Next-Generation Machine Intelligence and Future Warfare

1
Next-Generation Machine Intelligence and Future
Warfare

• Selmer Bringsjord
• Professor of Logic, Cognitive Science, Computer
  Science
• Rensselaer AI Reasoning (RAIR) Laboratory
  (Director)
• Department of Cognitive Science (Chair)
• Department of Computer Science
• Department of Decision Sciences Engineering
  Systems
• Rensselaer Polytechnic Institute (RPI)
• Troy NY 12180 USA
• TRADOC 11.18.03

2
The Rensselaer AI Reasoning Lab(The RAIR Lab)
Intelligent Tutoring Systems (mathematical logic)
Over 1million internal seeding
Intelligence Analysis
Item generation (theorem proving-based generation)
synthetic characters/psychological time
3
RAIR Web and RD
Advanced Synthetic Characters
MARMML
PERI
Savant
PAI
Slate
CDs
Super Teaching
4
An Unpleasant Future
EFFECTIVENESS
nuclear biological ... (see Livermore List)
Terrorism/Asymmetrical Warfare (enemy)
Machine Intelligence
9/11
today
TIME
5
A Rather Better Future
EFFECTIVENESS
nuclear biological ... (see Livermore List)
Terrorism/Asymmetrical Warfare (enemy)
Machine Intelligence
9/11
today
TIME
6
A Rather Better Future
EFFECTIVENESS
Terrorism/Asymmetrical Warfare (enemy)
Machine Intelligence
9/11
today
TIME
7
AI is back in a big way(e.g., ARDA,
IPTO).Thats a damn good thing.Because its
one future or the other.(And theres no turning
back from an IT-based defense too complicated to
manage without smart machines.)
(but nonetheless incomplete)
8
AI Foundations vs AI Engineering
super-rational adults
EVOLUTION?
EVOLUTION
rational adults
rats
dogs
dolphins
monkeys
children
adults

computers/robots simulate all behaviors
computers/robots simulate some reasoning
Infinitary Reasoning!
King-Ace and the like
Literary Creativity
9
The Paradox of Human vs. Machine Reasoning

• On the one hand, machine reasoners are getting
  faster, and can do some impressive things.
• On the other hand, as Herb Simon, one of the
  grandfathers of AI, admitted before he died,
  machine reasoning is absolutely nowhere when
  stacked against
• first-rate human reasoning
• and nowhere (Selmer explains) when stacked
  against
• the mathematical space of logical systems
  (Linströms Theorems).

10
ResponseNext-Generation Logicist AI
capture the genius of humans (psychology of
reasoning)
real, robust symbolic logic
11
Systems ...
12
MARMML RASCALS(new level of wargaming)
13
RASCALS
14
J-L 1
Suppose that the following premise is true If
there is a king in the hand, then there is an
ace in the hand, or else if there isnt a king
in the hand, then there is an ace. What can you
infer from this premise?
There is an ace in the hand.
NO!
NO!
In fact, what you can infer is that there isnt
an ace in the hand!
15
Cracked Easily in Natural Deduction
16
J-L 2
If one of the following assertions is true then
so is the other (1) There is a king in the hand
if and only if there is an ace in the hand. (2)
There is a king in the hand. Which is more
likely to be in the hand, if either the king or
the ace?
17
Solution
You can obviously prove K A in the cases
where (1) K ? A (2) K T
T K ? A, K - A T F
K ? A, K - A F T K ? A,
K - A so everything must boil down to
the F F case. It's easy here to do a proof that
can get A (constructive dilemma and reductio),
but K instead of K. Ergo, the answer is A.
18
King/Ace Proofs
19
MARMML Proofs in Chess Microworld CHOGIC
No other reasoning system in the world can do
this.
20
Slate ...
21
SLATE
22
Slate Hypothesis Generation in our Narrative
ScenarioWhat is the destination of the convoy?
Mary hits a dead end
How would this work??
SAGE offers
23
Slate Architecture-Sketch
Deduction
Abduction
Induction
24
Slate Architecture-Sketch
Use SNARK MARMML Athena Vampire Oscar ...
