The Summit will host 11 virtual workshops via ZoomGov to provide insight and discussion into current focus topics and areas of opportunity. Complete agendas for each workshop will be posted closer to the event date.
Tuesday, October 19, 2021
Memory-Centric Computing
3:15-5:30 PM
Dr. Ali Keshavarzi, DARPA/MTO
This workshop explores the state-of-the-art research and discusses the research gaps for the development of low latency, high throughput memory-centric computing. Matrix multiplication, linear algebra and other critical math operations accelerate the required computing for artificial intelligence/machine learning (AI/ML) workloads. This workshop investigates extending the reach to solving a broader class of problems such as optimizations and probes a wider range of applications for such computing tasks. The goal is to enable future research toward creation of memory-centric compute technologies and architectures.
3:15-5:30 PM
Dr. Ali Keshavarzi, DARPA/MTO
This workshop explores the state-of-the-art research and discusses the research gaps for the development of low latency, high throughput memory-centric computing. Matrix multiplication, linear algebra and other critical math operations accelerate the required computing for artificial intelligence/machine learning (AI/ML) workloads. This workshop investigates extending the reach to solving a broader class of problems such as optimizations and probes a wider range of applications for such computing tasks. The goal is to enable future research toward creation of memory-centric compute technologies and architectures.
Next Generation Sensing and Signal Processing
3:15-5:30 PM
Dr. James Wilson, DARPA/MTO
Agenda | Slides
This workshop will discuss the motivation and challenges with pushing advanced signal processing further into the front end of wireless systems. Focus will be given to the area of passive sensing. The workshop will also publish results from DARPA Artificial Intelligence Exploration (AIE) programs, Signal Processing in Neural Networks (SPiNN) and Hyper-Dimensional Data Enabled Neural Networks (HyDDENN), that explore AI-based signal processing and wireless communications for DOD communication as well as commercial 5G/6G wireless applications.
3:15-5:30 PM
Dr. James Wilson, DARPA/MTO
Agenda | Slides
This workshop will discuss the motivation and challenges with pushing advanced signal processing further into the front end of wireless systems. Focus will be given to the area of passive sensing. The workshop will also publish results from DARPA Artificial Intelligence Exploration (AIE) programs, Signal Processing in Neural Networks (SPiNN) and Hyper-Dimensional Data Enabled Neural Networks (HyDDENN), that explore AI-based signal processing and wireless communications for DOD communication as well as commercial 5G/6G wireless applications.
UWBG Materials for High Performance Applications
3:15-5:30 PM
Dr. Tom Kazior, DARPA/MTO
Agenda | Slides
Future DoD systems require higher performance (e.g., higher output power RF devices, higher power handling switching transistors and protection circuits, high temperature transistors) to meet mission requirements. Just as SiC and GaN technology provided a leap ahead in capability over legacy Si and GaAs device technology, the Ultra-Wide BandGap (UWBG) semiconductors (e.g., AlN, cBN, diamond, Ga2O3) show promise as the next leap in RF and power electronics. However, these materials and associated devices are in their infancy. The objective of this workshop is to define technical challenges and develop community awareness of these challenges. The workshop will include an overview of potential applications and ongoing research in UWBG materials and devices, including new DARPA initiatives (UWBG seedlings, Heterogeneous Heterostructures (H2) Microelectronics Explorations program), applications and research funded by the service branches (Air Force, Army, and Navy), and also a commercial perspective. The workshop concludes with a discussion on the unique UWBG materials and device challenges, as well as suggestions for potential paths forward.
3:15-5:30 PM
Dr. Tom Kazior, DARPA/MTO
Agenda | Slides
Future DoD systems require higher performance (e.g., higher output power RF devices, higher power handling switching transistors and protection circuits, high temperature transistors) to meet mission requirements. Just as SiC and GaN technology provided a leap ahead in capability over legacy Si and GaAs device technology, the Ultra-Wide BandGap (UWBG) semiconductors (e.g., AlN, cBN, diamond, Ga2O3) show promise as the next leap in RF and power electronics. However, these materials and associated devices are in their infancy. The objective of this workshop is to define technical challenges and develop community awareness of these challenges. The workshop will include an overview of potential applications and ongoing research in UWBG materials and devices, including new DARPA initiatives (UWBG seedlings, Heterogeneous Heterostructures (H2) Microelectronics Explorations program), applications and research funded by the service branches (Air Force, Army, and Navy), and also a commercial perspective. The workshop concludes with a discussion on the unique UWBG materials and device challenges, as well as suggestions for potential paths forward.
