uiux
Regular
Department of Defense
Fiscal Year (FY) 2023 Budget Estimates
April 2022
Air Force
Justification Book Volume 1 of 4
Research, Development, Test & Evaluation, Air Force
Vol−1
---
Title: Passive Electro-Optical/Infrared Sensing in Contested Environments
Description: Develop innovative passive optical sensing technology to support surveillance and reconnaissance in contested environments. Develop high performance focal planes, aperture technologies, sensing architectures, and imaging techniques capable of long range target detection and characterization for intelligence, surveillance, reconnaissance and air-to-air sensing.
FY 2022 Plans: Conduct flight test evaluation of the staring infrared search and track system against military relevant targets. Complete flight testing of compact, low-cost, low volume, real-time hyperspectral sensor for attritable platforms in preparation for operational demonstration. Continue development of low-earth orbit sensing systems for critical Air Force needs, including event-based sensors and passive interferometry.
FY 2023 Plans: Continue refinement of advanced processing algorithms for hyperspectral imaging. Conduct demonstration of low-cost, compact hyperspectral imaging sensor with on-board, near real time processing software that utilizes advanced processing algorithms under development. Perform testing of new multi-spectral cameras and filters that allow more compact designs. Continue development of low-earth orbit sensing systems for critical Air Force needs, including event-based sensors and passive interferometry. Perform a field demonstration and evaluation of an event based/neuromorphic sensing system. Finalize development of large format, long wave infrared detector array for infrared search and track in preparation for future testing.
Title: Processing Technologies
Description: Develop automatic and dynamically reconfigurable, scalable, affordable distributed peta-flop processing technologies for real-time global information systems. Starting in FY 2021, the remaining non-cyber work that was performed under Project 625319, Cyberspace Dominance Technology, in the Processing Technologies effort within this PE will now be performed within this effort.
FY 2022 Plans: Advance the application of novel neuromorphic systems for robust machine learning. Continue to advance research and development of the neuromorphic processor and validate capabilities for dynamic learning on mobile and power-constrained platforms. Initiate the development of a prototype integrated with existing embedded high performance computing systems. Commence the development and delivery of a Neuromorphic High-Performance-Computing (Brain-in-the-Box).
FY 2023 Plans: Continue to advance the application of novel neuromorphic systems for robust machine learning. Continue to advance research and development of the neuromorphic processor and validate capabilities for dynamic learning on mobile and power-constrained platforms. Decrease the development of a prototype integrated with existing embedded high performance computing systems. Continue the development and delivery of a Neuromorphic High-Performance-Computing (Brain-in-the-Box).
---
Project 625319
Title: Advanced Architectural Technologies
Description: Develop the architectural mechanisms that form the basis for predictable software and high assurance systems.
FY 2018 Plans: Continue research and development of a cyber hardened processor for embedded weapon systems. Develop a runtime environment that can monitor and maintain a trusted and resilient envelope of operation. Continue research and development on neuromorphic processing technologies to assess the feasibility of autonomy on mobile and power-constrained platforms.
FY 2019 Plans: Continue research and validation of a cyber hardened (robust, secure) processor for embedded weapon systems. Continue applied research to create trusted and resilient embedded systems that are capable of identifying, localizing, and automatically repairing previously unknown and/or unintended vulnerabilities. Continue research and development of the neuromorphic processor and validate capabilities for dynamic learning on mobile and power-constrained platforms. Initiate development of software using evolutionary approaches to make embedded systems tolerant to unexpected and unforeseen situations.
Title: Processing Technologies
Description: Develop automatic and dynamically reconfigurable, scalable, affordable distributed peta-flop processing technologies for real-time global information systems.
FY 2018 Plans: Research and develop a novel neuromorphic system for visual object detection using Google's open source deep learning framework, TensorFlow. Continue research and development to establish the memory-based network nodes, to further evolve and adapt the photon-based interconnects, and to develop an integration scheme to interface a quantum network with the existing freespace optical link between the Air Force Research Laboratory, Information Directorate laboratory facility in Rome and the Stockbridge remote test site.
FY 2019 Plans: Continue to research the application of novel neuromorphic systems for robust machine learning. Continue research and development in the area of supreme and quantum computing information sciences to establish the memory-based network nodes, to further evolve and adapt the photon-based interconnects, and to develop an integration scheme to interface a quantum network. Test the ability to teleport quantum information between network nodes, and to establish two-way quantum communication between two memory nodes. Conduct an analysis of conventional/quantum channel interface for long-distance communication
---
What I find interesting to the above information is the specific mention of "Dynamic Learning" eg.
"Continue to advance research and development of the neuromorphic processor and validate capabilities for dynamic learning on mobile and power-constrained platforms"
"Continue research and development of the neuromorphic processor and validate capabilities for dynamic learning on mobile and power-constrained platforms"
BrainChip has previously tried to trademark this phrase:
DYNAMIC LEARNING Trademark - Serial Number 87252419 :: Justia Trademarks
Computer hardware and software systems for artificial general intelligence, using spiking neural network devices that connect to dynamic synapses, and are capable of autonomous learning from input streaming data
trademarks.justia.com
