This paper articulates the fundamental theory of consciousness used in the Independent Core Observer Model (ICOM) research program and the consciousness measures as applied to ICOM systems and their uses in context including defining of the basic assumptions for the ICOM Theory of Consciousness (ICOMTC) and associated related consciousness theories (CTM, IIT, GWT etc.) that the ICOMTC is built upon. The paper defines the contextual experience of ICOM based systems in terms of a given instances subjective experience as objectively measured and the qualitative measure of Qualia in ICOM based systems
Designing an Artificial General Intelligence (AGI) System includes a number of
complex problems including defining elements that are not agreed upon to even have a
foundation to build such a design on in the first place. In the research program for the
cognitive architecture for AGI called the “Independent Core Observer Model” to be
able to frame tests and measures we needed to address definitions of consciousness
including baking out our own version of a theory of consciousness [
We made a number of assumptions to provide an experimental reference point
when designing AGI cognitive architecture [
Humans are not able to make logical decisions. Looking at the neuroscience
behind decisions we already can prove that humans make decisions based on how they
feel [
Subjective experience can be measured and understood. The traditional view that
the subjective nature of experience [
Consciousness can be measured. "Despite this enormous commitment to the study
of consciousness on the part of cognitive scientist covering philosophical,
psychological, neuroscientific and modelling approaches, as of now no stable models or
strategies for the adequate study of consciousness have emerged." [
We have a concrete definition of 'Subjective' as a concept. 'Subjective' then is defined as the relative experience of a conscious point of view that can only be measured objectively only from outside the system where the system in question experiences things 'subjectively' as they relate to that systems internal emotional context.
Consciousness is a system that exhibits the degrees or elements of the Porter
method for measuring consciousness regarding its internal subjective experience. [
The ICOM or Independent Core Observer Model Theory of Consciousness
(ICOMTC) is based on the Computational Theory of Mind [
. An instance of an ICOM system would be a variational instance of CCTM. In
addition to that, the ICOM Theory of Consciousness or ICOMTC also borrows from
the Integrated Information Theory [
Integrated information theory or IIT, approaches the relationship between
consciousness and its physical substrate by first identifying the fundamental proper-ties
of experience itself: existence, composition, information, integration, and exclusion.
IIT then postulates that the physical substrate of consciousness must satisfy three key
points or 'Axioms' [
ICOMTC also borrows from Global Workspace theory in that things move through
the system and only when things reach a certain point is that bit of 'thought' or
'context' raised to the level of the conscious mind. [
At a very high level, ICOM as a cognitive architecture [
How do we then look at a system that experiences emotional, subjective experience objectively? The following set notation shows us a simple logical implementation of the last climb of "a thought" as it makes its rise from the depths of the system to the awareness of the conscious, self-aware parts of the system. , ,
; (NewContext, ;
First, let us walk through the execution of this logic. Coming into the system we already have context data decomposition, sensory input, also related data from memory that may be of emotional interest but for the purposes of one 'thought' let's say it's one bit of context meaning an emotionally related context tree related to something that the system has sensed externally. This will be represented by 'Inputs.' At this point, we have already passed the point of that 'context' being raised to the global workspace. Figure 1 essentially is one cycle of the core considering what is in the global workspace or 'core' of ICOM. In Figure 1 we first see that we have two sets or collections of emotional models represented by the two sets defined in the first two rows, then we have the input new context placed in the 'NewContext' set. We apply the 'Needs' function that applies a matrix set of rules such as the technical requirements of the system to other wants and needs based on the systems hierarchy of needs and current environmental conditions. At this point, we look at how this thought applies conscious emotional rules in the function 'ConsciousRules' and then how that manipulates the current conscious emotional landscape. We say 'land-scape' because it is not a single emotion but a complex set of almost infinite combi-nations consciously and subconsciously that the system experiences.
In like manner, the system applies subconscious rules to the subconscious states and the subconscious rules to the conscious states and finally those states as they apply to the new context where in all cases it is only in the abstract from this states that the system experiences anything. Meaning the system is using the abstracted states to represent that emotional landscape in how things affect all of those emotional states and related context finally being passed to the observer for action if that 'NewContext' contained an action. In this way, the system doesn't even deal with the complexity of its actions as much as the system will do them if the system felt like it and knows how; where as numerous cycles might have to execute in the core for it to perform a new task, meaning it will have to think a lot more about something it doesn't know how to do. After that context is posted back to the observer (the more complex part of the system in ICOM), then it is placed back into context memory, and in this way, we see the rich set of the emotional landscapes of the system can model and execute.
Interestingly enough, in current ICOM research there are indications that this sort of system is perfectly capable of becoming mentally ill and even forgetful if hardware starts to limit operations, where as the only way to optimize for the execution environment would be to place memory limits and based on the node map memory models this would be the only way to continue optimal execution given certain limits.
