1.
Theory of
AI4U++
Textbook Algorithm Steps: Code the Volition Mind-Module
Hearing Vision Concepts Volition Emotion Motor Output /iiiiiii\ /!i!i!i!\ /YYYYYYYYYYYY\ | ||||||| || ||||||| | + | |||||||||||| | | ||||||| || | ___ | | + | |||||||||||| | | ||||||| || /old\ | + | |S|||||||||| | | ||||||| || (image)-|---+_ | |H|||||||||| | | ||||||| || \___/ | / \ | |A|||||||||| | | ||||||| || | (idea) __ | |K|||||||||| | | | ||||| || | \__/---------------/ \ | |E|||R|||||| | | |d------||---------|---+ ____ (fear)-|--*|||U|||||| | | ||||o|| || _____ | +-------/ \----\__/ | |||||N|||P|| | | ||g|||| || / re- \-|---+ / de- \---------|------*|||E|| | | || |||| ||/entrant\| + ( ci- ) | |||||||||T|| | | ||||||| ||\ image /| + \ sion /---------|----------*|| | | ||||||| || \_____/ | + \____/ | |||||||||||| |
In the diagram above, the "decision" area of volition receives
various levels of input from the knowledge base (KB) of the mind
and from the emotion of fear, which urges action more immediate
than the strictly rational mind contemplates.
(Preliminary) Purpose: The computation of decisions as the outcome
of the conscious and subconscious interplay of all knowledge in the
artificial mind, all values held by the artificial mind, any
supergoal
dominating the AI mind, and all motor options available to the AI.
Tentative inputs: Numeric levels of urgency or desirability in
association with a proposed motor initiative.
A
bouleumatic accumulator (or "volitional accumulator") is a
neuronal device that accumulates activation towards a trigger
level, under the simultaneous dynamics of storage and drainage.
Contemplating a motor option deliberately over time causes
storage of a build-up of activation in the volitional accumulator,
while drainage occurs automatically over time. If a threshold
level of activation is reached, the motorium goes into action.
Otherwise, drainage continues and the motor option fades away.
It's like: CNS --> VA --> MO
where
CNS = the Central Nervous System contemplating an action; and
VA = Volitional Accumulator; and
MO = Motor Option (not unlike "modus operandi", no? :-)
The brain-mind needs a volitional accumulator (VA) in between
thought and action so that thought will not immediately
cause action, but will at first only *propose* action.
Since infancy, thoughts and actions have remained associatively
linked and have remained volitionally coupled by the
Volitional Accumulator mechanism so that planning may occur.
Returns: Firing signals to initiate motor activity.
2. Stimulus-Response
Before the evolution of volition, a central nervous system (CNS)
deals with its environment on a stimulus-response basis, with no
allowance for variation in a repertoire of genetically hard-wired
behaviors and with no provision for deliberation and free will.
If the CNS in question has no memory of its lifelong experiences,
then it obviously has also no option of choosing among behaviors.
The robot or organism with no record of experience must do only
what it is pre-programmed to do: move towards food or energy;
move away from pain or danger; and never learn from experience.
3. Random Firing of Infantile Motor Neurons
Imagine (or perhaps look around you and observe) your human baby
lying backside-down in a crib and randomly waving little arms and
kicking little feet. You are witnessing the emergence of loops
of motor control in the brain-mind of the infant organism, which
at first randomly activates muscles, then feels and remembers the
results through the input sensorium, and then gradually learns an
ability -- carried forward through time -- to choose freely the
motor options which fired at first randomly and then voluntarily.
4. Voluntary Selection of Motor Options
Your childlike robot needs to remember in software what its
motor options are and what its basic needs and
goals are.
Then you the programmer need to construct a software mechanism
for the interactive match-up of the momentarily most compelling
need or goal with the best motor strategy for its achievement.
By a process of trial-and-error as time goes by, your robot must
refine its accumulated motor strategies while it subconsciously
remembers the initially random but afterwards voluntary linkages
between thoughts about action and motor engrams that cause action.
Supposedly every computer program operates in a deterministic
fashion, where the results of the computation are pre-determined
by the ineluctable and inexorable logic of the computer program.
Well, think again, Mr. Spock. Logic dictates that when random
events enter into a computation, the computation itself becomes
random and non-deterministic. The computations in the Mind of
your robot have random interactions with a very unpredictable
environment, with the liberating consequence of robotic free will.
In the not so long run, you as a human being have perhaps much
greater restrictions on your freedom and dignity than your robot.
http://mind.sourceforge.net/ai4u_157.html
is an overview of Mind.
Volition -- not yet coded
Screen # 79 ram:robot
--------------------------------
0 \ VOLITION stub 980505atm
1
2 \ The robot will have free will inasmuch as it matches its motor
3 \ options with its perceived or instinctual values. The process
4 \ of thinking about values or goals and how to attain them will
5 \ initiate the formulation of goal-directed motor strategies.
6
7 \ A robotic free will can be designed to pursue any suggested
8 \ motor option only when all related mental associations have
9 \ died down below a threshold level. If the associations cause
10 \ concern, fear or alarm, they will not acquiesce in the action.
11
12 \ If all pertinent thought ratifies a proposed motor action,
13 \ the MOTORIUM initiates an option.
14
15 \ VOLITION will return to the Main Program Loop MIND.
--------------------------------
8. Analysis of the Modus Operandi
In the Mind-1.1 release, the AI Mind has begun to take actions by
asking questions when confronted with previously unknown concepts.
However, these interrogative actions are not yet motor actions.
Even before the AI begins to govern motor outputs, some form of
the Volition module ought to be coded and put in charge of asking
questions, because to ask questions is to interact voluntarily
with the world. If the AI talks,
speaking is a motor initiative.
9. Troubleshooting and Robotic Psychosurgery
Try in advance not to introduce any
evolutionary bugs.
The
AI Debugger program may shed some light in general on how to debug
and troubleshoot programs in artificial intelligence.
10. Volition Resources for
Seed AI Germination and Evolution
