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

Variations on the standard Kohonen feature map can enable an ordering of the map state space by using only a limited subset of the complete input vector. Also it is possible to employ merely a local adaptation procedure to order the map, rather than having to rely on global variables and objectives. Such variations have been included as part of a hybrid learning system (HLS) which has arisen out of a genetic-based classifier system. In the paper a description of the modified feature map is given, which constitutes the HLSs long term memory, and results in the control of a simple maze running task are presented, thereby demonstrating the value of goal related feedback within the overall network.

Object detecting artificial retina

This paper describes the novel use of agent and cellular neural Hopfield network techniques in the design of a self-contained, object detecting retina. The agents, which are used to detect features within an image, are trained using the Hebbian method which has been modified for the cellular architecture. The success of each agent is communicated with adjacent agents in order to verify the detection of an object. Initial work used the method to process bipolar images. This has now been extended to handle grey scale images. Simulations have demonstrated the success of the method and further work is planned in which the device is to be implemented in hardware.

Tissue engineering a fetal membrane

The aim of this study was to construct an artificial fetal membrane (FM) by combination of human amniotic epithelial stem cells (hAESCs) and a mechanically enhanced collagen scaffold containing encapsulated human amniotic stromal fibroblasts (hASFs). Such a tissue-engineered FM may have the potential to plug structural defects in the amniotic sac after antenatal interventions, or to prevent preterm premature rupture of the FM. The hAESCs and hASFs were isolated from human fetal amniotic membrane (AM). Magnetic cell sorting was used to enrich the hAESCs by positive ATP-binding cassette G2 selection. We investigated the use of a laminin/fibronectin (1:1)-coated compressed collagen gel as a novel scaffold to support the growth of hAESCs. A type I collagen gel was dehydrated to form a material mimicking the mechanical properties and ultra-structure of human AM. hAESCs successfully adhered to and formed a monolayer upon the biomimetic collagen scaffold. The resulting artificial membrane shared a high degree of similarity in cell morphology, protein expression profiles, and structure to normal fetal AM. This study provides the first line of evidence that a compacted collagen gel containing hASFs could adequately support hAESCs adhesion and differentiation to a degree that is comparable to the normal human fetal AM in terms of structure and maintenance of cell phenotype.