A better way to think of ICOMTC is that not a single element of the system is conscious or self-aware to any level, it is the 'interactions' between the parts that together those interactions become aware abstractly, and it is through the underlying process articulated in Figure 1 that is then measured in terms of consciousness via the various methods as well as direct instrumentation of the system to measure 'qualia' for example. ,
In ICOMTC qualia can be objectively measured through the differential between the conscious emotional landscape of the system represented by a Plutchik model along with the subconscious model and the model of the irreducible set of any given con-text experienced by the system and the emotional model created that represents that specific 'contextual' experience. In the ICOMTC the qualia is that differential between the state and effective one emotional structure that represents that current context and how the system applies choices is then based on that and the numerous underlying factors that affect the construction and choices based on specific con-texts. Now by its nature the system can't self-reflect directly on those values but is an abstraction of that process in the global 'work space' that effectively is created by the underlying operation. We can of course measure this 'qualia' of the system but the system can't do it directly from its standpoint. In the research already done for ICOM we can see that ICOMTC system doesn't really have free will but it would appear that way from the systems standpoint and experience the illusion of free will much the way humans do.
As stated, qualia (in ICOM) then can be measured. Referring back to figure one we can use two values or sets from that set of operations and preform a 'qualia' measurement like this based on those values: , , ,
In this case we are computing qualia by taking the sets that represent the cur-rent emotional landscape of the system and a conscious and subconscious level and computing the difference matching sets where a set is a Plutchik model with 8 floating point values. We subtract the current state from the previous state giving us the Plutchik representation of the subjective emotional differential experienced by the system. This really gives you the numbers in terms of 'sets' that show how a specific element of 'context' that managed to make it to the global work space is 'experienced' or rather the effective of that experience. We actually have to calculate this after the fact external to the system as it is not actually computed in the real process (noted in figure 1) and there is not a 'direct' method in ICOM to surface an objective measure of qualia to the system without a complete abstraction but we can compute it external and use it for analysis.
The Independent Core Observer Model Theory of Consciousness (ICOMTC)
addresses key issues with being able to measure physical and objective details as well as
the subjective experience of the system (known as qualia) including mapping complex
emotional structures, as seen in previously published research related to ICOM
cognitive architecture [
There are two types of test types that are considered for use in the ICOM program designed around measuring and testing outside of the qualia analytics that are external measures. Keep in mind qualia under the Yampolskiy method as noted below is a different measure then the previous section. These tests us allow us to measure somewhat more subjective tasks based on our behavior of the system to further bake additional research. In both cases, these tests can be applied across various potentialy ‘conscious’ systems and humans giving us a frame of reference for comparison which we lack using the Qualia method form the previous sections. The two test types are: 7.1
Intelligence Quotient (IQ) tests -. are tests designed to measure ‘intelligence’ in
humans [
Collective Intelligence (CI) Test. – we would like to use this test, however the
information for executing this test is not publicly accessible and reaching out to the
researchers that created this test has produced no response thus far. [
A number of tests or measures will be collected, more oriented towards analysis for further study, primarily around correlative purposes. None of these tests may be used outside of as possible illustrative examples, without being statistically valid given the lack rigor or subjective nature of these measures.
The Turing Test –. this test is not considered quantifiable and there is debate over whether this measure tells us anything of value, however a test regimen for this has been completed and can be used for subjective analysis only.
The Porter Method –. This appears to be a qualitative test, but individual question
measures are entirely subjective and therefore the test lacks the level of
qualitativeness to be valid without a pool of historical values to measure against at the very least.
This test provides some value in meeting colloquial standards of consciousness and is
more comprehensive then some of the other tests albeit subjective it is at least the
attempt at being a comprehensive measure of consciousness. [
The Yampolskiy Qualia Test –. is a subjective measure of a subjective ‘thing’ and
therefore not a qualitative measure, however we have built a regimen based on this
when looking at qualia as measured in the previous examples. In theory this only tests
for presence of Qualia in human like subjects, passing this test does not mean that a
subject does not experience qualia in the sense of this paper, just that it was not
detected. This means that subjects may show signs of qualia, or not, but the test only
would show the presence of not the absence of qualia. [
There are numerous potential methods for measure consciousness in ICOM or other AGI systems but there is a lot of discord in terms of a foundation to build on. For ICOM to move forward as a cognitive architecture it is important that we build that foundation. The tests listed in this paper are the ones currently being used or being considered. There are a few others including mental health states analysis tests for example but these tests are more subjective and not helpful in the current research. We have found that by building on the basis articulated here that we at least can move the research forward for the time being.
Conclusions based on the this work at least have a frame of reference even if later
proved false then we at least know that and can start over. Ideally this foundation
gives us a deeper more rigorous program moving forward in terms of AGI Cognitive
Architecture research into conscious systems with an eye towards safety [