Deduction
Abduction
Induction
25
(No Transcript)
26
Want Analysts To Be Able to Work With Arguments
Diagrammed in Intuitive Fashion
Formal Theory (S) of Visual Logic
Slate Argument Diagramming Tool
27
Slate v1.2 Complete(Bringsjord, Shilliday,
Taylor, Wodicka)
28
But ...why is the rebirth of AI in service of
national defense incomplete?
29
A Rather Better Future
EFFECTIVENESS
nuclear biological ... (see Livermore List)
But we/I have to get on the ground!!
Terrorism/Asymmetrical Warfare
Machine Intelligence
9/11
today
TIME
30
Wargaming and Intelligence Analysis are
fine...but AI is needed on the ground!
31
THE END
32
Slate Hypothesis Generation in our Narrative
ScenarioWhat is the destination of the convoy?
Mary is searching for a proof like
---------------- PROOF ---------------- 1
-Yar(x)Terrorists(x). 2 -Camp(x,aconvoy)
-Accessible(x,aconvoylocation). 3
-Village(x,aconvoy) -Accessible(x,aconvoylocation
). 6 -CaveSystem(x,y) -Convoy(y)
-Terrorists(y) -PresentLocation(y,z)
-Accessible(x,z)Village(f1(x,y,z),y)Camp(f2(x,
y,z),y)Destination(x,y). 7 -CaveSystem(x,y)
-Convoy(y) -Terrorists(y) -PresentLocation(y,z)
-Accessible(x,z)Village(f1(x,y,z),y)Accessible
(f2(x,y,z),z)Destination(x,y). 8
-CaveSystem(x,y) -Convoy(y) -Terrorists(y)
-PresentLocation(y,z) -Accessible(x,z)Accessible
(f1(x,y,z),z)Camp(f2(x,y,z),y)Destination(x,y)
. 9 -CaveSystem(x,y) -Convoy(y)
-Terrorists(y) -PresentLocation(y,z)
-Accessible(x,z)Accessible(f1(x,y,z),z)Accessib
le(f2(x,y,z),z)Destination(x,y). 10
-Destination(x,aconvoy). 11 Convoy(aconvoy). 12
Yar(aconvoy). 13 PresentLocation(aconvoy,ac
onvoylocation). 14 CaveSystem(acavesystem19,aco
nvoy). 15 Accessible(acavesystem19,aconvoylocat
ion). 17 hyper,12,1 Terrorists(aconvoy). 18
hyper,15,9,14,11,17,13,unit_del,10
Accessible(f1(acavesystem19,aconvoy,aconvoylocati
on),aconvoylocation)Accessible(f2(acavesystem19,
aconvoy,aconvoylocation),aconvoylocation). 19
hyper,15,8,14,11,17,13,unit_del,10
Accessible(f1(acavesystem19,aconvoy,aconvoylocati
on),aconvoylocation)Camp(f2(acavesystem19,aconvo
y,aconvoylocation),aconvoy). 20
hyper,15,7,14,11,17,13,unit_del,10
Village(f1(acavesystem19,aconvoy,aconvoylocation)
,aconvoy)Accessible(f2(acavesystem19,aconvoy,aco
nvoylocation),aconvoylocation). 21
hyper,15,6,14,11,17,13,unit_del,10
Village(f1(acavesystem19,aconvoy,aconvoylocation)
,aconvoy)Camp(f2(acavesystem19,aconvoy,aconvoylo
cation),aconvoy). 23 hyper,19,2,18,factor_simp
Accessible(f1(acavesystem19,aconvoy,aconvoylocati
on),aconvoylocation). 24 hyper,20,3,23
Accessible(f2(acavesystem19,aconvoy,aconvoylocati
on),aconvoylocation). 25 hyper,21,3,23
Camp(f2(acavesystem19,aconvoy,aconvoylocation),ac
onvoy). 27 hyper,25,2,24 F. ------------ end
of proof -------------
on the strength, say, of a key proposition like
If there is a cave system that's accessible
from the convoy's present location, and it's a
convoy of terrorists, and there's no terrorist
camp accessible from the convoy's present
location, and there's no village accessible
from its present location, then that cave system
is its destination all x all y all
z ((CaveSystem(x,y) Convoy(y) Terrorists(y)
PresentLocation(y,z) Accessible(x,z)
-(exists z1 (Village(z1,y) Accessible(z1,z)))
-(exists z2 (Camp(z2,y) Accessible(z2,z))))
-gt Destination(x,y)).
