CMOS current-mode outstar neural networks with long-period analog ratio memory

An on-chip learnable neural network called the outstar is implemented in CMOS current-mode circuits. The on-chip memory called the ratio memory can retain the stored ratio values. The outstar chip is fabricated and tested. It has shown that the storage time of the ratio memory can be as long as 5 minutes without refreshment. Moreover, the outstar can normalize the input pattern and enhance the contrast of the stored pattern within a long period. The application of the outstar in pattern classification has been demonstrated.

[1]  José Luis Huertas,et al.  A CMOS analog adaptive BAM with on-chip learning and weight refreshing , 1993, IEEE Trans. Neural Networks.

[2]  Chung-Yu Wu,et al.  A new neural associative memory with learning , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.

[3]  Tetsuro Itakura,et al.  Neuro chips with on-chip back-propagation and/or Hebbian learning , 1992 .

[4]  S. Tam,et al.  An electrically trainable artificial neural network (ETANN) with 10240 'floating gate' synapses , 1990, International 1989 Joint Conference on Neural Networks.

[5]  S. Grossberg.  Nonlinear difference-differential equations in prediction and learning theory. , 1967, Proceedings of the National Academy of Sciences of the United States of America.