Tutorial K

Agenda Tutorials

Tutorial K

 A Primer on Adversarial Machine Learning (AML)

Fees

 $250 USD each

Date

 Monday – February 24, 2025

Time

 1:00 PM – 4:30 PM PT

Overview

 Adversarial machine learning (AML) is a field at the intersection of machine learning (ML) and computer security (cybersecurity). While ML model development prioritizes accuracy and efficiency, AML focuses on model security and robustness. This tutorial introduces core AML concepts, relevant governance documents, the taxonomy of attack techniques, and mitigation strategies.

Outline:

    • Introduction
        • Overview
        • Key Concepts
        •  Generative Artificial Intelligence
        •  Testing Methodologies
    • Background
      • Data Modalities
      •  Stages of Learningo
      • Deep Learning Architectures
      •  Example Generation
    • Governance
      o Information Security Models
      o Ethical Artificial Intelligence
      o Trusted Artificial Intelligence
      o Adversarial Machine Learning Frameworks
    • Plans and Roadmaps
    • Attack and Defense
      o Goals and Objectives
      o Capabilities
      o Techniques
      o Transferability
      o Mitigation Strategies
    • Applications and Best Practices
      o Tools and Libraries
      o Development and Deployment
      o Metrics and Benchmarking
      o Case Studies o Hands-On Scenarios
    • Conclusion
      o Lessons Learned
      o Emerging Threats
      o Resources

 
Instructors Ronald Nussbaum, Brian Tung, Andrew Brethorst, Nehal Desai, Dominic Berry, The Aerospace Corporation

Biographies

 Ronald Nussbaum is a Project Leader in the Software Implementation and Integration Department at The Aerospace Corporation. In this role, he leads and supports a variety of programs related to artificial intelligence, data architectures, computer wargaming, and other areas. A selfdescribed generalist, Ronald’s research interests span the breadth of theoretical computer science. Possessing a strong affinity for hands-on programming, he has a passion for applying machine learning techniques to complex big data problems. Ronald holds a B.S. degree in computer science from Grand Valley State University, as well as M.S. and Ph.D.
degrees in computer science from Michigan State University.

 

 Brian Tung, is a Senior Engineering Specialist in the Data Science and Artificial Intelligence Department at The Aerospace Corporation. He has led projects in machine learning, data science, natural language processing, distributed systems, and network security. He is also the
administrator and “show runner” for the ARG-Math discussion group, a biweekly Aerospace colloquium sharing a variety of mathematical topics with the broader Aerospace community in an entertaining and engaging way. In his spare time, he plays jazz piano and teaches music theory. He is an avid amateur stargazer and authored an article on special relativity for Sky and Telescope magazine

Andrew Brethorst is a Senior Engineering Specialist in the Data Science and Artificial Intelligence Department at The Aerospace Corporation. Mr. Brethorst completed his undergraduate degree in cybernetics from UCLA, and later completed his master’s degree in computer science with a concentration in machine learning from UCI. Much of his work involves applying machine learning Techniques to image exploitation, telemetry anomaly detection, intelligent artificial agents using reinforcement
learning, as well as collaborative projects within the research labs.

Nehal Desai is a data scientist in the Data Science and Artificial Intelligence Department at The Aerospace Corporation who works on national security space (NSS) applications. Prior to Aerospace, Nehal worked at Los Alamos National Lab, Silicon Graphics, and IBM. Nehal has a Ph.D. in Mechanical engineering from North Carolina State University. In his spare time Nehal enjoys raising hedgehogs and skeet shooting.

Dominic Berry is a cyber security researcher with the Cyber Assessments and Research Department at The Aerospace Corporation. Before coming to the company Dominic earned his B.S. in computational modeling and data analytics with a focus in cyber security and cryptography at Virginia Tech. Outside of work Dominic enjoys board games and spending time with his wife and son.

Description of Intended Audience and Recommended Prerequisites

 This tutorial is intended for developers training, testing, or integrating artificial intelligence (AI) models, particularly generative AI (GenAI) models; developers designing, building, or maintaining data repositories; cybersecurity professionals; and personnel managing or providing oversight to such programs.

What can Attendees Expect to Learn

 Attendees will gain knowledge of adversarial machine learning concepts, an understanding of the importance of
cybersecurity in all aspects of the generative artificial intelligence (GenAI) life cycle, hands-on experience with realistic scenarios as both attacker and defender, insight into what trends they should be monitoring, and a list of resources if they wish to learn more about AML
Tutorials