At the Aspire Venture Lab, collaboration is an important component to fostering big ideas. The power of collective effort is usually unmatched when it comes to affecting real change; and having a supportive network of brilliant machine learning experts and data scientists doesn’t hurt either.
Earlier this month we decided to truly put the Aspire community’s collective human and technological resources to the test with a little friendly competition. Last week we launched a predictive data modeling contest for everyone in the Aspire community that we’re calling the Battle at the Nostradome—named after the 16th century physician and astrologer Nostradamus who was famous for his predictive powers.
Seven teams, with members from five different Aspire ventures and others from Aspire’s research and science team, will be working on a single, exciting problem: how to predict glucose levels for diabetes patients.
It’s a problem that one of our ventures, Tempo Health, has been working on for the past year. They’re harnessing Aspire’s machine learning tools to create a predictive mobile app, BGNow, that will help diabetics manage their disease.
After getting some robust data sets from several patients at an adult Diabetes Training Camp in June—data that includes everything from continuous glucose monitor readings and insulin intakes to physical activity and consumed carbohydrates—they went on to build a working predictive model from that data earlier this year.
Now they’re going to take a 2nd pass at hardening their predictive model in competition with six other Aspire teams this month. That’s one of the advantages to being in the Aspire nest, ventures don’t face their challenges alone—they have a pool of resources, technology, and brilliant minds at their disposal.
“No data scientist is ever going to ask for less data,” Aspire data scientist Peter Funke likes to say. In the same vein, it doesn’t hurt to have a wider set of perspectives to approach a problem.
At the end of the competition, which will take submissions every week for the next four weeks, the best predictive models will be integrated into Tempo Health’s mobile app in time for an Alpha test in December.
But beyond meeting the challenge of building a model that can predict patient outcomes, the Battle at the Nostradome competition is also a way for Aspire to put its technological resources to the test.
The teams will be testing and optimizing their predictive models with Aspire’s A2i model-building platform. Another Aspire venture, Disfa, built the platform from the ground up to predict outcomes in financial markets. For Disfa, A2i was foundational to their company. Now it’s being used to manage a $20 million Hedgefund.
Now, we’re putting our A2i model-building platform to the test to see if the technology can be easily applied to our other ventures, while at the same time refining the tool with more user feedback.
Aspire’s model for nurturing new ventures differs from other venture capital firms, and we believe our method is the way of the future. Beyond providing our ventures with funding, staff, and infrastructure; ventures can also draw from foundational technology tools that were developed within the Aspire community.
“Aspire continues to be a leading-edge company in that our employees are using our own tool, A2i, for complex problem solving,” Chief Architect Thomas Rogers says. “The lessons learned will benefit employees and Aspire as a whole as we feed back improvements to A2i.”
Stay tuned for more updates on our data modeling competition in the coming weeks. The Aspire community is sure to gain valuable insights about the role our core technologies can play in building predictive model tools.