Lateral Information-Propagation Neural Networks for Inter-node Interpolation

摘要

Lateral Information-Propagation Neural Networks (LIPN) is proposed for on-line interpolation among neural nodes. Each node of LIPN corresponds to a state in a quantized input space and is composed of a processing unit and fixed space and is composed of a processing unit and fixed weights from its neighbor nodes as well as its input terminal. Information of a neural node propagates to neighbor nodes laterally through weight connections. Thus, inter-node interpolation is achieved. To test the feasibility of the prposed concept, 1-D hardware has been implemented with general purpose analog ICs. Experiments with static and dynamic signals have been done upon the LIPN hardware.