درباره این اپیزود
Cybersecurity has spent decades in a reactive cycle: find a vulnerability, patch it and wait for the next one. Dr. Patrick Lincoln joins Frank Cilluffo to discuss DARPA's effort to break that cycle by building systems that are inherently secure, private and resilient from the start. They explore lessons from the AI Cyber Challenge, the role of mathematical proof in eliminating vulnerabilities and the challenge of making complex systems trustworthy even when individual components fail. The conversation also examines how research can move beyond isolated demonstrations to strengthen the technologies and infrastructure society increasingly depends on.
Main Topics
- Preventing and providing technological surprise
- IPTO's return to computer science foundations
- Inherent security and privacy
- AI-enabled vulnerability discovery and patching
- Formal methods and mathematical proof
- Trust and resilience in megasystems
- Moving research into real-world use
Key Quotes
"This back and forth or cat and mouse game has been going on a long time. I'm getting tired of that. And so we're trying to find new foundations to build inherently secure systems and inherently private systems." — Patrick Lincoln
"We need power tools to let people do more, better, faster, with quicker reaction time and higher assurance that what they do actually does lead to really fixing the problem and not creating new problems, not causing harm to the system." — Patrick Lincoln
"Software is difficult. And so how can you get to high assurance for a complex system involving the analog systems, the sensors and actuators, the digital systems and hardware, and the digital systems and software? And there are processes for this, some of those involving mathematical proof and formal methods." — Patrick Lincoln
"If you've got a million subsystems, there will be failures, and perhaps some of them even maliciously so. ... By building these compositional—think internal firewalls—within a complex system, and ways that we can understand the emergent properties of a large collective of systems, ... we can then predict and therefore give assurance about its behavior long term." — Patrick Lincoln
"My favorite answer to this problem, the world's most urgent critical problem, I'll say the DARPA answer is improving our collective ability to solve urgent critical problems." — Patrick Lincoln
Links and Resources
- DARPA Information Processing Techniques Office
- Patrick Lincoln biography
- DARPA AI Cyber Challenge
- PROVERS formal-methods program
- Resilient Software Systems Capstone
About the Guest:
Patrick Lincoln is director of the Information Processing Techniques Office at DARPA, where he leads work spanning artificial intelligence, cybersecurity and privacy, and resilient complex systems.
Before joining DARPA, Lincoln held senior research leadership roles at SRI, an independent nonprofit research and development institute, where he led multidisciplinary work across computing, cybersecurity, artificial intelligence, and advanced systems research.
Lincoln holds a Ph.D. in computer science from Stanford University and a B.S. in computer science from MIT.
رونوشت 🔗
جستجوی اپیزودهای گذشته
قسمت های دیگر در این پادکست
