George Mason University

Abstract:
“All your machine learning models are belong to us…” Machine learning models can be tricked using tiny data alterations, sometimes as small as a single pixel for image recognition. What happens when malicious data is used to produce a different outcome than expected? Is it possible to detect malicious data? Machine learning models can be reliably reversed engineered without direct access to the model. What happens when malicious actors can fully replicate your machine learning models? This presentation and discussion will address those and other questions around how we can protect our AI and ML systems from the enemy.

Abstract:
Dr. Saunders will lead a discussion regarding the implication of IoT for national security. Real world case studies will be discussed with a specific focus on securing IoT systems in a large scale distributed environment.

Abtract:
We will discuss advancements in AI expert systems, SDP, Spectrum agility, big data, IoT and Cloud in C4I for Integrated Mobility. The future of mobility will have a far more consolidated, cognitive, integrated and transformational impact on defense, security and next generation of first responders whereby new incarnations of ‘armed force’ start doing different things in novel ways. COTS implementation to the tactical edge will enhance decision-support system during wartime/emergency wherein the time is a significant factor to communicate with warfighters, responders, autonomous vehicles, weapons, machinery, emergency equipment, etc. that need support with intelligence and quick response in the zone. The current trajectory of technology, ubiquitous connectivity, sensors, location-based services, social networks, machine learning, big data and IoT will have an even more profound impact on Mobility Integration. These present risks but also opportunities especially benefitting the Defense and Government to become SMARTER!