A world of a million search engines â€“ created not by some mega-power but by the community at large, and with the aid of semantic search technologies. That’s the world envisioned by one start-up presenting at DEMO Fall 2010 today.
Semantifi wants to be the power behind those engines. To be clear, Semantifi isn’t encouraging folks to use its engine to create the next Google. Rather, it says it can bring a platform to the community to go beyond what Google does, enabling them to create applications that let others search inside the data sets within the Deep Web that fascinate them, or that they think will be a draw for others, or most likely both. “We think keyword crawlers and [traditional] search doesn’t work for data,” says Shree Pragada, founder and CEO. You may get tons of links delivered that have word matches for a Google search on “California senator earmarks,” but even though that data is out there in the Deep Web, you won’t get a table of results showing who got earmarks for what drawn from searching the data itself.
Using its step-by-step process for uploading a data set, which includes reviewing and confirming the meta-data Semantifi assigns to it, “those who know data and have access to it can very quickly configure a data set to this search engine and share it with the world,” he says. “You locate the data set, define the structure and then publish it as an application that can be free or as a paid application if you think it’s commercially valuable,” Pragada says. Semantifi explains the data structure to its search engine. When users ask natural English-language questions or use keywords on its portal, Semantifi can understand the terms, look for relevant information across published applications, and present to the user the relevant applications that will answer that question. “The platform searches all search applications, then says this is the best way to present [the particular] information,” the CEO says.
Technically, Financially Speaking
The company describes the full battery of its service as a “meaning based search platform to search both structured and unstructured content” across the ‘search stack.’ It incorporates a knowledge-based driven semantic search engine; automatic data summaries and visualizations to interpret database results; and the ability to index multi-terabyte databases for fast results and crawl any type of database to compose complex queries.
Providing its capabilities to publishers as a service, says founder and CFO Sanjay Gokhale, is the difference between spending a half-million dollars and long periods developing a custom app and site for searching data on the Deep Web that’s stored inside databases and which most search engines can’t see, and “democratizing search accessibility to data.” Anyone can use its centrally hosted search platform for free. It will monetize its effort through advertising for the most part, as it expects most applications to be developed for the site also to be available free to users. It will get a revenue share from developers who create for-fee search apps, and it also plans as a revenue source fees from private applications, where content providers use its technology to publish applications on their own web sites.
Semantifi expects government apps to be hot with its platform, as more people want more transparency in this election year (for more on that see Prep for Mid-Term Elections, With Linked Data’s Help . It’s already created some apps on that front among those available for users to try out, including:
â— Recovery, a search app detailing how the Recovery funds have been spent and the status of the projects;
â—An app to search government spending at the national, regional, state and county levels;
â— And of course the app to search on the aforementioned earmarks requested by the US lawmakers.
Speaking of the possibilities of opening up data within the Deep Web, Pragada says, “We are excited about all this data but we can’t do it all as one company. it has to be driven by the community.”
• Last chance: Don’t forget to propose your startup for our Semantic Web Impact Awards. The deadline is Sept. 15.