PACC Brief

1
Cognitive Information Processing
Technology Brief to High Performance Embedded
Computing Workshop 2002 Mr. Zachary J.
Lemnios Information Processing Technology
Office Deputy Director 24 September 2002
2
Acknowledgements

• Ron Brachman DARPA/IPTO
• Bill Dally Stanford University
• Bob Graybill DARPA/IPTO
• David Honey DARPA/ATO
• Mark Horowitz Stanford University
• Steve Keckler University of Texas
• Dave Koester Mitre
• Bob Leheny DARPA/MTO
• Christie Marrian DARPA/MTO
• Dan Radack IDA

3
Agenda

• Introduction
• DoD System Challenges
• DARPA Offices and Programs
• Technology Trends
• Device Performance
• High Performance Computing
• Cognitive Processing Technology
• Summary

4
Asymmetric AdvantageEnabled by Information
Superiority
See Further with Greater Clarity
Space Based RADAR
E-3 AWACS Airborne Early Warning
FLTSAT
Secure Comm
GMTI, SAR, STAP, HSI
Tier II UAV
RC-135V Rivet Joint
Space-Time Adaptive Processing
SIGINT
Standard Missile
E-2C Hawkeye
Cooperative engagement
Aegis Cruiser
Towed Sonar Array
Adaptive Matched-Filter processing
TEL
Wideband high linearity target discrimination mult
i-angle target discrimination
Small Unit Operations




Covert sensing Power constrained
operations Secure LPI A/J communications
Chemical biological threat detection
5
Information Networks Will Revolutionize Platforms

• TODAY - Federated Architecture (Baselines 1-5)
• Point-to-Point Mainframes (UYK-7/43)
• Limited Growth Capability
• Vulnerable to Damage

• FUTURE - Distributed Processing (Future B/L)
• Highly Distributed Network
• Redundancy Plus Reconfigurability
• Effectively Invulnerable to Battle Damage

The Upside Agile, Adaptable, Survivable Systems
The Downside Complexity in h/w, s/w, integration
From www.hypocrites.com/pictures
6
A New Class of Autonomous Systems
Autonomous Systems
Mission Complexity
Cooperative Systems

• Human-like sensor depth
• Persistent engagement
• Extreme power and volume constraints
• Effects-based tasking and performance

Legacy Systems
Environmental Complexity
7
DARPA Organization
Director, Tony Tether
Special Projects Amy Alving Joe Guerci Chem/Bio
Def Systems Counter Underground Facilities Space
Sensors/Structures Navigation/Sensors/ Signal
Processing
Information Exploitation Richard Wishner Steven
Welby / Robert Tenney Sensors Exploitation
Systems Command Control
Tactical Technology Allen Adler Art Morrish
Air/Space/Land Platforms Unmanned Systems Space
Operations Laser Systems Future Combat
Systems Planning / Logistics
Advanced Technology Tom Meyer Dave Honey Assured
C3ISR Maritime Early Entry/Special Forces IAS
Programs
Defense Sciences Michael Goldblatt Steven
Wax Bio Warfare Defense Technologies Biology Mate
rials Devices Mathematics
Information Awareness John Poindexter Robert
Popp Asymmetric Threat Prediction Behavior
Modeling
Information Processing Technology Ron
Brachman Zach Lemnios Cognitive
Systems Computational - Perception Representatio
n Reasoning
Learning Natural Communication
MicrosystemsTechnology Robert Leheny Electronics
Optoelectronics MEMS Combined Microsystems
Focus of this brief
8
DARPA/MTOPlatform Scale Information Systems
Process
Sense
Actuate
Memory
Memory
Controller
RF ?w MMW
Human Interface Machine Interface Weapon
Interface Network Interface
Human Interface Machine Interface Weapon
Interface Network Interface
IR
Switching
Switching
Sensor(s)
I/O
I/O
UV
Analog
Digital
Output
Output
Visible
A/D
A/D
A/D
Bio
Processor(s)
Processor(s)

• Highly capable sensors- sensors that are self
  adapting
• Enhanced extraction of signals from background,
  noise, and jamming
• Covert data into actionable knowledge in near
  real time
• Provide technology for assured communication links

9
The Challenge of Complexity
1.E11
Ops/sec/
1.E10
1.E09
doubles
every 1.0
years
1.E06
CMOS
Tubes/ Transistor
1.E03
Mechanical/ Relays
1.E00
doubles every
doubles every
2.3 years
7.5 years
1.E-03
Combination of Hans Moravac Larry Roberts
Gordon Bell
1.E-06
2030
1880
1900
1920
1940
1960
1980
2000
2010
2020

