Preliminary Investigation of Self-Organization Pattern Mapping System Based on Photochromism

An optical neural network system was proposed utilizing the photochromic materials as optically addressed spatial light modulator (OASLM) which preserves the analog data stably. By combining this OASLM and the conventional SLM (SLM-OASLM strategy), the system controls massive weights of interconnection, which are represented as the local photoisomerization in the OASLM, quite efficiently in a parallel and active manner. A series of computer simulations are carried out in terms of the self-organization pattern mapping (SOPM) as an example of application to optimize the model parameters and to check the noise stability. Based on the results of simulation and the physical properties of photochromic materials, the practical data processing performance of the system is estimated.