Dr' Dana L' Ulery

1
U.S. Army Research Laboratory
Defense Situational Understanding in the Struggle
Against Violent Extremism
Dr. Dana L. Ulery ARL Knowledge Fusion
Center Computational and Information Sciences
Directorate
2
Outline

• The Face of the Enemy
• KFRW06 Focus Defense Situational Understanding
  (DSU)
• B3AN A Promising DSU Approach
• A New DSU Policy Initiative
• Final Reminder

3
Far and away the best prize that life has to
offer is the chance to work hard at work worth
doing. Theodore Roosevelt
4
The Face of the Enemy
5
Al Qaedas Master Plan
Al-Zarqawi The Second Generation of Al Qaeda,
2005 a 20 year plan From an article by Lawrence
Wright, The New Yorker, Sep 11, 2006

• Stage 1 The Awakening Sept 11, 2001 to 2003
• Start Sept. 11, 2001
• To wake the Islamic nation from its state of
  hibernation
• A strike at America, the head of the serpent
• End 2003, when US entered Baghdad
• Stage 2 Eye-Opening ends 2006
• Iraq becomes the recruiting ground
• Jihad on the Web will propagate Al Qaedas ideas
• Muslims will be pressed to donate funds
• Stage 3 Arising and Standing Up from 2007 to
  2010
• Expand area of conflict in Middle East, causing
  US to overextend its forces
• Will begin to directly confront Israel to gain
  more Muslim support
• Stage 4 Bring about the demise of Arab
  governments ends 2013
• Will launch electronic attacks to undermine US
  economy
• Will promote using gold as international
  standard, to collapse US dollar
• Stage 5 Declare Islamic caliphate ends 2016
• Stage 6 Total Confrontation
• Islamic Army formed to instigate a worldwide
  fight between believers and non-believers
• Definitive Victory achieved by 2020

For the new theorists of jihad, Al Qaeda is just
the beginning
6
Army G2 Report to Senate Committee on Armed
Services
All of our initiatives seek to leverage all
collected data to conduct fusion analysis to
generate relevant intelligence for decision
makers, both Commanders and Soldiers. Today, our
Soldiers in the fight have the greatest local
knowledge but the poorest situational knowledge
outside their immediate area of operations (only
hundreds of meters). The opposite situation
exists at the Joint, Theater and National levels,
where the greatest operational and strategic
knowledge resides, with little to no local
knowledge available.
Lt. General Keith B. Alexander April 2004
7
Army Defense Situational Understanding Problem
HUMINT - considered key to defeating the
insurgency is now largely disconnected with
SIGINT, with comm networks, or with social
military networks. More sensors generate more
data more options, but overwhelming
information overload problem.

• WARRIOR ON PATROL
• Needs information that is
• sharply focused
• relevant to immediate situation
• context specific
• real-time.

• 173rd Airborne Brigade
• Carrying 3 days of supply
• Situational awareness limited to last brief
  and current contact
• Fusion done manually today by analysts

8
Defense Situational Understanding
4th Annual KFRW Context-Aware Fusion Applications
for Tactical Overwatch

• YOU are whats most important to this workshop!
• We want your thoughts on identifying
  articulating
• DSU critical technical issues
• DSU high-potential technical opportunities
• Important new fusion research directions
• Are we primarily in agreement on the definition
  of the DSU problem and how to address it? If
  not, what do we need to resolve ?

9
Focus DSU Technical Issues

• Critical Requirements Characteristics
• Solution Approaches
• Major Barriers

10
DSU Critical Requirements Characteristics

• Relationships Matter cultural, social, economic,
  organizational, hunter-prey, objects, data
• Domain Context Are Critical
• Information Must Be Current Dynamic
• Information from Anywhere Everywhere
• Information Must Be Secure Trusted
• Distributed Information Collection Delivery
• Lightweight Presentations

Others ?
11
Domain Context Highlights

• KFRW04 Panel Perspectives on Fusion Systems,
  Complexity, and Context
• Moderators Dana Ulery Gene DeLoatch
• Panelists Jim Hendler, Jim Llinas, Ed Waltz,
  Richard Antony, V.S. Subrahmanian, Craig Scott,
  Roy George
• BORG Context Powered Fusion Research, by D.
  Ulery, KFRW04
• Network Science the Embedded Threat, by Ed
  Waltz, ARL March 2006
• 5th Annual CMIF Workshop on Critical Issues in
  Information Fusion, Sept. 2006
• Jim Llinas, Chair
• Thrust 1 Utility Representation of Context in
  Fusion Architectures Leads Otto Kessler
  Mitch Kokar

12
Solution Approaches Business EnterprisesA
useful analogy for tactical DSU

• Increasing effort to understand multiple factors
  that impact the situation
• Social, behavioral, environmental, financial,
  political
• Rely on deep automated application solutions to
  understand routine situations
• Enterprise application frameworks are often the
  cornerstone of enterprise business activities
• Unambiguous domain context terminology
• Fast but usually not flexible
• Examples HR Management, Financial Management,
  Customer Relationship Management, Time Labor
• Use human systems enabled by machine tools to
  understand non-routine situations
• Typically slow with considerable ambiguity, but
  some flexibility
• Examples risk management, emergency management
• Real-time application systems support critical SU
  for operations
• Sales customer transactions (Amazon.com,
  Fidelity)
• Manufacturing Process Control
• Supply Chain Control

13
Brigade Below Battlefield Awareness Network
(B3AN) A DSU Solution Approach

• B3AN is a DSU Application Services Framework
• Prototype for fusion research in
• Fusion Services
• Information fusion
• Software fusion
• Data fusion
• Component integration platform for fusion
  components, services, algorithms
• B3AN labs at ARL KIMCOE
• B3AN Application Services for Military Domains
• Threat Prediction Tracking
• Tactical Overwatch in Urban Operations/ Video