• While computational performance is increasing,
  productivity and effectiveness are not keeping
  up
• Users must adapt to system interfaces, rather
  than vice versa
• Systems have become more rigid and more fragile
• Systems have become increasingly vulnerable to
  attack
• The cost of building and maintaining systems is
  growing out of control

10
DARPA/IPTOCognitive Systems

• DARPA IPTO will create a new generation of
  cognitive computational and information systems
  with capability to
• reason, using substantial amounts of
  appropriately represented knowledge
• learn from their experience so that they perform
  better over time
• explain themselves and be told what to do
• be aware of their own capabilities and reflect on
  their own behavior
• respond robustly to surprise

Systems that know what theyre doing
11
Why Now?

• Human-level scaling of HW technology is on the
  horizon
• Foundations established for human neural systems
• Cognitive technology (from AI) is being applied
  to initial problems

12
The Result Will Enable a Revolution in
Capability for DoD
More Aggressive Threats
Adaptive and Intelligent Data-Fused Sensors
Threats are more dynamic and in deeper hide
(collapsing time lines) System performance is
outpaced by changing threat environments Cooperati
ve battle management requires robust information
backbone
Sensor Data Flow Overwhelming Human Analyst
Cognitive Information Exploitation
Sensor bandwidth is increasing faster than
processor capability Target classification has
become a multi sensor
The next revolution in sensing Autonomous
Adaptation The next revolution in computing
Cognitive Processing
13
Agenda

• Introduction
• DoD System Challenges
• DARPA Offices and Programs
• Technology Trends
• Device Performance
• High Performance Computing
• Cognitive Processing Technology
• Summary

14
Beyond CMOS The Road Beyond Bulk Silicon Field
Effect Transistors
Self-directed assembly
Quantum dots/Cell. Automata
NDR, RTD
SET
Nanotubes
Molecular devices
Nonplanar Switches
1/size
Schottky source/drain FET
Double-Gate CMOS
Bulk CMOS
High k gate oxide
SiGe/Ge FETs and Structures (strained layers)
Metal gate
After Chart by P. Wong, IBM
15
Memory Wall is Growing
100
Memory access times (cycles) are increasing based
on SIA clock frequency roadmap
180nm
130nm
100nm
70nm
50nm
10
1
1
10
100
1000
10000
Cache Capacity (KB)
Source Keckler Univ. of Texas
16
SIA Roadmap Impact on Computer Architectures
Single Clock Area
New architectures are required to accommodate
smaller clock regions
17
Novel Architectures are Required to Extend
Performance Productivity
X86 (52/year)
19/year
Moores Law (74/year)
301
30,0001
Opportunity for Cognitive Architectures
Source ISAT Summer 2001 Study- Last Classical
Computer Prof. Bill Dally (Stanford U) Study
Lead
18
Embedded Computing Performance Regions
10000
Custom VLSI
1000
Polymorphous Architectures
100
Reconfigurable
Computation Density (MOPS/cm3)
10
1
0.1
0
.
0
1
0
.
1
1
1
0
1
0
0
1
0
0
0
Computational Efficiency (GOPS/Watt)
19
Polymorphous ComputingArchitectures Program
Goal Computing systems (chips, networks,
software) that will morph to changing missions,
sensor configurations, and operational
constraints during a mission or over the life of
the platform.
Mission Aware Embedded Computing
20
Agenda

• Introduction
• DoD System Challenges
• DARPA Offices and Programs
• Technology Trends
• Device Performance
• High Performance Computing
• Cognitive Processing Technology
• Summary

21
DARPA/IPTOCognitive Systems

• DARPA IPTO will create a new generation of
  cognitive computational and information systems
  with capability to
• reason, using substantial amounts of
  appropriately represented knowledge
• learn from their experience so that they perform
  better over time
• explain themselves and be told what to do
• be aware of their own capabilities and reflect on
  their own behavior
• respond robustly to surprise

Systems that know what theyre doing
22
Cognitive Systems Thrusts
Systems
Core Cognition
Representation Reasoning
Learning
Communication Interaction
Perception
Robust Software and Hardware
Foundational Science and Mathematics
Bio-inspired Computing, new approaches to Trust
Management
Foundation
23
Anatomy of a Cognitive Agent
Reflective Processes
LTM (knowledge base)
Cognitive Agent
STM
Concepts
Deliberative Processes
Other reasoning
Sentences
Communication (language, gesture, image)
Prediction, planning
Action
Perception
Reactive Processes
Sensors
Effectors
External Environment
24
Initial Challenge Context

