Home News Events DARPA seeks ‘radical innovation’ in data analysis

DARPA seeks ‘radical innovation’ in data analysis

US: The Defense Advanced Research Projects Agency (DARPA) is planning an unusual eight-week residency programme in which teams of researchers will brainstorm over ways to use visual and geospatial images to identify people, places, objects, and activities.

The project, called the Innovation House Study, is being done in conjunction with George Mason University. Teams qualify for up to USD 50,000 in funding—USD 10,000 in startup costs and USD 20,000 for each of two phases. DARPA said that the work will be done in a “short-fuse, crucible-style environment.”

“We are inviting a new generation of innovators to try out ideas in an environment that encourages diverse solutions and far-out thinking,” said DARPA programme manager Michael Geertsen in a written statement. “If this model proves to be as successful as we believe it could be, it represents a new means for participating in government-sponsored research projects.”

The goal of the programme is to design and demonstrate “radical, novel” approaches to extracting meaningful content from visual and geospatial media, including the software to make it work. Teams will be given access to unclassified data, including aerial and ground-level video, high-resolution LiDAR (light detection and ranging) of urban and mountainous terrain, and unstructured amateur photos and videos. Teams can also use public data obtained from open-source and commercial repositories, the Web, and mobile phones.

The programme will emphasise collaboration among participants, rather than competition, according to DARPA. Teams will have access to experts from academia and defense and intelligence agencies.

Participants must be US citizens and at least 18 years old. There are no specific academic or professional requirements, but teams may be asked to submit a statement making their case to be included in the experiment. Selected teams will be announced in August.

The residential sessions will take place in two 4-week sessions, beginning in September and ending in November. In the first session, teams will design and demonstrate proof-of-concept software capabilities. Those that advance will work on developing functional software.

Source: Information Week