Symbolic and Sub-Symbolic Representations in Computational Models of Human Cognition

Kelley, Troy D.

doi:10.1177/0959354303136005

Cited by 37 publications

(23 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Procedural memory was unconscious, episodic, and instance-based while declarative memory was more conscious, semantic, and fact based. This type of separation also maps well onto the symbolic and subsymbolic distinctions of knowledge organization [34]. Procedural memory is easily characterized as subsymbolic, while declarative memory is more easily characterized as symbolic.…”

Section: B Psychologymentioning

confidence: 60%

Review of Consciousness and the Possibility of Conscious Robots

Long

Kelley

2010

Journal of Aerospace Computing, Information, and Communication

Self Cite

View full text Add to dashboard Cite

This paper discusses the psychological, philosophical, and neurological definitions of consciousness and the prospects for the development of conscious machines or robots in the foreseeable future. Various definitions of consciousness are introduced and discussed within the different fields mentioned. A conscious machine or robot may be within the realm of engineering possibilities if current technological developments, especially Moore's law, continue at their current pace. Given the complexity of cognition and consciousness a hybrid parallel architecture with significant input/output appears to offer the best solution for the implementation of a complex system of systems which functionally approximates a human mind. Ideally, this architecture would include traditional symbolic representations as well as distributed representations which approximate the nonlinear dynamics seen in the human brain.

show abstract

Section: B Psychologymentioning

confidence: 60%

Review of Consciousness and the Possibility of Conscious Robots

Long

Kelley

2010

Journal of Aerospace Computing, Information, and Communication

Self Cite

View full text Add to dashboard Cite

show abstract

“…This appearance should arise naturally from the organization of the network and the way of information representation and not from any separable layers of a composite system (As an attempt towards this goal Smolensky et al used tensor product representation; this approach is not discussed here.). A similar integrated symbolic and sub-symbolic approach is also advocated by Kelley [8].…”

Section: Sub-symbolic and Symbolic Cognitionmentioning

confidence: 94%

The Role of Associative Processing in Cognitive Computing

Haikonen

2009

Cogn Comput

View full text Add to dashboard Cite

The traditional approaches-of symbolic artificial intelligence (AI) and of sub-symbolic neural networks-towards artificial cognition have not been very successful. The rule-based symbolic AI approach has proven to be brittle and unable to provide any real intelligence (Mckenna, Artificial intelligence and neural networks: steps toward principled integration, Academic Press, USA, 1994). On the other hand, traditional artificial neural networks have not been able to advance very much beyond pattern recognition and classification. This shortcoming has been credited to the inability of conventional artificial neural networks to handle syntax and symbols. Hybrid approaches that combine symbolic AI and sub-symbolic neural networks have been tried with results that fall short of the ultimate goal. It has been argued that traditional AI programs do not operate with meanings and consequently do not understand anything (Searle, Minds, brains & science, Penguin Books Ltd, London, 1984; Searle, The mystery of consciousness, Granta Books, London, 1997). It seems that in this way some essential ingredient is missing, but there may be a remedy available. Associative information processing principles may enable the utilization of meaning and the combined sub-symbolic/symbolic operation of neural networks.

show abstract

“…However, we are then faced with the problem of how to integrate stochastic information with logic; the classical problem of symbolic/subsymbolic integration [8], [9].…”

Section: Symbolic Processing Within P-a Learning Systemsmentioning

confidence: 99%

A Framework for Hierarchical Perception–Action Learning Utilizing Fuzzy Reasoning

Windridge

Felsberg

Shaukat

2013

IEEE Trans. Cybern.

View full text Add to dashboard Cite

Abstract-Perception-Action (P-A) learning is an approach to cognitive system building that seeks to reduce the complexity associated with conventional environment-representation/actionplanning approaches. Instead, actions are directly mapped onto the perceptual transitions that they bring about, eliminating the need for intermediate representation and significantly reducing training requirements. We here set out a very general learning framework for cognitive systems in which online learning of the P-A mapping may be conducted within a symbolic processing context, so that complex contextual reasoning can influence the P-A mapping. In utilizing a variational calculus approach to define a suitable objective function, the P-A mapping can be treated as an on-line learning problem via gradient descent using partial derivatives. Our central theoretical result is to demonstrate topdown modulation of low-level perceptual confidences via the Jacobian of the higher levels of a subsumptive P-A hierarchy. Thus, the separation of the Jacobian as a multiplying factor between levels within the objective function naturally enables the integration of abstract symbolic manipulation in the form of fuzzy deductive logic into the P-A mapping learning. We experimentally demonstrate that the resulting framework achieves significantly better accuracy than using perception-action learning without top-down modulation. We also demonstrate that it permits novel forms of context-dependant multi-level P-A mapping, applying the mechanism in the context of an intelligent driver assistance system.

show abstract

Symbolic and Sub-Symbolic Representations in Computational Models of Human Cognition

Cited by 37 publications

References 31 publications

Review of Consciousness and the Possibility of Conscious Robots

Review of Consciousness and the Possibility of Conscious Robots

The Role of Associative Processing in Cognitive Computing

A Framework for Hierarchical Perception–Action Learning Utilizing Fuzzy Reasoning

Contact Info

Product

Resources

About