scholarly journals Artificial Consciousness as a Platform for Artificial General Intelligence

2019 ◽  
Author(s):  
Ryota Kanai ◽  
Ippei Fujisawa ◽  
Shinya Tamai ◽  
Atsushi Magata ◽  
Masahiro Yasumoto

In this paper, we propose a hypothesis that consciousness has evolved to serve as a platform for general intelligence. This idea stems from considerations of potential biological functions of consciousness. Here we define general intelligence as the ability to apply knowledge and models acquired from past experiences to generate solutions to novel problems. Based on this definition, we propose three possible ways to establish general intelligence under existing methodologies for constructing AI systems, namely solution by simulation, solution by combination and solution by generation. Then, we relate those solutions to putative functions of consciousness put forward, respectively, by the information generation theory, the global workspace theory, and a form of higher order theory where qualia are regarded as meta-representations. Based on these insights, We propose that consciousness integrates a group of specialized generative/forward models and forms a complex in which combinations of those models are flexibly formed and that qualia are meta-representations of first-order mappings which endow an agent with the ability to choose which maps to use to solve novel problems. These functions can be implemented as an ``artificial consciousness''. Such systems can generate policies based on a small number of trial and error for solving novel problems. Finally, we propose possible directions for future research into artificial consciousness and artificial general intelligence.

1976 ◽  
Vol 98 (1) ◽  
pp. 133-138 ◽  
Author(s):  
A. Maewal ◽  
T. C. Bache ◽  
G. A. Hegemier

Using a method developed for studying wave propagation problems, a continuum theory is developed for diffusion-type processes in a laminated composite with periodic micro-structure. Construction is based upon an asymptotic scheme in which a typical macrodimension is assumed large compared to a microdimension. The order of truncation of the asymptotic sequence so obtained defines a hierarchy of models. Solutions are given for the lowest-order models and compared with the results from a finite difference code. For most cases the zeroth-order “effective conductivity” theory yields good results. For exceptional problems requiring a higher-order theory, a modified version of the first-order theory is shown to suffice. For many applications these elementary equations may offer an attractive alternative to other means for obtaining solutions.


1973 ◽  
Vol 57 (1) ◽  
pp. 1-25 ◽  
Author(s):  
Noor Afzal

A higher order theory for two-dimensional turbulent boundary-layer flow of a compressible fluid past a plane wall is formulated, for moderately large values of the Reynolds number, by the method of matched asymptotic expansions. The parameters (γ − 1) M2∞and the molecular Prandtl number are assumed to be of order unity. The analysis deals with the set of Reynolds equations of mean motion (which are underdetermined without an additional set of closure hypotheses) and assumes that the non-dimensional fluctuations in velocity, temperature and density are of orderU*, (friction velocity divided by free-stream velocity a t some designation point), while fluctuations in pressure are of orderU2*.The first-order results of the present study lead to asymptotic laws for velocity and temperature distributions which correspond to the law of the wall, logarithmic law and defect law, and also to skin friction and heat-transfer laws. It turns out that the first-order defect law depends upon the gradient of entropy and stagnation enthalpy and the law of the wall is independent of viscous dissipation. The second-order terms of the present work (accounting for mean convection due to turbulent mass flux, viscous dissipation in the inner flow and displacement effects in the outer flow) describe the necessary corrections to first-order terms due to low Reynolds number effects. In the overlap region the second-order results, for the law of the wall and the defect law, show bilogarithmic terms along with logarithmic terms.


AI Magazine ◽  
2012 ◽  
Vol 33 (1) ◽  
pp. 25-42 ◽  
Author(s):  
Sam Adams ◽  
Itmar Arel ◽  
Joscha Bach ◽  
Robert Coop ◽  
Rod Furlan ◽  
...  

We present the broad outlines of a roadmap toward human-level artificial general intelligence (henceforth, AGI). We begin by discussing AGI in general, adopting a pragmatic goal for its attainment and a necessary foundation of characteristics and requirements. An initial capability landscape will be presented, drawing on major themes from developmental psychology and illuminated by mathematical, physiological and information processing perspectives. The challenge of identifying appropriate tasks and environments for measuring AGI will be addressed, and seven scenarios will be presented as milestones suggesting a roadmap across the AGI landscape along with directions for future research and collaboration.


2019 ◽  
Author(s):  
Richard Brown ◽  
Hakwan Lau ◽  
Joseph LeDoux

Critics have often misunderstood the higher-order theory (HOT) of consciousness. Here we clarify its position on several issues, and distinguish it from other views such as the global workspace theory (GWT) and early sensory models, such as first-order local recurrency theory. The criticism that HOT overintellectualizes conscious experience is inaccurate because in reality the theory assumes minimal cognitive functions for consciousness; in this sense it is an intermediate position between GWT and early sensory views, and plausibly accounts for shortcomings of both. Further, compared to other existing theories, HOT can more readily account for complex everyday experiences, such as of emotions and episodic memories, and make HOT potentially useful as a framework for conceptualizing pathological mental states.


