DARPA 101: Engaging with DARPA

Transcription

1 DARPA 101: Engaging with DARPA Dr. Bill Regli, Acting DSO Director Discover DSO Day June 15, 2017

2 DARPA s Mission Breakthrough Technologies for National Security Precision Guidance & Navigation Communications/Networking IR Night Vision Stealth Radar Arrays UAVs 1960s 1970s 1980s 1990s 2000s 2010s Microelectronics: VLSI, CAD, manufacturing, IR, RF, MEMS ARPAnet/Internet Information Technology: timesharing, client/server, graphics, GUI, RISC, speech recognition Materials Science: semiconductors, superalloys, carbon fibers, composites, thermoelectrics, ceramics New capabilities require a healthy ecosystem across Service S&T, universities, and industry DARPA s role: pivotal early investments that change what s possible 2

3 DARPA Inside Stellar program managers Technology leadership Adventurous spirit Conviction and drive to change the world Active engagement with technology community Universities Labs Companies small and large Military services and agencies DARPA Culture Off-scale impact Risk taking Honor in public service 3

4 DARPA Technical Offices 4

5 DSO is DARPA s DARPA Accelerating breakthrough discoveries to create new enabling technologies for national security 5

6 National Security Challenges Informing DSO New physical, mathematical and social phenomena Asymmetric, commercially acquired, non-kinetic, and unconventional threats Existing structures for dominance and force projection are costly, complex and increasingly threatened Deficits in attention, imagination, and agility limit our options 6

7 Megatrends Transforming Science & Engineering Speed of scientific discovery is opening new opportunities to explore foundational questions Ubiquitous sensing and manipulation of the world provides new insights into understanding complexity Revolution in computing, data, and AI enables humans to change intractable problems to ones they can solve 7

8 Science in Service to National Security Threat Landscape New physical, mathematical and social phenomena Asymmetric, commercially acquired, non-kinetic, and unconventional threats Existing structures are costly, complex and increasingly threatened Deficits in attention, imagination, agility limit our options Investment Strategy New Foundations Develop the basis for altering our approach Complexity Engineering Shift or harness complexity to enable new approaches Noosphere Improve our understanding of people, society and human behavior Fundamental Limits Advancing traditional scientific questions and inquiry Science of Design Cognitive amplification to advance our human creative and decision making capacity 8

9 Program Managers Fariba Fahroo Mathematics Anne Fischer Chemical Systems James Gimlett Physics Jan Vandenbrande Math, Design, & Production Automation John Paschkewitz Systems, Design, & Materials Michael Fiddy Inverse Problems in Electromagnetics John Main Material System Innovation Adam Russell Behavioral/Social Sciences Ale Lukaszew Physics/Materials Predrag Milojkovic Vincent Tang Imaging & Optics Applied Physics 9

10 Focus Areas Physical Systems 2007 Ned Batchelder Math, Modeling & Design Human-Machine Systems Social Systems Credit: Detroit Institute of Arts The Economist, April

11 DARPA Program Gestation Ideas from inside & outside DARPA Seedlings, RFIs, & workshops Internal planning, wrangling & approval New Program! 11

12 Seedling vs. Programs Seedlings Usually submitted through DSO Office-Wide BAA Small short duration (6-9 months) projects Move concepts from disbelief to mere doubt May lead to the next generation of program ideas Programs Proposals solicited through specific DSO program BAAs Often multi-year, multi-disciplinary efforts Technology development to move from possibility to capability 12

13 How We Think: Heilmeier Catechism Important questions to consider when approaching DARPA with ideas: What are you trying to do? (no jargon!) How does this get done today? What is new about your approach? If you succeed, what difference do you think it will make? How long do you think it will take? Can your work transition (to the DoD or others)? How much will it cost? 13

