A novel associative memory for high level control functions

A novel associative memory (AM) knowledge base for real-time diagnostics and high-level control functions for complex large-scale dynamic systems is presented. The proposed AM architecture is called ASDM as it is an adaptive architecture based on the sparse distributed memory (SDM) first introduced by Kanerva (1988). ASDM is proposed to overcome many limitations of SDM, while keeping most of the advantages of the original SDM. The new model is adaptive since the memory cells are renamed according to a learning/storing process. The analysis of the best match problem can be carried out in a deterministic setting. Also introduced is the concept of time-varying intensity of memory, and generalized metrics are used for determining distance between two data objects. One major application of this ASDM is that of a neural expert system. With the ASDM, the machine can `learn´ rules on top of the knowledge acquired from human experts