1. 3 bit automaton
for artificial humane intelligence
Takahashi Toshiki
2012/12/01
wyc@tmtoc.com
2. 1. 5-value logic
• Y -- Yes
Classical Logic
• N -- No
• W -- Wait
• E -- Executable Process State of OS
• R -- Run
3. 2. 2 types of computing system
Neumann Type Non Neumann Type
・Stored program ・Flat network architecture
・Manager controls processes ・No manager
4. 3. 8 states automaton
Judge Action
B -- Begin
5 values + F -- Finish = 8 states
S -- Stop
5. 4. 3 bits allocation
D S1 S2 C
D S1 S2 0 0 0 B
XXX 0
0
0
1
1
1
Y
W
Data Bit State Bit 0 1 0 N
1 1 0 R
1 1 1 S
1 0 1 E
1 0 0 F
6. 5. logical chart between states
∧ B Y N W F E R S
B B Y N W F E R S Ex:
Y Y N W F E R S B Y∧N = W
N N W F E R S B Y
W W F E R S B Y N 001 Y
F F E R S B Y N W + 010 N
E E R S B Y N W F
_____
R R S B Y N W F E
011 W
S S B Y N W F E R
7. 6. future for OS in SoC
• Implementation
・Classical system (0 & 1)
・Quantum system (4 entangled states)
・DNA system (4 types of base, AGCT)
• “HW + OS → SoC” = Free Device Design
・elastic device
・edible device
・creature device
8. 7. Implementation of Intelligence
• Automaton + Probability = Intelligence
• Criteria: “Select option which has the biggest
information.”
Selection = Max(Π|p(i)|).
• Setting complex probability in this automaton,
it can discover and predict rules,
then learn them.
→this is exactly artificial humane intelligence!
9. Appendix: complex probability
Probability p(i) -- information we can get from event i
• p(i) = 1
we know the event occurs without any observation (universal truth)
• p(i) = 1/2
we observe the event occurs (measurement)
• p(i) = 0
we know the event does not occurs without any observation, so
we cannot know the event occurs with any observation (mystery)
• p(i) = -1/2
we know the event can occur with some observation (discovery)
• p(i) = -1
we can know the event can occur without observation (prediction)