2021 ◽  
Author(s):  
◽  
Wilfred Gordon Malcolm

<p>The programme of work for this thesis began with the somewhat genenal intention of parallelling in the context of higher order models the ultraproduct construction and its consequences as developed in the literature for first order models. Something of this was, of course, already available in the ultrapower construction of W.A.J. Luxemburg used in Non Standand Analysis. It may have been considered that such a genenal intention was not likely to yield anything of significance oven and above what was already available from viewing the higher order situation as a 'many sorted' first order one and interpreting the first order theory accordingly. In the event, however, I believe this has proved not to be so. In particular the substructure concepts developed in Chapter II of this thesis together with the various embedding theorems and their applications are not immediately available fnom the first order theory and seem to be of sufficient worth to warrant developing the higher order theory in its own terms. This, anyway, is the basic justification for the approach and content of the thesis.</p>


2011 ◽  
Vol 26 (26) ◽  
pp. 4661-4686
Author(s):  
S. A. MARTÍNEZ ◽  
R. MONTEMAYOR ◽  
L. F. URRUTIA

We present an alternative method for constructing a consistent perturbative low energy canonical formalism for higher-order time-derivative theories, which consists in applying the standard Dirac method to the first-order version of the higher-order Lagrangian, augmented by additional perturbative Hamiltonian constraints. The method is purely algebraic, provides the dynamical formulation directly in phase space and can be used in singular theories without the need of initially fixing the gauge. We apply it to two paradigmatic examples: the Pais–Uhlenbeck oscillator and the Bernard–Duncan scalar field with self-interaction. We also compare the results, both at the classical and quantum level, with the ones corresponding to a direct perturbative construction applied to the exact higher-order theory, after incorporating the projection to the space of physical modes. This comparison highlights the soundness of the present formalism.


1996 ◽  
Vol 307 ◽  
pp. 135-165 ◽  
Author(s):  
M. A. Jog ◽  
P. S. Ayyaswamy ◽  
I. M. Cohen

The evaporation and combustion of a single-component fuel droplet which is moving slowly in a hot oxidant atmosphere have been analysed using perturbation methods. Results for the flow field, temperature and species distributions in each phase, inter-facial heat and mass transfer, and the enhancement of the mass burning rate due to the presence of convection have all been developed correct to second order in the translational Reynolds number. This represents an advance over a previous study which analysed the problem to first order in the perturbation parameter. The primary motivation for the development of detailed analytical/numerical solutions correct to second order arises from the need for such a higher-order theory in order to investigate fuel droplet ignition and extinction characteristics in the presence of convective flow. Explanations for such a need, based on order of magnitude arguments, are included in this article. With a moving droplet, the shear at the interface causes circulatory motion inside the droplet. Owing to the large evaporation velocities at the droplet surface that usually accompany drop vaporization and burning, the entire flow field is not in the Stokes regime even for low translational Reynolds numbers. In view of this, the formulation for the continuous phase is developed by imposing slow translatory motion of the droplet as a perturbation to uniform radial flow associated with vigorous evaporation at the surface. Combustion is modelled by the inclusion of a fast chemical reaction in a thin reaction zone represented by the Burke–Schumann flame front. The complete solution for the problem correct to second order is obtained by simultaneously solving a coupled formulation for the dispersed and continuous phases. A noteworthy feature of the higher-order formulation is that both the flow field and transport equations require analysis by coupled singular perturbation procedures. The higher-order theory shows that, for identical conditions, compared with the first-order theory both the flame and the front stagnation point are closer to the surface of the drop, the evaporation is more vigorous, the droplet lifetime is shorter, and the internal vortical motion is asymmetric about the drop equatorial plane. These features are significant for ignition/extinction analyses since the prediction of the location of the point of ignition/extinction will depend upon such details. This article is the first of a two-part study; in the second part, analytical expressions and results obtained here will be incorporated into a detailed investigation of fuel droplet ignition and extinction. In view of the general nature of the formulation considered here, results presented have wider applicability in the general areas of interfacial fluid mechanics and heat/material transport. They are particularly useful in microgravity studies, in atmospheric sciences, in aerosol sciences, and in the prediction of material depletion from spherical particles.


2019 ◽  
Author(s):  
Hakwan Lau

I introduce an empirically-grounded version of a higher-order theory of conscious perception. Traditionally, theories of consciousness either focus on the global availability of conscious information, or take conscious phenomenology as a brute fact due to some biological or basic representational properties. Here I argue instead that the key to characterizing the consciousness lies in its connections to belief formation and epistemic justification on a subjective level.


Sign in / Sign up

Export Citation Format

Share Document