ASIC Functional Verification
3:15-5:30 PM
Mr. Serge Leef, DARPA/MTO
Agenda | Slides
Brief presentations from industry and academia on the state-of-the-art and commercial needs in the ASIC functional verification space, followed by a panel discussion moderated by Program Manager Serge Leef that will encourage audience/panelist brainstorming on future functional verification challenges and solutions.
3:15-5:30 PM
Mr. Serge Leef, DARPA/MTO
Agenda | Slides
Brief presentations from industry and academia on the state-of-the-art and commercial needs in the ASIC functional verification space, followed by a panel discussion moderated by Program Manager Serge Leef that will encourage audience/panelist brainstorming on future functional verification challenges and solutions.
Best Practices for Measuring Security at the System Level
3:15-5:30 PM
Mr. Keith Rebello, DARPA/MTO
Agenda | Slides
As technologies for hardware and system security are developed, the question naturally arises: "How much more secure are systems that use this technology?" This question has implications not just for determining the success of R&D programs, but also for informing decisions by business leaders and policy makers when weighing the costs of incorporating new security technologies against the realizable security benefits. Currently, security improvements are approximated using synthetic benchmarks based on known attacks methods, red team evaluations that provide feedback from attackers, or formal methods to prove correctness of key parts of the system architecture. All of these techniques provide an incomplete answer to the question. This workshop will focus on establishing the benefits and limitations of various security evaluation mechanisms and exploring new approaches or combinations of approaches that might be used to evaluate security.
3:15-5:30 PM
Mr. Keith Rebello, DARPA/MTO
Agenda | Slides
As technologies for hardware and system security are developed, the question naturally arises: "How much more secure are systems that use this technology?" This question has implications not just for determining the success of R&D programs, but also for informing decisions by business leaders and policy makers when weighing the costs of incorporating new security technologies against the realizable security benefits. Currently, security improvements are approximated using synthetic benchmarks based on known attacks methods, red team evaluations that provide feedback from attackers, or formal methods to prove correctness of key parts of the system architecture. All of these techniques provide an incomplete answer to the question. This workshop will focus on establishing the benefits and limitations of various security evaluation mechanisms and exploring new approaches or combinations of approaches that might be used to evaluate security.
Wednesday, October 20, 2021
The Future of Open-Source Chip Design Tools
3:15-5:30 PM
Mr. Serge Leef, DARPA/MTO
Agenda | Slides
This workshop will review ongoing research and transition opportunities under the Intelligent Design of Electronic Assets (IDEA) and Posh Open Source Hardware (POSH) programs, including presentations on open-source transition, the role of commercial consortiums, ASIC engineering in a no-human-in-the-loop software environment, and government user perspectives on design automation. Following the presentations, a moderated panel discussion will encourage audience Q&A.
3:15-5:30 PM
Mr. Serge Leef, DARPA/MTO
Agenda | Slides
This workshop will review ongoing research and transition opportunities under the Intelligent Design of Electronic Assets (IDEA) and Posh Open Source Hardware (POSH) programs, including presentations on open-source transition, the role of commercial consortiums, ASIC engineering in a no-human-in-the-loop software environment, and government user perspectives on design automation. Following the presentations, a moderated panel discussion will encourage audience Q&A.
New Opportunities for Lifelong Learning Machines
3:15-5:30 PM
Mr. Ted Senator, DARPA/DSO
Agenda | Slides
DARPA's Lifelong Learning Machines (L2M) Program is now in Phase 2. This phase focuses on the development of practical systems that address complete lifelong learning solutions covering continuous learning, transferability and mission life sustainability as applied to various domains. The workshop will be an opportunity for the L2M Phase 2 Systems Group teams to describe their L2M systems and targeted applications, and enable attendees to better understand how to make use of outcomes from the program, transitioning L2M program results to deployed self-learning systems.
3:15-5:30 PM
Mr. Ted Senator, DARPA/DSO
Agenda | Slides
DARPA's Lifelong Learning Machines (L2M) Program is now in Phase 2. This phase focuses on the development of practical systems that address complete lifelong learning solutions covering continuous learning, transferability and mission life sustainability as applied to various domains. The workshop will be an opportunity for the L2M Phase 2 Systems Group teams to describe their L2M systems and targeted applications, and enable attendees to better understand how to make use of outcomes from the program, transitioning L2M program results to deployed self-learning systems.
