2017 Year in Review
2017 was another banner year for IARPA. Here are the highlights:
- 9 new multi-year research programs covering diverse technical fields, including biosecurity, event forecasting, space situational awareness, homomorphic encryption, human identification, and superconducting electronics
- 13 single-year seedling studies researching topics, such as Bloom filters, a dual comb spectrometer, and infra-red imaging sensors
- 4 new public prize challenges offering cash awards to innovators across the globe
- 48 technical workshops with over 3,000 attendees
- Over 190 peer-reviewed publications from IARPA-funded research
- 34 transition agreements to transfer IARPA-funded technologies to other government agencies
- Over 250 research publications from current and past IARPA programs, uploaded to the Defense Technical Information Center registered user site. This includes more than 70 publications available to the public through DTIC’s public user site
IARPA-funded research highlights:
The Babel program has become a goldmine for speech scientists, with over 50 scientific publications citing use of Babel data in 2017 alone. A total of 757 Babel speech data sets have been distributed to 136 different organizations with a goal to continue to advance speech technology research under low-training conditions.
The SILMARILS program demonstrated trace explosive detection capabilities beyond the program’s difficult targets. Phase 1 of the program developed record-setting component technologies for hypercube acquisition speed and power, as well as wavelength coverage and flatness for infrared supercontinuum sources. For this breakthrough, and others, the SILMARILS PM, Dr. Kristy DeWitt, was awarded the Intelligence Community’s Award for Individual Achievement in Science and Technology.
One of the FUSE program’s software and models for prediction of emerging technologies is now available to the public. Meta is a tool that helps researchers understand what is happening globally in science and shows them where science is headed.
The MORGOTH’S CROWN prize challenge used machine-learning approaches to develop algorithms that improve chemical detection on complex surfaces and in cluttered environments.
The QEO program demonstrated the feasibility of quantum machine learning by using a quantum annealing system to train multiple classes of artificial neural networks for image recognition and reconstruction tasks.
The N2N prize challenge collected fingerprint data from 354 human subjects during a week-long event, producing over 40,000 rolled fingerprint images and 30,000 black powder and chemically processed latent fingerprint lifts. In addition to facilitating evaluation of challenge prototypes, this data will be made available to the research community to improve latent fingerprint matching.
The EMBERS event forecasting system, developed under the OSI program, is now a commercial offering of Virginia Tech Applied Research Corporation. Organization can now purchase the system’s daily forecasts for a particular set of countries.
The fMoW prize challenge provides a satellite imagery dataset with one million points of interest annotated to researchers and entrepreneurs, enabling them to better understand satellite imagery through novel learning frameworks and multi-modal fusion techniques.
The interagency team of IARPA’s Finder and NGA’s HUNTER programs won the Intelligence Community Science and Technology team award. The IARPA/NGA team developed a revolutionary approach to geolocation of arbitrary ground level images. The application has been successful world-wide.
The Janus program has made significant advances in face recognition. The technology is now capable of finding the correct subject (rank 1) 95% of the time on a challenging gallery with over 1 million subjects.
The MICrONS program delivered the largest-ever maps of neocortical circuit structure and function to inform the development of next-generation machine learning algorithms. In parallel with this work, MICrONS performers launched seven startup companies with over $30M of private funding to commercialize technologies developed through the program.
The HFGeo program has advanced ionospheric modeling by developing and applying new assimilation algorithms. Ionospheric data assimilation is a technique to evaluate the 3-D time-varying distribution of free electron density, using a combination of physics-based modeling and observations.
The LogiQ program’s efforts to achieve an error-corrected quantum bit (aka, a qubit) helped to advance quantum computing technology. This led to new quantum-algorithm demonstrations, an improved understanding of noise and errors in these systems, and realization of systems with more than 15 qubits.
Research data sets from ACE and OSI programs have been posted to Harvard University’s Dataverse Project. This information is now publically searchable and easily citable in the world’s largest collection of social science research, and has been downloaded over 700 times. Two papers authored by researchers not part of the original IARPA-funded program were published in 2018 using this data.
IARPA media highlights: