuters. "Hopefully people won't need that one too long."
What they should need for the long-term is one of the most significant features of the upgrade, Social Tags. The service has used heavy-duty natural language processing to generate tags and now adds the ability to go beyond news categories to include more "human-like" tags to mix and match with semantic ones. Imagine, for example, OpenCalais extracting from an article about Porsche and BMW and a race between two makes of these cars, and perhaps the location of the race.
"But from a consumption point of view, what do I really want to tag that--sports cars or racing, or a human-like tag," he says. "Social tags give you back those human-like tags to mix and match with semantic tags as well." What's exciting about this is where this can take things in the future: Today it can leverage knowledge bases such as Wikipedia to this end, but Thomson Reuters has access to very large database of other types of content, such as all the patent filings ever and reams of scientific information. "We're starting with open data assets but will branch out in the future to include others," he says, bringing on board domain-specific social tagging that would expand OpenCalais' reach into richer publishing domains such as life sciences.
The service now also supports Spanish entity extraction, opening up its use to about 500 million more people. It's also been engineered for greater reliability and scalability.
Other briefings at the conference this week will come from Top Quadrant, Collibra, Expert System, hakia, iQuest Analytics, Metatomix, Ontotext, Saltlux, STI International, SYSTAP, and zAgile.
Tune in to
SemanticWeb.com and
InternetNews for full coverage.
Tom Dunlap reported from San Jose. Jennifer Zaino reported from Long Island.
Comment and Contribute