33
Slate Hypothesis Generation in our Narrative
ScenarioWhat is the destination of the convoy?
But there are no cave systems nearby either!!
Hence Slate models Marys fruitless search
start of search given
clause 1 (wt2) 11 Convoy(aconvoy). . given
clause 2 (wt2) 12 Yar(aconvoy). . given
clause 3 (wt2) 15 Bioagents(aconvoy). . give
n clause 4 (wt2) 16 USBase(amilbase33). . gi
ven clause 5 (wt2) 17 hyper,12,1
Terrorists(aconvoy). . given clause 6 (wt3) 13
PresentLocation(aconvoy,aconvoylocation). . giv
en clause 7 (wt3) 14 CaveSystem(acavesystem1
9,aconvoy). . Search stopped because sos
empty. end of search
34
Slate Hypothesis Generation in our Narrative
ScenarioWhat is the destination of the convoy?
Slate introduces the concept of an attack
position, which must be accessible from the
convoys present location, and an associated
target, which must be wind-accessible from the
attack position to bio-agents, and automatically
produces a proof that supports Mary being told in
simple English that a particular US base is
threatened. Here is the proof in a simple
simulation, obtained in 1.73 seconds of CPU time.
The proof shows that mountain46 is
the convoys destination, and usmilbase33 is
the target
35
Slate Hypothesis Generation in our Narrative
Scenario (? v)What is the destination of the
convoy?
---------------- PROOF ---------------- 1
-Yar(x)Terrorists(x). 2 -WindAccessible(x,y)
-USBase(x) -Bioagents(z) -Terrorists(z)AttackPo
sition(y,z,x). 3 -CaveSystem(x,aconvoy)
-Accessible(x,aconvoylocation). 4
-Camp(x,aconvoy) -Accessible(x,aconvoylocation).
5 -Village(x,aconvoy) -Accessible(x,aconvoyloc
ation). 6 -AttackPosition(x,y,z) -Convoy(y)
-Terrorists(y) -PresentLocation(y,u)
-Accessible(x,u)CaveSystem(f1(x,y,u,z),y)Villag
e(z2,y)Camp(z3,y)Destination(x,y). 7
-AttackPosition(x,y,z) -Convoy(y)
-Terrorists(y) -PresentLocation(y,u)
-Accessible(x,u)CaveSystem(f1(x,y,u,z),y)Villag
e(z2,y)Accessible(z3,u)Destination(x,y). 8
-AttackPosition(x,y,z) -Convoy(y)
-Terrorists(y) -PresentLocation(y,u)
-Accessible(x,u)CaveSystem(f1(x,y,u,z),y)Access
ible(z2,u)Camp(z3,y)Destination(x,y). 9
-AttackPosition(x,y,z) -Convoy(y)
-Terrorists(y) -PresentLocation(y,u)
-Accessible(x,u)CaveSystem(f1(x,y,u,z),y)Access
ible(z2,u)Accessible(z3,u)Destination(x,y). 10
-AttackPosition(x,y,z) -Convoy(y)
-Terrorists(y) -PresentLocation(y,u)
-Accessible(x,u)Accessible(f1(x,y,u,z),u)Villag
e(z2,y)Camp(z3,y)Destination(x,y). 11
-AttackPosition(x,y,z) -Convoy(y)
-Terrorists(y) -PresentLocation(y,u)
-Accessible(x,u)Accessible(f1(x,y,u,z),u)Villag
e(z2,y)Accessible(z3,u)Destination(x,y). 12
-AttackPosition(x,y,z) -Convoy(y)
-Terrorists(y) -PresentLocation(y,u)
-Accessible(x,u)Accessible(f1(x,y,u,z),u)Access
ible(z2,u)Camp(z3,y)Destination(x,y). 13
-AttackPosition(x,y,z) -Convoy(y)
-Terrorists(y) -PresentLocation(y,u)
-Accessible(x,u)Accessible(f1(x,y,u,z),u)Access
