This model shows how spatial representations in 3D space could emerge using unsupervised neural networks. Model is a hierarchical one which means that it has multiple layers, where each layer has got a specific function to achieve. This architecture is more of a generalised one i.e. it gives rise to different kinds of spatial representations after training.
Model Type: Connectionist Network
Cell Type(s): Abstract rate-based neuron
Simulation Environment: MATLAB
Implementer(s): Soman, Karthik [karthi.soman at gmail.com]