Abstract
We propose a novel learning strategy to train the optical neural networks based on neuroevolution. The accuracy of modulation formats recognition reach to 93%, indicating that the proposed method is competitive with other learning algorithms.
© 2019 The Author(s)
PDF ArticleMore Like This
Tyler W. Hughes, Momchil Minkov, Ian A. D. Williamson, Yu Shi, and Shanhui Fan
JF3F.2 CLEO: Applications and Technology (CLEO:A&T) 2019
Francis T. S. Yu, Taiwei Lu, and Don A. Gregory
MB4 Spatial Light Modulators and Applications (SLM) 1990
Xing Lin, Yair Riverson, Nezih T. Yardimci, Muhammed Veli, Yi Luo, Mona Jarrahi, and Aydogan Ozcan
SF1N.2 CLEO: Science and Innovations (CLEO:S&I) 2019