A biologically motivated mechanism for self-organizing a neural network with modifiable lateral connections is presented. The weight modification rules are purely activity-dependent, unsupervised and local. The lateral interaction weights are initially random but develop into a Mexican hat'' shape around each neuron. At the same time, the external input weights self-organize to form a topological map of the input space. The algorithm demonstrates how self-organization can bootstrap itself using input information. Predictions of the algorithm agree very well with experimental observations on the development of lateral connections in cortical feature maps.

PDF, PS

In Proceedings of the IEEE International Conference on Neural Networks (San Francisco, CA), 1360-1365, 1993. Piscataway, NJ: IEEE.

@InProceedings{sirosh:lateral, title={How Lateral Interaction Develops In A Self-Organizing Feature Map}, author={Joseph Sirosh and Risto Miikkulainen}, booktitle={Proceedings of the IEEE International Conference on Neural Networks (San Francisco, CA)}, publisher={Piscataway, NJ: IEEE}, pages={1360-1365}, url="http://nn.cs.utexas.edu/?sirosh:icnn93", year={1993} }