Spiking Cellular Associative Neural Networks for Pattern Recognition

The Cellular Associative Neural Network (CANN) is a novel symbolic pattern match-ing system, currently used for both the identi¯cation of objects in noisy images and for graph similarity searching of chemical structures. Objects are de¯ned by a set of symbolic rules, which iteratively combine low level features into higher level con-structs, until object level de¯nitions can be obtained. The °ow of information follows a cellular automata based model to aid parallel implementation and rules are stored in AURA associative memories, which provide e±cient storage, fast retrieval and the ability to identify partially matching rules in constant time. This thesis reviews a series of investigations into the CANN, identifying weaknesses in the original design and proposing and evaluating a range of potential improve-ments to the architecture. These improvements include the incorporation of a spike based representation of uncertainty into the system, which in addition to locating the objects present in the recall data, also enables the CANN to associate a relative measure of con¯dence with each output label produced. A Spiking Neural Edge De