RFI: DEVELOPING ALGORITHMS ON GEOSPATIAL DATA

REQUEST FOR INFORMATION: DEVELOPING ALGORITHMS ON GEOSPATIAL DATA The following text has been posted at the GEOINT Solutions Marketplace (www.geoint.community/problems) and cross-posted on FedBizOpps (www.fedbizopps.gov). PROBLEM: What marketable artificial intelligence would you develop with access to new libraries of geospatial data? What data would you need, and how might you commercialize your algorithms outside of the U.S. Government market? The National Geospatial-Intelligence Agency (NGA) is conducting a market research study on the demand for a public-private partnership (PPP) to produce artificial intelligence algorithms trained on geospatial data to help defend national security. Responses are requested by 11:59 PM Eastern Time on October 4th, 2017. AUDIENCE: Companies, universities, and other inventors that a) have a proven capability to create computer vision algorithms on geospatial training data, including but not limited to overhead video and still photography, and b) seek additional data sets for the purpose of enhancing existing algorithms or creating new ones, and c) would seek to commercialize or otherwise apply such algorithms in other market segments beyond the Federal Government. At this time, we are only looking for responses from organizations that develop algorithms in-house; we are not looking for partners to broker introductions to other companies. BACKGROUND: NGA proposes to invest geospatial data into the most innovative companies in America in order to drive intelligence, defense innovation, and spur the U.S. geospatial data economy. The agency would do this through the creation of a "data investment fund" which would be a private corporation, similar to a Federally-Funded Research & Development Center. Through the PPP, NGA would share data with U.S. inventors, companies and universities. These inventors, companies and universities would make a return on this investment by inventing new algorithms, creating artificial intelligence, or otherwise adding innovative value to this data. The benefit to the U.S. government would be to receive back licenses and government purpose rights on these new technologies. The benefit to the inventors, companies and universities would be to produce new commercial products. The benefit to the U.S. economy would be to exploit a government owned resource through tapping into the unrealized value of data for the public good and U.S. economic benefit. QUESTIONS: 1. About You: 1a. Organization name: 1b. Organization description: 1b. Are you a certified Small Business? 1c. Do you qualify for any SBA-recognized socioeconomic programs? If so, please list. 2. Existing Technology: 2a. What technology does your organization currently produce using machine learning on geospatial training data? 2b. What additional technology is your organization developing with geospatial data? Do you have prototypes? 2c. On what type of geospatial data did you train your technology (e.g. drone footage, hand-held photos, vector files)? 2d. How did you obtain the training data (e.g. open source, from a customer, self-produced), and how difficult was it to acquire? 2e. What is your past performance with other customers in this field? 3. Future Technology: 3a. What kinds of new training data might enhance the performance of your technology? 3b. What kinds of new training data might enable you to generate new technology? 3c. Where might you otherwise look for such data? Please submit responses to the GEOINT Solutions Marketplace at www.geoint.community/problems by 11:59 PM Eastern Time on October 4th, 2017. Contact Charles.E.Chapin@nga.mil with any questions.