Holographic, Nonlinear Optical Implementations of Neural Network Models

这项工作涉及联想记忆的光学实现，因此属于神经工程领域。基本的联想记忆已经被证明了，尽管围绕这一特殊方法还存在一定的争议。尽管如此，首席调查员已经证明，可以使用光学技术从图像的一部分提取完整图像，并使用非线性和反馈来实现这一点。这项工作还将解决设计新奇过滤器的可能性。这是一种对被寻址对象的运动敏感的内容可寻址处理器。

This work relates to optical implementation of associative memory and as such is within the area of neuro-engineering. The basic associative memory has been demonstrated although there is a certain amount of controversy surrounding this particular approach. The princial investigator has nonetheless shown that optical techniques can be used to extract a full image from a portion of that image and uses nonlinearity and feedback to do that. The work will also address the possibility of designing a novelty filter. This is a content addressable processor which is sensitive to the motion of the object addressed.