Post on 06-Apr-2018
8/3/2019 Almaden Institute James Albus
1/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
A National Program
forUnderstanding the
Mechanisms of MindJames S. Albus
Senior NIST Fellow
Intelligent Systems Division
National Institute of Standards and TechnologyBldg 220, Rm B-124
Gaithersburg, MD 20899
james.albus@nist.gov
8/3/2019 Almaden Institute James Albus
2/25
8/3/2019 Almaden Institute James Albus
3/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Integrate elements from existing programs in:
Neurosciences neurophysiology, brain modeling
Cognitive sciences psychology, reasoning
Computer sciences AI, simulation & modeling
Control theory mechanisms and controlGame theory decision making, cost/benefit analysis
Robotics perception, world modeling, behavior
Visualization computer graphics, video games
Bring together researchers from top laboratoriesaround the country with a common focus for a
"Decade of the Mind"
Approach
8/3/2019 Almaden Institute James Albus
4/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Why Now?
Neurosciences computation and representation in the brain
-biochemistry, synaptic transmission, functional modules, brain imaging
Cognitive Modeling representation and use of knowledge- mathematics, logic, language, learning, problem solving
Intelligent Control making machines behave appropriately- manufacturing, autonomous vehicles, manipulation, locomotion
Depth Imaging geometrical modeling of 3-D world
- image & map segmentation, classification, symbol grounding
Computational Power speed and memory that rival the brain
- >1010 ops today, heading for >1015 ops.
The science & technology is ready
Integration across disciplines reference model architecture
8/3/2019 Almaden Institute James Albus
5/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
We are at a tipping point
Analogous to where nuclear physics was in 1905
Fundamental processes are understood in principle
Technology is emerging to conduct definitiveexperiments
Perception
World modelingReasoning
Planning
Control
Brain structure and function
Cognitive & control architecturesComputational equivalence
Language
Learning & memory
Significant military and economic applications
will develop early in the century
8/3/2019 Almaden Institute James Albus
6/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
We live at a unique point in the history
of science. The technology to discover
and characterize how the subjective mindemerges out of the objective brain is
within reach. The next years will prove
decisive.
-- Christof Koc
hfrom The Quest for Consciousness2004
8/3/2019 Almaden Institute James Albus
7/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Computational power
will soon be available
Todays supercomputer >1014 ops
Computing power of human brain ~ 1013 - 1016 ops
1010
10111012
1013
1014
1015
1016
2005 2010 2015 2020 2025 2030 2035
Computational powerx10 every 5 years
Single board
Cluster of 10
Computing
Power
(ops)
Date
Supercomputer
8/3/2019 Almaden Institute James Albus
8/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Military Future Combat System,
UGV, UAV, UUV, UGS
Commercial autos, trucks
Entertainment video games
Academic AI, robotics
Intelligent MachinesWill Be Critical for
Military Security and Economic Prosperity
Money Is Flowing
Progress is rapid
Billions of $ will be invested over the next decade
8/3/2019 Almaden Institute James Albus
9/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
NIST/ARL Roadmap to 2025
2005 Robust autonomous road-followingand off-road driving
2010 LADAR cameras provide the range,
resolution, and speed to cope with dense traffic
2015 Cognitive reasoning capabilities enable
useful tactical behaviors on the battlefield
2020 Cognitive reasoning and tactical behaviors
approach human levels of performance
2025 Autonomous combat vehicles surpass human
levels of performance in most, if not all, areas
8/3/2019 Almaden Institute James Albus
10/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
A Plan forA National Program for
Understanding theMechanisms of Mind
1. Theory and Fundamental Science
2. Experimental Test Environment
3. Practical Applications
4. Performance Metrics
5. Social, Ethical, and Legal Issues
8/3/2019 Almaden Institute James Albus
11/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
First Draft of a Plan
1. Theory and Fundamental Science
Goals:
a) To develop theoretical models of the brain that
describe the inputs and outputs of all of the major
neural modules and systems of the brain, and specifythe functional transformations that take place therein.
b) To develop theoretical models of the mind that
generate the functional equivalent of the phenomenaof perception, cognition, intention, imagination,
memory, learning, feeling, emotion, and behaviors
of manipulation, locomotion, and language.
8/3/2019 Almaden Institute James Albus
12/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
First Draft of a Plan
2. Experimental Test Environment
Goals:
a) To develop experimental models of the brain
that mimic the inputs and outputs of functional
modules in the brain, and mimic the functionaltransformations that take place therein.
b) To demonstrate the performance of brain models
in controlling systems applied to real-world tasksof locomotion, manipulation, imagination, reasoning,
and natural language conversation.
