![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
Abstract
Artificial general intelligence is a field of research aiming to distil the principles of intelligence that operate independently of a specific problem domain and utilise these principles in order to synthesise systems capable of performing any intellectual task a human being is capable of and beyond. While “narrow” artificial intelligence which focuses on solving specific problems such as speech recognition, text comprehension, visual pattern recognition and robotic motion has shown impressive breakthroughs lately, understanding general intelligence remains elusive. We propose a paradigm shift from intelligence perceived as a competence of individual agents defined in relation to an a priori given problem domain or a goal, to intelligence perceived as a formative process of self-organisation. We call this process open-ended intelligence. Starting with a brief introduction of the current conceptual approach, we expose a number of serious limitations that are traced back to the ontological roots of the concept of intelligence. Open-ended intelligence is then developed as an abstraction of the process of human cognitive development, so its application can be extended to general agents and systems. We introduce and discuss three facets of the idea: the philosophical concept of individuation, sense-making and the individuation of general cognitive agents. We further show how open-ended intelligence can be framed in terms of a distributed, self-organising network of interacting elements and how such process is scalable. The framework highlights an important relation between coordination and intelligence and a new understanding of values.
Acknowledgements
This work was supported by the PhD scholarships from The Global Brain Institute under the Grant WDVO121 from Yuri Milner Fundation.
Notes
No potential conflict of interest was reported by the authors.
1 This is possibly what Goertzel means in his emphasis on complex environments in the definition.
2 Observations can be entirely passive but interactions are necessary in order to observe the effects of the agent’s actions on the environment.
3 A short introduction to assemblage theory can be found in: http://wikis.la.utexas.edu/theory/page/assemblage-theory.
4 Capacities also mean capacities to be destroyed. Certain interactions can amplify the internal unsettled intensities within an individual to the effect of its disintegration.
5 The scaling is not only structural but also temporal. Changes at different scales do not occur at the same frequency, also the stability of elements varies with complexity.
6 The designation “given” here is a simplification made for clarity. In fact, the elements in are never fully individuated and are affected by interactions taking place in its two adjacent strata as well.
7 Development generally means increase in intelligence in correlation to the complexity of situations and objects the system can make sense of. But the process is not necessarily monotonous; disintegration of already integrated structures can take place as well in the course of development. For example, when a theory is being replaced by a different, better theory that can explain and cohere more observations.
8 The tendency to form habits or repeating patterns of interaction is philosophically profound. It seems to indicate an ontological bias towards coordination over disparity, and more generally, of order over disorder. This goes back to transcendental empiricism being our point of departure. The co-emergence of observer and observed necessarily reflects an intrinsic bias (though temporary and local) towards order over disorder, otherwise neither observers nor observations could possibly emerge. Order, therefore is both self-evident and self-generative and so is the intelligence manifesting in it.
9 The distinction made here is clear only in the context of a single stratum but is much less apparent considering multiple strata as topological changes in one stratum lead to behavioural changes in the stratum above it.
10 For an early fascinating account of the idea of self-organisation in the sense described here, see Ashby (Citation1962).
11 All forms of conditioning including self-conditioning belong to this category as they establish correlations between an agent’s input and output signals.
12 The cybernetic nature of individuation was already discussed in Section 4.3, but here it is introduced in the more specific context of our model.
13 Such processes of individuation can become extremely complex. This example also demonstrates that considering a single price for a good is often a gross oversimplification. Prices of goods undergo an individuation process that is never exhausted especially if demand and supply are distributed and fluctuating.
15 Note that these are only simplified formulas that do not take into account the different sizes of subsets.