The self-organized map, an architecture suggested for artificial neural networks, is explained by presenting simulation experiments and practical applications. The self-organizing map has the property of effectively creating spatially organized internal representations of various features of input signals and their abstractions. One result of this is that the self-organization process can discover semantic relationships in sentences. Brain maps, semantic maps, and early work on competitive learning are reviewed. The self-organizing map algorithm (an algorithm which order responses spatially) is reviewed, focusing on best matching cell selection and adaptation of the weight vectors. Suggestions for applying the self-organizing map algorithm, demonstrations of the ordering process, and an example of hierarchical clustering of data are presented. Fine tuning the map by learning vector quantization is addressed. The use of self-organized maps in practical speech recognition and a simulation experiment on semantic mapping are discussed. >

The self-organizing map

Criticism and the Growth of Knowledge. Edited by I. Lakatos and A. Musgrave. Pp. viii, 282. £3·50, paperback £1. 1970. (Cambridge University Press.)

Dual-process and dual-system theories in both cognitive and social psychology have been subjected to a number of recently published criticisms. However, they have been attacked as a category, incorrectly assuming there is a generic version that applies to all. We identify and respond to 5 main lines of argument made by such critics. We agree that some of these arguments have force against some of the theories in the literature but believe them to be overstated. We argue that the dual-processing distinction is supported by much recent evidence in cognitive science. Our preferred theoretical approach is one in which rapid autonomous processes (Type 1) are assumed to yield default responses unless intervened on by distinctive higher order reasoning processes (Type 2). What defines the difference is that Type 2 processing supports hypothetical thinking and load heavily on working memory.

/pdf/dual-process-theories-of-higher-cognition-advancing-the-5k4q1zq7ey.pdf

Dual-Process Theories of Higher Cognition: Advancing the Debate

Many current neurophysiological, psychophysical, and psychological approaches to vision rest on the idea that when we see, the brain produces an internal representation of the world. The activation of this internal representation is assumed to give rise to the experience of seeing. The problem with this kind of approach is that it leaves unexplained how the existence of such a detailed internal representation might produce visual consciousness. An alternative proposal is made here. We propose that seeing is a way of acting. It is a particular way of exploring the environment. Activity in internal representations does not generate the experience of seeing. The outside world serves as its own, external, representation. The experience of seeing occurs when the organism masters what we call the governing laws of sensorimotor contingency. The advantage of this approach is that it provides a natural and principled way of accounting for visual consciousness, and for the differences in the perceived quality of sensory experience in the different sensory modalities. Several lines of empirical evidence are brought forward in support of the theory, in particular: evidence from experiments in sensorimotor adaptation, visual \"filling in,\" visual stability despite eye movements, change blindness, sensory substitution, and color perception.

A sensorimotor account of vision and visual consciousness-Authors' Response-Acting out our sensory experience

Language, Thought, and Other Biological Categories

Successful recognition of natural signals, e.g., speech recognition, requires substantial statistical pattern recognition capabilities. This is at odds with the fact that the bulk of work on applying neural networks to pattern recognition has concentrated on non-statistical problems. Three basic types of neural-like networks (Backpropagation network, Boltzmann machine, and Learning Vector Qumtization), were applied in this work to two representative artificial statistical pattern recognition tasks, each with varying dimensionality. The performance of each network's different approach to solving the tasks was evaluated and compared, both to the performance of the other two networks, and to the theoretical limit. The Learning Vector Quantization was further benchmarked against the parametric Bayes classifier and the k-nearestneighbor classifier using natural speech data. A novel Learning Vector Quantization classifier (LVQ2) is introduced the first time in this work.

/pdf/statistical-pattern-recognition-with-neural-networks-5f0bqs3koz.pdf

Statistical pattern recognition with neural networks: benchmarking studies.