• Persistent, personal partner/associate systems
• Learn from experience
• Learn what you like and how you operate
• by observation
• by direct instruction or guidance, in a natural
  way
• Imagine possible futures, anticipate problems and
  needs
• Omnipresent / always available
• Examples
• Commanders (C2) assistant
• (Intelligence) Analysts associate
• Personal executive assistant/secretary
• Disaster response captains RAP
  (robot/agent/person) team

25
Summary

• DoD is facing immense challenges
• New and dynamic threats in much deeper hide
• Collapsing timeliness, rapidly changing threat
  environments
• New classes of autonomous systems require
• Platform Scale Integration (DARPA/MTO)
• Cognitive Capabilities (DARPA/IPTO)
• Systems that Know what they are doing
• can reason
• can learn from their experience
• can explain themselves
• can be aware of their own capabilities
• can respond robustly to surprise

26
Challenge

• Send us your best ideas
• IPTO BAA 02-21, http//www.darpa.mil/ipto/Solicita
  tions/PIP_02-21.html
• MTO BAA Solicitation http//www.darpa.mil/mto/soli
  citations/index.html
• Take a tour as a DARPA Program Manager
• rleheny_at_darpa.mil [phone]
• rbrachman_at_darpa.mil [phone]
• zlemnios_at_darpa.mil [phone]

27
Backup
28
Memory Issues Lead to Inefficient Performance
STREAMS ADD Computes A B for long vectors A
and B (historical data available)
29
Todays Systems - Collection of Rigid Embedded
Subsystems
System-of-Systems
Static Mission Scripts
Micro-Systems
Micro-Systems
Mode Driven Configuration
.
Sub- System
Sub- System
Sub- System
Sub- System
Sub- System
Sub- System
Pre-defined Functional Capability
Sense
Actuate
Process
Sense
Actuate
Process
Single Instance, Point-Design Implementations
DoD Systems must Move from Integration Driven
Architectures to Capability Driven Architectures
30
High Performance Embedded Cognitive Systems
System-of-Systems Cognitive Information Extraction
Goal(s) Driven Missions
Cognitive micro-Systems
Cognitive micro-Systems
Cognitive Guided
.
Sub- System
Sub- System
Sub- System
Sub- System
Sub- System
Sub- System
Cognitive Enabled Agile Subsystem (e.g. PCA)
Sense
Actuate
Process
Sense
Actuate
Process
Local Hard Real-Time Cognitive Agents
Systems with Human Like Capability
31
Intelligent Systemsfor Mission Agility
Cognitive Processing Adaptation
u
Sense
Actuate
Process
Systems with Human Like Capability
32
DARPA/IPTOCognitive Systems

• While computational performance is increasing,
  productivity and effectiveness are not keeping up
• Users must adapt to system interfaces, rather
  than vice versa
• Systems have become more rigid and more fragile
• Systems have become increasingly vulnerable to
  attack
• The cost of building and maintaining systems is
  growing out of control

33
Information Technology

• Information Exploitation Office
• Detection, Precision ID, Tracking, and
  Destruction
• of Elusive Surface Targets
• Information Awareness Office
• Detect Defeat Terrorist Networks
• Information Processing Technology Office
• Cognitive Systems
• Advanced Technology Office
• Robust, Secure Self-Forming Tactical Networks

34
Feature Size Trends
Intel8080
1 million transistors, 40MHz
1000nm
Intel386
100 million transistors, 3GHz
Intel486
Pentium
PentiumPro
10 billion transistors?, 30GHz?
100nm
Pentium IV
Synchronous (clocked)
Feature size (nanometers)
10nm
1nm
1970
1980
1990
2000
2010
2020
2030
2040
2050
CMOS
35
High ProductivityComputing Systems Program

• Goal Provide a new generation of
  economically viable high productivity computing
  systems for the national security and industrial
  user community (2007 2010)
• Approach Implement productive high-end systems
  with high effective bandwidth, low latency,
  balanced system architecture, robustness,
  application responsive tailorability,
  performance measurement and prediction

• Applications
• Intelligence/surveillance, reconnaissance,
  cryptanalysis, weapons analysis, airborne
  contaminant modeling and biotechnology

Fill the Critical Technology and Capability
Gap Today (late 80s HPC technology)..to..Future
(Quantum/Bio Computing)