8/3/2019 Almaden Institute James Albus
13/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
First Draft of a Plan
3. Practical Applications
Goal:
To apply intelligent systems technology to social
and economic problems such as:
Manufac
turing
autos, applianc
es, planes, drugs, textilesConstruction roads, bridges, homes, businesses, factories
Transportation trucks, cars, buses, planes, trains
Agriculture planting, harvesting, tending, aquaculture
Mining and drilling digging, hauling, undersea ops
Recycling and environmental restorationRenewable sources of energy
Education and entertainment
Aids to handicapped and elderly
Medical and nursing care
8/3/2019 Almaden Institute James Albus
14/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
First Draft of a Plan
4. Metrics
a) To develop methods and measures for verifying,
validating, and evaluating models of mind and brain.
b) To develop methods and measures for measuring
the performance of intelligent machines and systems.
Goals:
8/3/2019 Almaden Institute James Albus
15/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
First Draft of a Plan
5. Social & Ethical Issues
Goal:
a) To confront the social, ethical, legal, and philosophical
issues related to investigating the human mind,
inc
luding the implic
ations for mental health.
b) To provide a forum for public debate of the potential
costs, risks, and benefits of understanding the mind,
inc
luding possible religious andc
ivil liberties objec
tions.
c) To address issues of unemployment, economic growth,
and environmental implications of intelligent machines.
8/3/2019 Almaden Institute James Albus
16/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
How Much ShouldWe Invest?
NIST Study of Economic Impact
on Manufacturing and Construction
DOD Studies on Military Impact
of Robotic Vehicles on the Battlefield
DOT Studies on Safety Impact
of DriverWarning and Collision Avoidance Systems
Investment should be commensurate
with expected benefits
8/3/2019 Almaden Institute James Albus
17/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Similar National Investments
Human Genome Program ~$3 Billion
Hubble Space Telescope ~$3 Billion
Apollo Moon Expedition ~ $20 Billion
A Rational National Investment:
$4 Billion over a Decade in
Understanding the Human Mind
International Space Station ~$100 Billion
Iraq war ~$2 Billion/week
8/3/2019 Almaden Institute James Albus
18/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
DARPA interest is high and will continue-- Dr. Tony Tether
DARPA Director
Biologically Inspired Cognitive Architectures (BICA)
Biomemetic Computing
Personalized Assistant that Learns (PAL)
Transfer Learning
Integrated Learning
Architectures for Cognitive Information Processing (ACIP)
Global Autonomous Language Exploitation (GALE)
Advanced Soldier Sensor Information System and Technology(ASSIST)
Real-World Reasoning (REAL)
Coordination Decision Support Assistants (Coordinators)
8/3/2019 Almaden Institute James Albus
19/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
DARPA interest is high and will continue-- Dr. Tony Tether
DARPA Director
Improving Warfighter Information Intake Under Stress
Human-Assisted Neural Devi
c
es
Revolutionizing Prosthetics
Neurotechnology for Intelligence Analysts
About to begin a program to understand how the brain and
vision system work together to process and recognize images
8/3/2019 Almaden Institute James Albus
20/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Krasnow Institute at George Mason Universitywill host a workshop in Spring of 2007
to Ascertain the Advisability of a
Decade of the Mind
Plans for Implementation
Plans to enlist other agencies:
National Institutes ofHealth
National Science Foundation
NASAArmy Research Laboratory
Office of Naval Research
National Academy of Science
8/3/2019 Almaden Institute James Albus
21/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Or snail-mail to:James Albus
National Institute of Standards and TechnologyBldg 220, Rm B-124
Gaithersburg, MD 20899
James.Albus@nist.gov
Feedback Questions
If you wish to register your opinion on any of
these issues, please e-mail me at:
8/3/2019 Almaden Institute James Albus
22/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Feedback Questions
2) Do you believe that a scientific theory of mind is achievable:
within a decade?
within two decades?before 2050?
before 3000?
never?
1) Do you believe that a scientific theory of mind is
a desirable goal?
3) In your field of expertise,
if you had a $ 50 million budget,and a 10 year planning horizon:
what program of research would you propose?
4) How much could you usefully spend?
8/3/2019 Almaden Institute James Albus
23/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
5) What areas of the brain would you choose to model?
6) What phenomena of the mind would you choose to model?
7) What parameters would you include in your model?
8) How would propose to test your model?
Feedback Questions
9) What kinds of experimental apparatus would be required to validate your
model?
10) How would you demonstrate and evaluate results?
11) What are the fundamental metrics and measures?
8/3/2019 Almaden Institute James Albus
24/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
12) What applications could be expected to result from success in what you propose?
in medicine?in clinical practice?in health care?in manufacturing?in transportation?in construction?in services?
in other areas?13) In each area, estimate the economic and social benefits, costs, and risks.
Feedback Questions
14) What do you think is the best approach to raising money and garnering politicalsupport?
15) What are the pitfalls one should anticipate?
16) What are the downside risks?
17) What agencies are likely to provide funding?
18) What other questions need to be asked?
8/3/2019 Almaden Institute James Albus
25/25
NIST Manufacturing Engineering Laboratory Intelligent Systems Division
Handouts are Available