Multi-Sensor Information Processing
3:15-5:30 PM
Dr. Timothy Hancock, DARPA/MTO
Dr. Ali Keshavarzi, DARPA/MTO
Traditional improvements in sensing have often relied on building "better" sensor hardware. Instead, this workshop will focus on the use of multiple sensing modalities and leveraging data fusion at the edge. Examples of multi-mode sensing include combinations of active and/or passive sensing such as lidar/EOIR or radar/EM spectrum. We see this today in commercial applications where lidar and passive imagery is used in vehicle autonomy and augmented reality in mobile devices, or the use of Bluetooth/WiFi for device tracking or contact tracing during the COVID-19 pandemic. Most of these applications require some amount of local real-time processing to present actionable information to a local user, but may also require a persistent cloud interface. This workshop will hear from experts in this area about the benefits of multi-modal sensing, the information processing requirements to fuse the data early, and specifically ask the question, "Are there innovations needed in processor architectures that are unique to fusing multi-modal data?"
3:15-5:30 PM
Dr. Timothy Hancock, DARPA/MTO
Dr. Ali Keshavarzi, DARPA/MTO
Traditional improvements in sensing have often relied on building "better" sensor hardware. Instead, this workshop will focus on the use of multiple sensing modalities and leveraging data fusion at the edge. Examples of multi-mode sensing include combinations of active and/or passive sensing such as lidar/EOIR or radar/EM spectrum. We see this today in commercial applications where lidar and passive imagery is used in vehicle autonomy and augmented reality in mobile devices, or the use of Bluetooth/WiFi for device tracking or contact tracing during the COVID-19 pandemic. Most of these applications require some amount of local real-time processing to present actionable information to a local user, but may also require a persistent cloud interface. This workshop will hear from experts in this area about the benefits of multi-modal sensing, the information processing requirements to fuse the data early, and specifically ask the question, "Are there innovations needed in processor architectures that are unique to fusing multi-modal data?"
Next Generation Mixed-Mode Microelectronics
3:15-5:30 PM
Dr. James Wilson, DARPA/MTO
Agenda | Slides
This workshop will review of ongoing research under the DARPA T-MUSIC program, including:
3:15-5:30 PM
Dr. James Wilson, DARPA/MTO
Agenda | Slides
This workshop will review of ongoing research under the DARPA T-MUSIC program, including:
- Efforts from foundries to develop advanced mixed-mode process technologies for integrated circuits combining both digital CMOS and advanced analog/RF devices
- Efforts to create state-of-the-art mixed-mode circuits using the above advanced technologies
- An overview of government and commercial needs in the mixed-mode microelectronics space
MEMS Next
3:15-5:30 PM
Dr. Ben Griffin, DARPA/MTO
Agenda | Slides
The DARPA Microsystems Technology Office (MTO) has a history of innovations in microelectromechanical systems (MEMS) that can be separated into three eras: (1) MEMS enablement, (2) MEMS-specific system solutions, and (3) MEMS as a tool. This workshop will briefly review DARPA MTO's history of MEMS investments and then discuss potential new topics to define the 4th wave of MEMS within DARPA MTO.
3:15-5:30 PM
Dr. Ben Griffin, DARPA/MTO
Agenda | Slides
The DARPA Microsystems Technology Office (MTO) has a history of innovations in microelectromechanical systems (MEMS) that can be separated into three eras: (1) MEMS enablement, (2) MEMS-specific system solutions, and (3) MEMS as a tool. This workshop will briefly review DARPA MTO's history of MEMS investments and then discuss potential new topics to define the 4th wave of MEMS within DARPA MTO.
DOE Office of Science Microelectronics Co-Design Research
3:15-5:30 PM
Dr. Andrew Schwartz, DOE
Agenda | Slides
This workshop is designed to provide an overview of a new Microelectronics Co-Design Research program at the US Department of Energy (DOE) Office of Science. Ten newly-awarded projects are led by DOE National Laboratories with participation from experts in academia and industry. The ten projects seek to increase energy efficiency and functionality in microelectronics, while embracing a co-design approach of multi-disciplinary collaboration that takes into account the interdependencies among materials, chemistry, physics, architectures, software, and algorithms. In addition, key findings from two prior co-design workshops will be shared. Finally, the workshop will conclude with a live discussion with the DOE Lab project leads intended to identify common interests and potential partnerships with the ERI community.
3:15-5:30 PM
Dr. Andrew Schwartz, DOE
Agenda | Slides
This workshop is designed to provide an overview of a new Microelectronics Co-Design Research program at the US Department of Energy (DOE) Office of Science. Ten newly-awarded projects are led by DOE National Laboratories with participation from experts in academia and industry. The ten projects seek to increase energy efficiency and functionality in microelectronics, while embracing a co-design approach of multi-disciplinary collaboration that takes into account the interdependencies among materials, chemistry, physics, architectures, software, and algorithms. In addition, key findings from two prior co-design workshops will be shared. Finally, the workshop will conclude with a live discussion with the DOE Lab project leads intended to identify common interests and potential partnerships with the ERI community.