ible(z2,u)Accessible(z3,u)Destination(x,y). 14
-Destination(amountain46,aconvoy). 15
Convoy(aconvoy). 16 Yar(aconvoy). 17
PresentLocation(aconvoy,aconvoylocation). 19
Accessible(amountain46,aconvoylocation). 20
WindAccessible(amilbase33,amountain46). 21
Bioagents(aconvoy). 22 USBase(amilbase33). 23
hyper,16,1 Terrorists(aconvoy). 24
hyper,20,2,22,21,23 AttackPosition(amountain46,a
convoy,amilbase33). 25 hyper,24,13,15,23,17,19,un
it_del,14 Accessible(f1(amountain46,aconvoy,acon
voylocation,amilbase33),aconvoylocation)Accessibl
e(z2,aconvoylocation)Accessible(z3,aconvoylocatio
n). 26 hyper,24,12,15,23,17,19,unit_del,14
Accessible(f1(amountain46,aconvoy,aconvoylocation
,amilbase33),aconvoylocation)Accessible(z2,aconvo
ylocation)Camp(z3,aconvoy). 27
hyper,24,11,15,23,17,19,unit_del,14
Accessible(f1(amountain46,aconvoy,aconvoylocation
,amilbase33),aconvoylocation)Village(z2,aconvoy)
Accessible(z3,aconvoylocation). 28
hyper,24,10,15,23,17,19,unit_del,14
Accessible(f1(amountain46,aconvoy,aconvoylocation
,amilbase33),aconvoylocation)Village(z2,aconvoy)
Camp(z3,aconvoy). 29 hyper,24,9,15,23,17,19,unit_
del,14 CaveSystem(f1(amountain46,aconvoy,aconvoy
location,amilbase33),aconvoy)Accessible(z2,aconvo
ylocation)Accessible(z3,aconvoylocation). 30
hyper,24,8,15,23,17,19,unit_del,14
CaveSystem(f1(amountain46,aconvoy,aconvoylocation
,amilbase33),aconvoy)Accessible(z2,aconvoylocatio
n)Camp(z3,aconvoy). 31 hyper,24,7,15,23,17,19,un
it_del,14 CaveSystem(f1(amountain46,aconvoy,acon
voylocation,amilbase33),aconvoy)Village(z2,aconvo
y)Accessible(z3,aconvoylocation). 32
hyper,24,6,15,23,17,19,unit_del,14
CaveSystem(f1(amountain46,aconvoy,aconvoylocation
,amilbase33),aconvoy)Village(z2,aconvoy)Camp(z3,
aconvoy). 33 hyper,26,4,25,factor_simp,factor_sim
p Accessible(f1(amountain46,aconvoy,aconvoylocat
ion,amilbase33),aconvoylocation)Accessible(z2,aco
nvoylocation). 34 hyper,27,5,33,factor_simp
Accessible(f1(amountain46,aconvoy,aconvoylocation
,amilbase33),aconvoylocation)Accessible(z3,aconvo
ylocation). 35 hyper,28,5,33,factor_simp
Accessible(f1(amountain46,aconvoy,aconvoylocation
,amilbase33),aconvoylocation)Camp(z3,aconvoy). 36
hyper,35,4,34,factor_simp Accessible(f1(amount
ain46,aconvoy,aconvoylocation,amilbase33),aconvoyl
ocation). 37 hyper,29,3,36 Accessible(z2,aconvoy
location)Accessible(z3,aconvoylocation). 38
hyper,30,3,36 Accessible(z2,aconvoylocation)Cam
p(z3,aconvoy). 39 hyper,38,4,37,factor_simp
Accessible(z2,aconvoylocation). 40
hyper,31,5,39 CaveSystem(f1(amountain46,aconvoy
,aconvoylocation,amilbase33),aconvoy)Accessible(z
3,aconvoylocation). 41 hyper,40,3,36
Accessible(z3,aconvoylocation). 42
hyper,32,5,39 CaveSystem(f1(amountain46,aconvoy
,aconvoylocation,amilbase33),aconvoy)Camp(z3,acon
voy). 43 hyper,42,3,36 Camp(z3,aconvoy). 44
hyper,43,4,41 F. ------------ end of proof
-------------
customary destinations ruled out
Shows that mountain46 is convoys destination