Meca Sapiens

From HandWiki

Meca Sapiens (from the Latin words mēchanicus, which means "mechanical", and sapiens which means "wise", i.e. "wise machine ") is a framework, in Artificial General Intelligence (AGI), whose aim is to implement synthetic consciousness. The framework is based on an understanding of consciousness as a system capability; its design stages followed a top-down development process; and it utilizes standard software engineering tools, structures and techniques.

A complete system architecture based on the Meca Sapiens framework [1] and suitable for design and implementation was published in late 2015. The creator of the Meca Sapiens architecture is J.E. Tardy.

Overview

Meca Sapiens differs from other research in artificial consciousness with respect to: its understanding of consciousness, the development process it follows and the tools and techniques it employs.

Consciousness as system capability

Meca Sapiens defines consciousness as an observable system capability whose formal component is the capacity of a system to generate, communicate and utilize absolute cognitive (non-sensory) representations of itself and its environment. In this understanding, the human experience does not define consciousness but is viewed as a particular instance of (a more general) category of "system consciousness".

This differs from a prevalent understanding of consciousness[2][3][4] as ontologically subjective[5] phenomena (see consciousness, hard problem of consciousness, artificial consciousness, neural correlates of consciousness) that motivate attempts to replicate, in synthetic structures, the subjective experiences (Qualia) of living creatures.[6][7][8] In Meca Sapiens, the subjective sensations of consciousness, experienced by humans, are of anecdotal importance; what matters is the observed behavior of a system in response to certain types of information.

Top-down design

The development of the Meca Sapiens architecture followed a top-down (or Waterfall) process where the objective (in this case synthetic consciousness) is first defined as achievable requirements and a complete system architecture of the solution is outlined before any implementation begins. This approach is characterized by a refusal to produce any software coding before a complete solution is first outlined at the system architecture level.

This development strategy diverges from the bottom-up approach adopted in other research projects (i.e. IDA-LIDA,[9] Novamente-OpenCog[10]) where coding of a partial version begins early and is followed by continuing attempts to expand the initial prototype in generality and scope.

Standard software techniques

The Meca Sapiens architecture is based on the software engineering tools and techniques used in other information systems. It describes a purposefully designed embedded autonomous agent that will be conscious. The architecture makes no use of the structures or processes that take place in human (or animal) brains.[11] In particular, Artificial Neural Networks, whose role is central in many AGI related projects, are treated as stochastic optimization mechanisms.

Status

The intent to define consciousness as a system capability and implement it using standard techniques was first stated in 1989[12]

2008: launch of the Meca Sapiens project to develop the system architecture of a conscious machine.

2011: publication of a System Requirements Specification document defining machine consciousness in terms of specification objectives.[13]

2015: publication of a complete system architecture to implement synthetic consciousness.[1]

2016: a number of implementation have begun and are currently under way.

References

  1. 1.0 1.1 Tardy, J.E., The Meca Sapiens Blueprint, Sysjet, 2015, ISBN:978-2-9812184-6-9
  2. "Proponents of [Artificial Consciousness] believe it is possible to construct machines (e.g., computer systems) that can emulate this [Neural Correlates of Consciousness] interoperation." artificial consciousness
  3. John Searle (2005). "Consciousness". In Honderich T. The Oxford companion to philosophy. Oxford University Press. ISBN:978-0-19-926479-7.
  4. Baars, B. J. (1988). A cognitive theory of consciousness. Cambridge: Cambridge University Press
  5. "My own philosophy for understanding consciousness has been to [...] ask what would be necessary for a machine [...] to report similar internal sensations." Igor Aleksander in: http://footnote1.com/machine-consciousness-fact-or-fiction/, February 2014.
  6. Tononi, Giulio, Consciousness as Integrated Information: a Provisional Manifesto, Biological Bulletin 215: 216–242. (December 2008)
  7. Haikonen, Pentti (2003), The Cognitive Approach to Conscious Machines, Exeter, UK: Imprint Academic, ISBN:0-907845-42-8
  8. Graham-Rowe, Duncan. "Mission to build a simulated brain begins", NewScientist, June 2005.
  9. Franklin, S., & Patterson, F. G. J. (2006). The LIDA Architecture: Adding New Modes of Learning to an Intelligent, Autonomous, Software Agent IDPT-2006 Proceedings (Integrated Design and Process Technology): Society for Design and Process Science
  10. Hart, D; B Goertzel (2008). OpenCog: A Software Framework for Integrative Artificial General Intelligence (PDF). Proceedings of the First AGI Conference
  11. Thaler, S. L. (2014). Synaptic Perturbation and Consciousness, Int. J. Mach. Conscious., 06, 75 (2014). DOI: 10.1142/S1793843014400137.
  12. Monterège, J.T., The creation of digital consciousness, SIGART Newsletter, July 1989, No 109.
  13. Tardy, J.E., The Creation of a Conscious Machine, Createspace, 2009, ISBN:2981218417

External links