Print
Email
Print Edition Stories
Friday, January 2, 2009
Data crunching firms predict strong market in life sciences
By Brendan Lynch
MHT managing editor Rodney Brown on NECN: 22 tech companies give back to community
How is the economic crisis affecting tech companies?
MHT Editor Doug Banks on NECN: Decoding the Gen Y workforce
Cognika Inc., a startup with an artificial-intelligence computer model developed at MIT to organize large amounts of data, is testing its predictive technology on patient dosing with information from a Boston hospital.
The company uses a computer model to analyze numbers, text and images, according to CEO Ram Srivastav, who added that the company plans to incorporate the analysis of audio and video. Using that model, Cognika is partnering with Massachusetts General Hospital to prove its technology. Using patient records from another Boston-area hospital, Cognika’s model predicts how a patient would react to a certain dose of medication. That prediction is then compared to the actual record of the patient’s reaction. Using the model, a hospital could improve the speed and accuracy of dosing decisions, Srivastav said.
The model could also have applications on projects such as predicting a company’s revenue or the success of a marketing campaign. Ultimately, the model takes a client’s data and predicts what happens when a new event is introduced, Srivastav said, noting the larger the data set, the more accurate the prediction.
Cognika calls its technology “cognitive computing,” and the model is intended to mimic human learning, Srivastav said. The model uses semantic database technology, but Srivastav said he doesn’t like the term.
“‘Semantic’ covers the whole gamut, so it’s an oxymoron,” he said.
Cognika’s chief software architect, Shashi Kant, developed the model while at the MIT Sloan School of Management from 2004 to 2006. The two-man operation, founded in 2006, has been funded by friends and family so far, but is actively seeking venture funding, Srivastav said. The company uses contract developers in India and the United States, but it plans to hire about four people in early 2009, Srivastav said. The company is also looking for office space in Cambridge. Srivastav and Kant work remotely from Brookline and Cambridge, respectively.
Cognika isn’t the only local player using semantic technology to analyze huge amounts of data. Cambridge Semantics Inc. — counter-intuitively located in Boston — plans to launch three products later this month or early in February, according to vice president of technology and standards Lee Feigenbaum. The three versions of the company’s semantic database technology — a web version, an Excel integration and a middleware product — are intended to be complementary, Feigenbaum said.
Cambridge Semantics has made some changes since the 10-person company’s launch in March. It has changed the product name from Shape to Anzo, which means “four” in the Aladian language of the Ivory Coast, Feigenbaum said. Anzo is the fourth iteration of the software since Feigenbaum began working on it at IBM Corp., in the former Lotus Development Corp. building in East Cambridge.
After some “combination of decision making and circumstances being what they are,” the angel-funded company has also refined its target markets — initially life sciences, finance, government and insurance — to pharmaceutical companies and oil companies. Finance and insurance companies may not have extra cash for semantic technology, but biotechs and oil companies have plenty of clinical trial and oil production information, respectively.
“They have thousands of spreadsheets and no handle on how to get at it,” Feigenbaum said.
RSS Feeds
Comments
Please Login/Register to post comments.
No comments have been added or approved.