Abstract: Usually, generalization is considered as a function of learning from a set of examples. In present work on the basis of recent neural network assembly memory model (NNAMM), a biologically plausible 'grandmother' model for vision, where each separate memory unit itself can generalize, has been proposed. For such a generalization by computation through memory, analytical formulae and numerical procedure are found to calculate exactly the perfectly learned memory unit's generalization ability. The model's memory has complex hierarchical structure, can be learned from one example by a one-step process, and may be considered as a semi-representational one. A simple binary neural network for bell-shaped tuning is described.

Keywords: generalization, 'grandmother' model for vision, neural network assembly memory model, one-step learning, learning from one example, neuron receptive field, bell-shaped tuning, semi-representation.

ACM Classification Keywords: Memory structures (B.3), associative memories; reliability, testing, and faulttolerance (B.8.1); learning (I.2.6), connectionism and neural nets; vision and scene understanding (I.2.10), representations, data structures, and transforms; image representation (I.4.10), hierarchical.

Link:

GENERALIZATION BY COMPUTATION THROUGH MEMORY

Petro Gopych

http://www.foibg.com/ijita/vol13/ijita13-2-p07.pdf