We have constructed an optical neural-network system with learning cap
ability by using a Pockels readout optical modulator. The system has a
two-dimensional structure that permits easy optical alignment and can
handle images without scanning. Learning signals are calculated optic
ally with two Liquid-crystal devices by a matrix-matrix outer-product
method. The calculated learning signals are added directly to the weig
hts memorized on the Pockels readout optical modulator. A two-layer ne
twork is implemented, and the learning and recognition of four alphabe
tical characters are realized according to the delta rule. (C) 1998 Op
tical Society of America.