Replication or even modelling of consciousness in machines requires some clarifications and refinements of our concept of consciousness. Design of, construction of, and interaction with artificial systems can itself assist in this conceptual development. We start with the tentative hypothesis that although the word 'consciousness' has no well-defined meaning, it is used to refer to aspects of human and animal information processing. We then argue that we can enhance our understanding of what these aspects might be by designing and building virtual- machine architectures capturing various features of consciousness. This activity may in turn nurture the development of our concepts of consciousness, showing how an analysis based on information processing virtual machines answers old philosophical puzzles as well enriching empirical theories. This process of developing and testing ideas by developing and testing designs leads to gradual refinement of many of our pre-theoretical concepts of mind, showing how they can be construed as implicitly 'architecture-based' concepts. Understanding how human-like robots with appropriate architectures are likely to feel puzzled about qualia may help us resolve those puzzles. The concept of 'qualia' turns out to be an 'architecture-based' concept, while individual qualia concepts are 'architecture-driven'.

/pdf/virtual-machines-and-consciousness-3gztlmpisi.pdf

Virtual Machines and Consciousness

http://www.cs.bham.ac.uk/research/projects/cogaff/chrisley-embodied-ai.pdf

Embodied artificial intelligence

Much discussion of emotions and related topics is riddled with confusion because different authors use the key expressions with different meanings. Some confuse the concept of "emotion" with the more general concept of "affect", which covers other things besides emotions, including moods, attitudes, desires, preferences, intentions, dislikes, etc. Moreover researchers have different goals: some are concerned with understanding natural phenomena, while others are more concerned with producing useful artifacts, e.g. synthetic entertainment agents, sympathetic machine interfaces, and the like. We address this confusion by showing how "architecture-based" concepts can extend and refine our pre-theoretical concepts in ways that make them more useful both for expressing scientific questions and theories, and for specifying engineering objectives. An implication is that different information-processing architectures support different classes of emotions, different classes of consciousness, different varieties of perception, and so on. We start with high level concepts applicable to a wide variety of types of natural and artificial systems, including very simple organisms, namely concepts such as "need", "function", "information-user", "affect", "information-processing architecture". For more complex architectures, we offer the CogAff schema as a generic framework which distinguishes types of components that may be in a architecture, operating concurrently with different functional roles. We also sketch H-Cogaff, a richly-featured special case of CogAff, conjectured as a type of architecture that can explain or replicate human mental phenomena. We show how the concepts that are definable in terms of such architectures can clarify and enrich research on human emotions. If successful for the purposes of science and philosophy the architecture is also likely to be useful for engineering purposes, though many engineering goals can be achieved using shallow concepts and shallow theories, e.g., producing "believable" agents for computer entertainments. The more human-like robot emotions will emerge, as they do in humans, from the interactions of many mechanisms serving different purposes, not from a particular, dedicated "emotion mechanism".

/pdf/the-architectural-basis-of-affective-states-and-processes-3b1xbhpmor.pdf

The Architectural Basis of Affective States and Processes

Some have suggested that there is no fact to the matter as to whether or not a particular physical system realizes a particular computational description. This suggestion has been taken to imply that computational states are not "real", and cannot, for example, provide a foundation for the cognitive sciences. In particular, Putnam has argued that every ordinary open physical system realizes every abstract finite automaton, implying that the fact that a particular computational characterization applies to a physical system does not tell one anything about the nature of that system. Putnam's argument is scrutinized, and found inadequate because, among other things, it employs a notion of causation that is too weak. I argue that if one's view of computation involves embeddedness (inputs and outputs) and full causality, one can avoid the universal realizability results. Therefore, the fact that a particular system realizes a particular automaton is not a vacuous one, and is often explanatory. Furthermore, I claim that computation would not necessarily be an explanatorily vacuous notion even if it were universally realizable. Key words. Computation, philosophy of computation, embeddedness, foundations of cognitive science, formality, multiple realization.

/pdf/why-everything-doesn-t-realize-every-computation-voixwc8kd9.pdf

Ron Chrisley

Papers

Statistical pattern recognition with neural networks: benchmarking studies.

Virtual Machines and Consciousness

Embodied artificial intelligence

The Architectural Basis of Affective States and Processes

Why everything doesn't realize every computation