14 DSO Recent Programs Accelerated Computation for Efficient Scientific Simulation (ACCESS) Transformative Design (TRADES) Extreme Optics and Imaging (EXTREME) Young Faculty Award (YFA) Gamifying the Search for Strategic Surprise (GS3) Agile Teams (A-Teams) RadioBio Computational architectures that will achieve the equivalent of petaflops performance in a benchtop form-factor and be capable of what traditional architectures would define as strong scaling for predictive scientific simulations of interest Develop/exploit new mathematics to incorporate advanced materials and manufacturing techniques into the design of solid parts and structures Optical systems capable of extreme performance and/or capabilities, which utilize Engineered Optical Materials (EnMats) Identify and engage rising stars in junior faculty positions in academia and equivalent positions at non-profit research institutions and expose them to Department of Defense (DoD) and National Security challenges and needs Apply a unique combination of online game and social media technologies and techniques to engage a large number of experts and deep thinkers in a shared analytic process to rapidly identify, understand, and expand upon the potential implications and applications of emerging science and technology Discover, test and demonstrate generalizable mathematical abstractions for the design of agile human-machine teams and to provide predictive insight into team performance Establish if purposeful communication via electromagnetic waves between biological systems exists, and if it does, will determine what mechanisms are involved and what information is being transferred BAA release: 5/6/2016 BAA release: 5/11/2016 BAA release: 8/24/2016 RA release: 9/29/2016 SN release: 11/30/2016 BAA release: 12/5/2016 BAA release: 2/15/2017 Molecular Informatics Explore new approaches to store and process information with molecules BAA release: 3/31/2017 Ground Truth Topological Excitations in Electronics Lagrange Use artificial but plausible simulations with known causal rules (aka ground truth ) as testbeds to validate social science modeling methods Explore ways to overcome fundamental limits in conventional electronic memory and logic by using topologically protected states Develop novel mathematical methods on both the theoretical and algorithmic fronts to solve high-dimensional, dynamic, data-driven optimization and decision-making problems BAA release: 4/28/2017 BAA release: coming soon BAA release: coming soon 14

15 DSO Recent RFIs Nanoweaving (RFI) Assessing the state of the art in nanoweaving and nanobraiding Released 5/11/2016 Fundamental Limits of Learning (RFI) Delta (RFI) Forensic Social Science Supercolliders (FS3) (RFI) RadioBio (RFI) Molecular Code and Compute (RFI) Rapid Screening Technologies in Organic Synthesis (RFI) Foundations for Generalizable Representation and Automated Extraction of Knowledge (RFI) Innovative Concepts for Multiresolution Interactive Wargames (RFI) Metamaterial-based Optical System Design (RFI) What are the fundamental limitations inherent in machine learning systems? Leveraging multiple sources of individual differences and variability in order to better understand and design for the complex, changing, and dynamic interactions among individuals and their social groups, including teams and organizations. Developing interactive simulations that can be used to calibrate the validity of different social science research methods and tools in drawing strong inference about causal mechanisms that can lead to emergent complex behaviors in human social systems. State-of-the-art in experimentally observing, modeling and understanding communication between biosystems exploiting electromagnetic (EM) waves from low frequencies (khz) to high frequencies (THz) Exploring ideas and research that directly couples information storage and computation using non-natural, sequence-controlled polymers in a generalizable, adaptable format that may lead to fundamentally new computational architectures in the future Research concepts or existing technologies that enable fully automated, rapid screening of reaction conditions across synthetic organic chemistry Methods for representation and extraction of knowledge in scientific data analysis and modeling to support computational reasoning over diverse sources of evidence Scalable, interactive gaming or wargaming approaches simultaneously spanning a large number of space and time scales with the goal of assessing a wide range of possible competitive outcomes and strategies using a range of human decision making strategies Design and optimization of metamaterial-based optical imaging and nonimaging systems Released 5/12/2016 Released 8/22/2016 Released 9/6/2016 Released 10/17/2016 Released 12/5/2016 Released 1/31/2017 Released 3/17/2017 Released 4/7/2017 Released 4/18/

16 Some Recent Seedlings Data-driven Inverse Design Paradigm for Part Qualification in Additive Manufacturing Advanced Reservoir Learning Interoperability Challenges and Scenarios in Computational Design Theoretical Foundations for Design of Human Machine Systems All Together Now: Superradient Laser and Clock High-level Provably-Correct Control of Robotic Swarms 16

17 Three Ways to Engage with DSO Talk to a Program Manager (PM) PMs here today and tomorrow /phone/face to face throughout the year Submit ideas to the DSO Office-Wide BAA (HR001117S0040) Concepts New Ideas Seedlings: Disbelief Mere Doubt Respond to DSO program BAAs Programs: Possibility Capability Find DARPA announcements at and 17

18 We Look Forward to Your Ideas 18

19 19