14
Solving DSU-related Problems
B3AN is a promising approach for automating a
Family of Application Services to cover the DSU
Solution Space

• a sphere of activity, interest, or function
• concepts, activities rules are central

• Attributes strongly associated with domain
  include
• Concepts
• Activities
• Rules
• Context
• Terminology
• Communities of Interest
• Urgency
• Granularity of detail
• Authority
• Policy
• Practices
• Complexity

Domain
Network of Applications
Solution Space
15
Benefits

• Use-driven models design
• Speed of use
• Sharp reduction of information overload
• Focus on specific problem domain
• Specialized databases tools
• Selectively utilizes public/private resources
• Takes advantage of users domain knowledge to
  achieve situational understanding
• Encourages common/ best practices
• B3AN Framework enhances software modularity,
  reusability, extensibility 1,2

• M. Fayad and D. Schmidt, Object-Oriented
  Application Frameworks, Comm. ACM, 1997
• M. Papazoglou and D. Georgakopoulaos,
  Service-Oriented Computing Introduction, Comm.
  ACM, 2003
• D. Ferguson, Model Driven Architecture Meets
  Service-Oriented Architecture, IBM, 2003.

16
B3AN Research Goals

• Research Application Services Approach to DSU
• Identify re-useable fusion capabilities within
  specific Army problem domains
• Explore fusion interoperability across a solution
  space (multiple levels, horizontal vertical)
• Examine relationship of various fusion
  algorithms, agents, models working to solve a
  targeted complex problem.

17
B3AN Architecture
Service Providers
B3AN Clients
Data
Soldier
Threat Prediction
Analyst
Services
Inter-Agent Communications Ontology Mapping
Agent Registry
Transforms
Video Feeder
Tracking
Component App Processes
Integration Re-useable Components, Fusion
Algorithms
B3AN Framework
Source Integration Point
Policy
Fusion Node
Object/Class Models
SOA Infrastructure
18
B3AN Contributors
Policy Driven B3AN (Kaste)
CSUSB ITTN POLICY
Army Fusion Research Center
Real-Time News Analysis (Forester)
Knowledge Captain (Hanratty)
Brigade Below Battlefield Awareness Network
(B3AN)
WFL
MIT
UMD
Soldiers As Effective Sensors (Smith, Ohio State)
Policy Aware Web (Hendler, Berners-Lee, Weitzman)
KAoS (Bradshaw)
KIMCOE
Human Agent Automation (Penn State, SA Tech)
NJ Inst Technology
Morgan State U
U of Virginia
General Dynamics
Event Based Fusion (Oria)
IED Prediction (Brown)
Interoperability Architecture (Lewthwaite)
Intelligent Agents Peer-to-Peer Failover
Recovery (Watkins)
Insurgency Information Framework (Bronner)
Visualization Motion Tracking (Scott)
19
The Cross-Domain Problem A Barrier to DSU

• Critical to DSU is the ability to fuse and share
  information among the Top Secret analyst, the
  Secret commander, and the uncleared warfighter.
• Cross-Domain policy implementation is a major
  barrier to many B3AN application domains.
• Policy makers are calling for a new approach to
  information security classification
• 2004 House Hearing Too Many Secrets
  Overclassification As A Barrier to Critical
  Information Sharing
• 2004 DOD Directive 8320.2, Data Sharing in a
  Net-Centric Department of Defense, and
  subsequent Chairman of the Joint Chiefs of Staff
  Policy CJCSI 3340.02, Horizontal Integration of
  Warfighter Intelligence.

20
Innovation Policy Analysis Program

• A new initiative at Cal State U San Bernardino
  ARL to address the DSU cross-domain problem
• Program goals
• Develop a new cross-domain policy model driven by
  policy aware web concepts capabilities.
• Explore the potential to enable social and
  cultural policy considerations to be wholly or
  partially automated within the context of B3AN
  application domains.
• Manage the introduction of the new technology in
  a planned way that integrates defense policy
  management, organizational behavior and culture,
  and computing research.
• Form a Working Policy Group from the policy
  science and computer science research communities
  to spearhead the plan.
• Form a Policy Council from the defense security
  community and the research communities to refine
  requirements and get solutions fielded
  successfully.

21
A Final Reminder
How does your research help the warfighter to
better understand his situation?
22
Details
23
BRIGADE BELOW BATTLEFIELD AWARENESS NETWORK
(B3AN) - Architecture
Granularity Levels Level 1/Level 3
fusion Interfaces Agent handoffs
Test Eval Orders efficacy using Soldier-as-senso
r
UGS
UAV
S2 Client
Raw Video
Ohio State Dr. Ertin
Ohio State Dr. Smith
Level 1 Fusion
Video
B3A Agent
Source Integration Point
Offline Analysis Visualization
Agent Registry
Auto Annotation Processes
Video
Starlight Analysis
Associated Data ID, Position, Security
Associated Data ID, Position, Security
Filter Processes
Map Overlay
IED Prediction Client
Transform Processes
Auto Library
Fused edited Video Metadata
Streaming Video, Audio
Threat Model
Initial Security Transform Levels
Fusion Node
Internet
Siprnet
Security Rules Engine
Security Client
Intelligence
Intel Officer
Real-time Cache
Real-time semantics

• News Sites
• Defense Sites

Archives
Analysis Agent
Goal-directed Real-time Video
Test Eval Cognitive overload Trust Human/agent
Interaction
B3A Network FY06 Project CTA Collaboration
Reduction Processes
Soldier Client
Unclassified In Context
SA Technologies