Agenda

Tuesday, November 27, 2018

8:00 a.m. - 8:10 a.m.
AFCEA WELCOME

Mr. Ray Cross
Vice President for Intelligence
AFCEA International
LTG Bob Noonan, USA (Ret.)
Chair, AFCEA Intelligence Committee

8:10 a.m. - 8:45 a.m.
KEYNOTE: MODERATED CONVERSATION WITH PDDNI

The Hon. Sue Gordon
Principal Deputy Director of National Intelligence
Office of the Director of National Intelligence 

8:45 a.m. - 9:30 a.m.
SESSION ONE: GLOBAL AI THREAT BRIEFING

The US is not alone in recognizing the potential value of AI. To achieve a competitive advantage, the US must understand global trends in adversarial tactics, techniques and procedures in using AI and, where necessary, defend against them. This session will provide an understanding of adversarial capabilities. 

Moderator
Mr. Kurt Miller

Acting NIO for S&T
U.S. Government
Panelists
Mr. Bob Lawton

Analyst
FBI
Ms. Carmen Holliday
Analyst
U.S. Government

Mr. Brady Genz
Analyst
U.S. Army
Chair, STIC AI Working Group

Focus Questions
How are adversaries using AI to threaten US interests at home and abroad? 
What can the US government do to improve the detection and response to these threats that leverages industry and academic partners?

9:30 a.m. - 10:15 a.m.
SESSION TWO: IDENTITY INTELLIGENCE

Identity intelligence (I2) uses multiple data sources to include biometric, biographic and behavioral data to establish identity for targets of interest. I2 approaches can be used to discover the existence of previously unknown threat actors, connecting individuals to other persons, and to characterize potential threats. This panel will discuss current and future efforts to improve the accuracy and scope of I2 systems. 

Moderator
Mr. Greg Soter
Office of the Director of National Intelligence
Panelists
Mr. Brian Baughman

U.S. Government
Ms. Kathy Lane 
Identity Intelligence Execuitve
U.S. Government
Ms. Scott McCallum
Researcher
U.S. Governement
Focus Questions
Where are AI methods currently being used in I2 approaches and what are the gap areas? 
How can industry best engage with AIM to improve I2 performance? 

10:15 a.m. - 10:45 a.m.
NETWORKING BREAK


10:45 a.m. - 11:30 a.m.
SESSION THREE: AUTHENTICITY OF DATA

Recent developments in AI have made it possible to generate synthetic video, audio and images where people, events and activities appear to occur but that are not real. Quickly and accurately separating these fabricated from real data is the mission of the IC. Understanding the current state and potential future state of AI methods for determining the authenticity of data is the focus of this session.  

Moderator
Ms. Michelle Brennan

Principal Systems Engineer
MITRE
Panelists
Mr. Ray Bogusz
Researcher
U.S. Government
Ms. Wendy Dinova-Wimmer
Tehcnical Director for Customer Media Operations
U.S. Government 
Ms. Candice Gerstner
U.S. Government
Mr. Ed Grogan
Technical Director for Custom Media Operations
U.S. Government
Mr. Chris Vogeler
Computer Forensics Analyst
DHS
Focus Questions
How can AI methods be used to improve the detection of fake data? 
How can these methods be scaled to handle the large volumes required to be analyzed? 
How can AI methods be updated as new approaches for generating fake data are developed?

11:30 a.m. - 12:30 p.m.
NETWORKING LUNCH


12:30 p.m. - 1:15 p.m.
SESSION FOUR: DARPA AND IARPA RESEARCH

Two of the AIM strategy objectives are to identify and invest in addressing gaps and to support basic research. DARPA and IARPA share these goals and are actively participating in efforts that support AIM initiatives, including identity intelligence, authenticity of data, assurance, digital foundation and human-machine teaming. This session will focus on the research programs in DARPA and IARPA, how they support the AIM goals and how industry, academia and other partners can engage.

Moderator
Dr. David Honey

Panelists
Mr. John Beieler
Intelligence Advanced Research Projects Activity (IARPA) 
Dr. Valerie Browning
Deputy Director
Defense Advanced Research Projects Agency (DARPA) Defense Science Office
 
Focus Questions
What are the goals, capabilities and limits for efforts in:
        o identity intelligence;
        o determining the authenticity of data; 
        o building more robust AI systems; and,
        o human-machine teaming (especially explainability)?
How would you characterize the national and global research context surrounding your agency’s work in these areas as a whole, and your ability to compete scientifically in today’s environment?
Outside of traditional BAAs, how can industry, academia and other groups work with DARPA and IARPA? 
Are there efforts going on that indirectly support AIM such as accreditation process acceleration, low power chips or improved storage?

1:15 p.m. - 2:00 p.m.
SESSION FIVE: DIGITAL FOUNDATION

A robust foundation is fundamental to success in the application of AI to mission needs. For AIM, the foundation consists of more than just access to elastic compute and storage. A digital foundation also includes methods for data discovery, data fusion, cleaning, and workflow management. It must support both the rapid exploration of new ideas on relatively small data samples, as well as the scaling of test processes to the enterprise-level data feeds and users. This panel will discuss current capabilities as well as plans for achieving a digital foundation that meets these challenges.

Moderator
Mr. Tom Drayer

U.S. Government
Panelists
Ms. Rebecca King

Deputy IC Chief Data Officer
Mr. Joshua Marcuse
Executive Director 
Defemse Innovation Board

 
Mr. Jason Wang
Researcher
U.S. Government
Focus Questions
How can new software environments be securely but rapidly provisioned? 
What are the current impediments to data discovery and how can they be addressed while maintaining appropriate controls? 
What architectures and processes are being used to bring efforts all the way from lab to factory in a timely way?

2:00 p.m. - 2:15 p.m.
STRETCH BREAK


2:15 p.m. - 3:00 p.m.
SESSION SIX: AI ASSURANCE & SECURITY
Many AI approaches have achieved remarkable, even super-human capabilities. But often these successes are very narrow and susceptible to surprising failures. Fielding such brittle and poorly understood capabilities won’t gain the trust of users and will fail to meet the needs of the IC. This session will focus on developing approaches for measuring how well AI systems meet mission needs, understanding how an AI approach fits into a larger environment and discussing how to address both natural and adversarial changes to the environment that will impact this performance.

Moderator
Mr. John Beieler

Intelligence Advanced Research Projects Activity (IARPA) 
Panelists
Mr. Daniel Felps

National Geospatial-Intelligence Agency
Mr. Mikel Rodriguez
Computer Vision Researcher
MITRE
 
Focus Questions
How can a system that continuously changes based on new data be maintained? 
What are the ways an adversary might attack an AI system and how can they be stopped? 

3:00 p.m. - 3:15 p.m.
SESSION SEVEN: NEXT STEPS & ENGAGEMENT STRATEGY

Mr. Dean Souleles
Chief Technology Advisor
Office of the Director of National Intelligence 

Ms. Corin Stone
Deputy Director, Strategy and Engagement
Office of the Director of National Intelligence

3:15 p.m. 
CLOSING REMARKS