Using self-organising feature maps for the control of artificial organisms

Ball, Nigel R.; Warwick, Kevin

doi:10.1049/ip-d.1993.0025

Cited by 9 publications

(5 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Operation of the HLS system has been documented in a number of papers [2,4]. Recent research has concentrated on a comparison of HLS's performance against Holland and Reitman's CS-1 system [6].…”

Section: Application Resultsmentioning

confidence: 99%

“…Both these variations have been incorporated into an hierarchical set of feature maps used to produce learning behaviour in a modified classifier system controlling artificial organisms [2][3][4].…”

Section: Kohonen Feature Mapsmentioning

confidence: 99%

“…The effective application of Kohonen self-organizing Feature Maps (KFM) to the control of artificial organisms has been demonstrated using a single map to provide an associative memory of the task environment [2][3][4]. The nature of the control problem demands the concurrent exploration and exploitation of the domain by the organism.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Self-organising neural networks for adaptive control

Warwick

Ball

1996

Journal of Intelligent and Robotic Systems

View full text Add to dashboard Cite

Self-organizing neural networks have been implemented in a wide range of application areas such as speech processing, image processing, optimization and robotics. Recent variations to the basic model proposed by the authors enable it to order state space using a subset of the input vector and to apply a tocal adaptation procedure that does not rely on a predefined test duration limit. Both these variations have been incorporated into a new feature map architecture that forms an integral part of an Hybrid Learning System (HLS) based on a genetic-based classifier system. Problems are represented within HLS as objects characterized by environmental features. Objects controlled by the system have preset targets set against a subset of their features. The system's objective is to achieve these targets by evolving a behavioural repertoire that efficiently explores and exploits the problem environment. Feature maps encode two types of knowledge within HLS-long-term memory traces of useful regularities within the environment and the classifier performance data calibrated against an object's feature states and targets. Self-organization of these networks constitutes non-genetic-based (experience-driven) learning within HLS. This paper presents a description of the HLS architecture and an analysis of the modified feature map implementing associative memory. Initial results are presented that demonstrate the behaviour of the system on a simple control task.

show abstract

Section: Application Resultsmentioning

confidence: 99%

Section: Kohonen Feature Mapsmentioning

confidence: 99%

See 1 more Smart Citation

Self-organising neural networks for adaptive control

Warwick

Ball

1996

Journal of Intelligent and Robotic Systems

View full text Add to dashboard Cite

show abstract

“…Further sensors will be added, allowing the insects to learn, for example, to follow specific objects or to be attracted to certain other insects and to change their behavioural pattern dependent on external factors such as heat and light. It should be stressed that the insects are autonomous, physical entities and have Taking things a stage further [12], a hybrid learning system has been devised which includes a Kohonen net as a long-term memory device for an artificial organism living inside a computer-world maze. Essentially the organism must move around its maze and locate food and water sources.…”

Section: Neural Network and Intelligencementioning

confidence: 99%

“…The artificial organism can learn quickly [12] where food and water are in its maze, indeed its life depends on it, and further when the sources are moved to other locations it takes some time to adjust, returning at less frequent intervals to the previous food and water locations.…”

Section: Neural Network and Intelligencementioning

confidence: 99%

Cybernetics – The Modern Science of Systems

Warwick

1994

Kybernetes

View full text Add to dashboard Cite

IntroductionCybernetics has been taught as a degree subject at Reading University for over 30 years and the Department of Cybernetics has been in existence since 1977. Over the last few years the student intake has escalated, applications are continuing to increase and the quality of students entering is improving annually -this is very much against the overall national trend in applied science, but is directly indicative of the fact that an increasing number of potential university science students are excited by the prospect of studying cybernetics. Last year alone more students entered Reading University to study cybernetics and computer science than any other science discipline; also the quality of entry to cybernetics, in terms of examination results, was better than for any other science subject. The demand to study the subject is clear.Research too has escalated such that the figure for research income (finance) per faculty member per year is now higher in cybernetics than for any other department at Reading University. Indeed, on this basis, the Department is competitive with any other applied science department in the world. Apart from UK Research Council support much finance is received from industry, present examples being British Gas, British Petroleum, British Aerospace, National Grid, Brooke Bond, Sony, GEC, Molins, SUN Microsystems and Toshiba, together with a host of smaller concerns. Reasons for this support are clear -in each case a way has been found to make use of cybernetics within the company to make the company more profitable. In some cases this is in terms of a relatively low-level system, in some it is the interaction between humans and technology, and in others it is concerned with optimization of the company as a system.One exciting avenue of research in the Department of Cybernetics at Reading is that concerned with technical aids for disabled people, an area which, more than any other, encompasses just what cybernetics is all about. A project linked with Possum Controls and Siemens is developing a smart system to enable a disabled person to control many domestic functions from wherever the person is, despite their, possibly severe, disability. This project, although at present concerned solely with a system involving a disabled person, opens up a whole area of study towards houses of the future with voice-controlled functions, e.g.

show abstract

An Overview of SOM Literature

Kohonen¹

2001

Self-Organizing Maps

View full text Add to dashboard Cite

Using self-organising feature maps for the control of artificial organisms

Cited by 9 publications

References 1 publication

Self-organising neural networks for adaptive control

Self-organising neural networks for adaptive control

Cybernetics – The Modern Science of Systems

An Overview of SOM Literature

Contact Info

Product

Resources

About